Een Lamp Ophangen voor €75: Wat er achter 65Klusser zit
Op 20 oktober 2026 meldde Sleutelstad dat 65Klusser ook in Leiden en omgeving begint. Het concept is eenvoudig. Gepensioneerde vakmensen doen kleine klussen bij mensen thuis: een lamp ophangen, een kast in elkaar zetten, een beugel in de douche, een klemmende deur.
Dat voorziet in een echte behoefte. Er blijven in veel huizen klussen liggen. Er zijn veel 65-plussers met handen en ervaring die niet stil willen zitten.
Deze blog zoekt uit wat er achter de dienst zit, waar het geld blijft en welke alternatieven er zijn. Daarna volgt een vergelijking met een heel andere vorm: de MAZE-Bank, een kring waarin daden worden geboekt in plaats van euro’s.
Alle getallen zijn nagerekend. Wat niet gevonden is, staat er als open bij.
Hoe 65Klusser werkt
De dienst werkt in vier stappen.
De klant belt of vult een formulier in.
Een werkopnemer komt kijken en bespreekt de klus.
De klant krijgt een prijsopgave op basis van €75 per uur, inclusief btw. Bij een goed af te bakenen klus kan dat een vaste prijs zijn.
Na akkoord krijgt de klant naam en foto van de klusser, per mail of WhatsApp. De klusser komt, doet de klus, en de klant betaalt achteraf.
De klussers hebben een Verklaring Omtrent het Gedrag (VOG). Ze zijn persoonlijk gescreend en verzekerd. Ze komen uit de eigen regio.
De dienst is actief in Haaglanden, Leiden, Rotterdam en het Groene Hart. Daarnaast zijn er lijnen voor VvE’s, horeca en bedrijven.
Wie erachter zit
65Klusser is een handelsnaam. De drager is O. den Hoed Beheer B.V. in Voorburg.
Volgens het handelsregister is die B.V. in maart 2017 opgericht als advies- en consultancyholding, met één werkzame persoon. Dezelfde B.V. voert nog twee handelsnamen: “Digital Nomad Recruiters” en “IntelliFox”.
Die eerste handelsnaam staat als pagina op de site van het recruitmentbureau 24People. Onno den Hoed is volgens een openbaar profiel sinds 2012 directeur-eigenaar van 24People en sinds 2020 medeoprichter van OnJobs. Hij sprak in september 2026 in de lokale media namens 65Klusser.
In de eerste berichten, uit december 2025, heet Gert Barnhoorn (68) de oprichter. Hij vertelde toen op Radio 5 en Den Haag FM over het ontstaan. Hoe Barnhoorn en den Hoed zich tot elkaar verhouden, staat in geen enkele gevonden bron.
De dienst is jong. Begin december 2025 draaide hij “een paar weken”, met zes klussers en ongeveer twintig klanten. Dat waren vooral alleenstaande vrouwen. Flyers in de bus werkten; reclame op sociale media niet. Nu noemt de site ongeveer twintig plaatsen en zeven klusserprofielen.
Het is geen franchise. Uitbreiding loopt via regiopartners “zonder franchisefee”. Die werken ongeveer zestien uur per week voor indicatief €1.700 tot €4.000 per maand.
De klusser krijgt volgens de wervingspagina €35 bruto per uur, “all-in”.
Er blijft 61,98 − 35 = €26,98 per uur voor de B.V. Dat is 43,5% van de omzet exclusief btw.
Dat laatste getal is een bovengrens. De werkopname is gratis voor de klant. Er zijn volgens de site geen voorrijkosten. Werkgeverslasten, verzekering en werving gaan er nog af.
Twee punten zijn onduidelijk.
Voorrijkosten. De site zegt “geen voorrijkosten”. Sleutelstad schrijft “exclusief voorrijkosten”, zonder bedrag. Parkeerkosten komen in elk geval apart.
Contractvorm. Of de klusser zzp’er is, oproepkracht of via een payrollbedrijf werkt, staat nergens.
Ook de regiopartner is na te rekenen. Zestien uur per week is 16 × 52 / 12 = 69,3 uur per maand. Dat geeft €24,52 tot €57,69 per uur. Voor €4.000 per maand zijn bij de volle marge van €26,98 ongeveer 148 klusseruren per maand nodig.
Wat de dienst goed doet
Het model lost drie dingen tegelijk op.
Werk voor ouderen met vakmanschap. Veel gepensioneerde timmerlieden, schilders en monteurs willen nog iets doen. Een paar uur per week is genoeg. €35 bruto per uur is een redelijk bedrag voor een bijbaan na de AOW-leeftijd.
Vertrouwen voor wie alleen woont. De eerste klanten waren vooral alleenstaande vrouwen. Voor hen telt het om te weten wie er aanbelt. VOG, screening en een foto vooraf geven dat.
De kleine klus wordt gedaan. Een gewone aannemer komt niet voor een gordijnrail. Zo’n klus blijft vaak jaren liggen.
Dat is de kern van het succes. De prijs is de prijs van die zekerheid.
De alternatieven
Voor dezelfde behoefte bestaat veel. De verschillen zitten vooral in wie het werk doet en voor wie het bedoeld is. Hieronder een klus van twee uur, nagerekend.
Alternatief
Regio
Twee uur
Wie doet het
Voor wie
HipHelpt
Leiden, Oegstgeest
€0
vrijwilliger
geen netwerk én weinig geld; tot ca. 2 uur
Welzijnskwartier, klussendienst
Katwijk
€10
vrijwilliger
senioren
WOEJ
Leidschendam-Voorburg
€12,50
vrijwilliger
inwoners
Radius, klussen in huis
Leiden, Oegstgeest
€60 + materiaal
welzijnsorganisatie
ouderen, mensen met een beperking
Hups, met ANBO-PCOB-korting
landelijk
€96,95
klushulp via platform
leden
Klusup
landelijk
vanaf €110
zelfstandige, ook gepensioneerden
iedereen
De Klusbakfiets
o.a. Den Haag
€117,37
vakman
iedereen
Hups, standaard
landelijk
€121,95
klushulp via platform
iedereen
Zoofy
landelijk
vanaf €130
vakman via platform
iedereen
65Klusser
Haaglanden, Leiden, Rotterdam, Groene Hart
€150
65+-vakman
iedereen
GAMMA Klusservice
landelijk
€154,95
vakman via Zoofy
iedereen
Toelichting bij de rekensom:
Hups: gemiddeld €41 per uur, minimaal twee uur, plus een eenmalige matchingsfee van €39,95. Voor ANBO-PCOB-leden is die fee €14,95.
De Klusbakfiets: €48,50 exclusief btw, dus €58,69 inclusief btw.
GAMMA: €65 per uur plus €24,95 servicekosten.
Er is meer, maar zonder vergelijkbaar uurtarief:
Altijd Voorhanden werkt met klussers van 45+ in Leiden, Leiderdorp en Den Haag. Voorrijkosten in de regio zijn inbegrepen; het uurtarief staat niet op de site.
Stichting Present (Leiden, Den Haag, Gouda, Woerden) zet groepen vrijwilligers in, via een hulpverlener.
Burenhulp van Servicepunt XL in Den Haag heeft meer dan 250 vrijwilligers.
Gilde Zoetermeer rekent €5 per uur voor minima.
Het Bruisnest Gouda begint bij €5 per uur.
Incluzio Rotterdam is gratis, plus materiaal.
Werkspot is gratis voor de klant; de vakman betaalt €3 tot €75 per lead. Het markttarief daar is €35 tot €50 per uur.
Repair Café Leiden repareert wat mensen meebrengen, maar komt niet aan huis.
Wat ontbreekt, is ook informatief:
Een tijdbank in Leiden is niet gevonden. De Time Bank Den Haag toont als laatste datum 2018.
Mens voor Elkaar in Voorschoten is per 1 augustus 2026 gestopt.
Een coöperatie die klussen tussen buren regelt, bestaat in de regio niet. De voorbeelden liggen in Zeist, Utrecht en Amsterdam.
Het patroon is helder:
Wie weinig geld en geen netwerk heeft, kan bij een welzijnsorganisatie terecht.
Wie geld heeft, kan kiezen uit platforms en bureaus.
Daartussen zit een grote groep. Die heeft wat geld en een klein netwerk. Voor die groep is 65Klusser gemaakt.
Wat de keten kost
Het visnet helpt om te zien wat hier gebeurt. Mensen zijn knopen. Een klus is een maas die tussen twee knopen sluit: iemand helpt iemand.
Bij 65Klusser loopt die maas niet rechtstreeks. Ze loopt via een keten van rollen:
klant;
bureau;
werkopnemer;
klusser.
Er zijn vijf stappen, met drie tot vier geldoverdrachten:
van klant naar B.V.;
van B.V. naar klusser;
btw naar de fiscus;
eventueel loonheffing.
Elke schakel voegt iets toe: werving, inschatting, zekerheid. Elke schakel kost ook iets. Van de €61,98 die exclusief btw binnenkomt, bereikt €35 de man of vrouw die het werk doet.
De klant en de klusser wonen vaak een paar straten van elkaar. Ze leren elkaar kennen. Bij de volgende klus belt de klant toch weer het bureau. De maas sluit, maar niet tussen hen. Ze sluit via de keten.
Dat is geen verwijt aan de dienst. Het is hoe vrijwel elke dienst in de euro-economie werkt. Een overdracht kost geld. Wereldwijd kost een geldoverdracht naar familie in het buitenland nog altijd ongeveer zes procent van het bedrag.
De vergelijking: de MAZE-Bank
De MAZE-Bank is een vorm waarin buren hun onderlinge hulp boeken zonder geld. Hij is beschreven op constable.blog en draait op maze.swarp.nl/bank. Volgens de bronnen werkt hij zo:
Boeking. Iemand noteert een daad in één zin: wie deed wat voor wie, en wanneer.
Tegentekening. De daad telt pas als de ontvanger tegentekent met een tik op “klopt”. Daarvoor is geen account nodig.
Stand. Per relatie is de stand −1, 0 of +1. Die volgt uit de bevestigde daden en wordt nergens als saldo bewaard.
Verval. Een open stand vervalt naar nul en wordt een gift aan de kring. Er is geen rente. De vervalsnelheid stelt de kring zelf vast.
Geld. Binnen de kring gaat geen geld rond. Euro’s bestaan alleen aan de rand, voor wat niet in daden kan, zoals materiaal. Dat geld staat bij een vergunninghouder, nooit bij een lid.
Bestuur. De kring bestuurt zichzelf met consent. Een waarnemer leest de getallen en heeft geen macht.
Instap. Wie wil beginnen, plakt een bestaand rooster of zorgschema in. Elke regel wordt met een klik bevestigd.
Een 65Klusser-klus ziet er in zo’n kring zo uit:
Iemand vraagt in de kring om hulp bij een gordijnrail.
Een handige buur zegt toe.
De buur hangt de rail op en noteert de daad.
De ander tikt “klopt”.
De rail zelf wordt in euro’s gekocht, als grenspost.
Binnen de kring kost dat niets. Er is geen tussenpartij die een deel houdt. Verder is de relatie wederkerig. De klusser van vandaag kan over tien jaar zelf de hulp nodig hebben. Zijn bevestigde daden staan dan in de kroniek van de kring.
Wat de ene heeft en de andere niet
Het is een vergelijking, geen wedstrijd. Beide vormen dragen iets wat de andere mist.
65Klusser heeft:
Werving. De dienst vindt klussers en brengt ze bij klanten.
Een vakinschatting vooraf. De werkopnemer ziet wat een klus vraagt.
VOG en verzekering. Dat geeft zekerheid voor wie een vreemde binnenlaat.
Een werkende praktijk. Er komen echte klussers bij echte mensen.
De MAZE-Bank heeft:
Geen tussenpartij. De waarde blijft tussen de twee mensen die de maas sluiten.
Een bewijs per daad. De ontvanger bevestigt zelf wat er gedaan is.
Wederkerigheid. Wie helpt, wordt later geholpen.
Bestuur door de betrokkenen, met consent.
Vertrouwen via mensen. Bestaande leden tekenen voor nieuwe leden, en borgen tekenen mee.
Daar staat de stand van zaken tegenover. Volgens de handleiding zijn de kern, de koppeling, het bestuursgereedschap en de reconciliatie gebouwd. De live pagina meldt nog “Nog geen daden” en “Nog geen coöperatieve jas”. Het vooroverleg met de Belastingdienst is open. Wat een VOG en een verzekering in de kring vervangt, staat in geen bron.
Het concept rond de eenheid is ook verschoven. Op 8 september 2026 rekende een post nog met Seeds: tien per lid per week, 2 procent verval, een evenwicht van 10 / 0,02 = 500. Een week later spreekt de handleiding van “geen geld, geen saldo — daden”. Een veld seeds=1 bestaat nog als gewicht, niet als munt.
Waar ze elkaar raken
Het interessantste punt zit niet in het verschil. Het zit in de plek waar ze elkaar raken.
Een vaste klusser die steeds terugkomt bij dezelfde klant, is een afspraak die zich herhaalt. Juist zo’n afspraak gebruikt de MAZE-Bank als ingang: een rooster, een zorgschema, een vaste hulp.
65Klusser bouwt die relaties nu op, in twintig plaatsen. Er ontstaan knopen die elkaar kennen. Of de maas tussen hen ooit zonder bureau gaat sluiten, hangt af van wat zo’n kring niet heeft: werving, vakinschatting en de zekerheid die een VOG en een verzekering nu geven.
Wat er verdwijnt als dat lukt, is de marge. Wat er blijft, is de gepensioneerde vakman die een rail ophangt bij een buurvrouw drie straten verder. En de kans dat zij, of iemand anders in de kring, later iets terugdoet wat hij nodig heeft.
Wat open blijft
De inhoud van de algemene voorwaarden van 65Klusser: aansprakelijkheid, betaaltermijn, annulering.
De contractvorm van de klussers.
De verhouding tussen Gert Barnhoorn en Onno den Hoed.
Of er voorrijkosten zijn, en hoeveel.
De uurtarieven van Altijd Voorhanden, Onze Klusvrouw en Hups in Leiden.
In de MAZE-Bank: wat een VOG en een verzekering vervangt, of een offerte een ingang kan zijn, en hoe de Belastingdienst een daad behandelt.
Referenties
Over 65Klusser
65Klusser — website.https://www.65klusser.nlWaarom lezen? De werkwijze in vier stappen, het tarief en de regio’s, uit de eerste hand. Leesadvies: begin bij de veelgestelde vragen.
65Klusser — Word klusser.https://www.65klusser.nl/word-klusserWaarom lezen? Hier staat de vergoeding van €35 bruto all-in per uur. Dat is de enige bron voor de verdeling van de €75.
65Klusser — Regiopartner.https://www.65klusser.nl/regiopartnerWaarom lezen? Het uitbreidingsmodel: zestien uur per week, indicatief €1.700–€4.000 per maand, zonder franchisefee.
65Klusser — In de media.https://www.65klusser.nl/in-de-mediaWaarom lezen? Het overzicht van radio, krant en lokale media, met de namen van wie er namens de dienst spreekt.
Seniorenjournaal, “65Klusser met businessmodel” (6-12-2025, op basis van het AD). https://seniorenjournaal.nl/diensten/65klusser-met-businessmodel/Waarom lezen? Het ontstaan, de eerste zes klussers, de klantengroep en waarom flyers beter werkten dan sociale media. Leesadvies: de beste korte bron over de begintijd.
Stichting Present.https://stichtingpresent.nl/leiden/Waarom lezen? Een landelijk netwerk dat groepen vrijwilligers aan hulpvragen koppelt, via hulpverleners.
Leef3 — Hups klushulp.https://leef3.nu/dienst/hups-klushulp/Waarom lezen? De volledige prijsopbouw van Hups, inclusief matchingsfee en kilometervergoeding. Leesadvies: nuttig vóór elke vergelijking van platforms op hun “vanaf”-prijs.
Klusup — Klusjesman Leiden.https://www.klusup.nl/klusjesman/leidenWaarom lezen? Een platform waar gepensioneerden zich zelf als klusser kunnen aanmelden, zonder bureau ertussen.
“About the Cooperative” (3-4-2009). https://constable.blog/2009/04/03/about-cooperative/Waarom lezen? Rochdale en Raiffeisen, en waarom coöperaties op menselijke schaal moeten blijven om te slagen. Leesadvies: de historische achtergrond bij elke kring.
The Vacuum.Net theory rests on five axioms. There is one strand (N1). Only what closes persists (N2). A closed configuration carries memory (N3). The response of the medium is a state function (N4). Each step deeper multiplies the states by three (N5).
These axioms work. They give a ternary register, a closure rule, a depth ladder and numbers that land on hydrogen, room heat and the Earth’s cavity.
Yet they carry a hidden second actor. A strand that crosses, winds and closes is something that travels over places. Then there are two things: the places and the carrier that runs along them. The rewrite rule R has the same problem. It is applied to a ring, but by what?
This essay removes the second actor. It starts from one statement: the only thing there is, is light. It then follows that statement through four steps. Light is a location. A location is a point with a bound movement. That movement combines with itself into all three trit values. And the oldest Chinese form of the character 道, together with the oldest Daoist account of origin, already contains the result.
The outcome is a theory with three rules instead of five axioms. It is simpler. It is also older.
Every step carries a status. Proved means a mathematical result that holds under its hypotheses. Computed means a number checked by arithmetic. Source means a published text or inscription. Reading means an interpretation offered here. Open means not yet settled.
2. The units are already one photon of light
The Foundational Paper (Edition 8.1) fixes two conversion units from hydrogen.
The length unit is S₀ = 1/(2R_H). With the reduced-mass Rydberg constant R_H = 10,967,758.340 m⁻¹, this gives S₀ = 4.55882 × 10⁻⁸ m, or 45.588 nm.
The time unit is T₀ = S₀/c. With c = 299,792,458 m/s, this gives T₀ = 1.52066 × 10⁻¹⁶ s.
The paper names E₀ the Hartree energy of hydrogen. It is equally the energy of one photon of wavelength S₀. Length, time and energy of the net are together one quantum of light. Computed.
Two consequences follow at once.
First, the power ceiling of the net, one closure per mesh per layer time, is E₀/T₀ = 0.0287 W. That is one photon per mesh per period. Computed.
Second, the depth ladder 3ⁿ·S₀ becomes a ladder of one photon read at other depths. At depth 2 the length is 9·S₀ = 410.29 nm, the Balmer Hδ line. At depth 5 it is 243·S₀ = 11.08 µm, close to the peak of room-temperature radiation. At depth 31 it is 3³¹·S₀ = 2.816 × 10⁷ m. The lowest ideal mode of the Earth–ionosphere cavity has wavelength 2πR/√2 = 2.831 × 10⁷ m, for R = 6,371 km. The two differ by 0.5 %. Computed.
All three landings are light. The ladder does not pass through different kinds of phenomena. It passes through one spectrum.
3. Light is a location, not a motion
The first impulse is to call light the unit motion of the net. One mesh per layer time gives S₀/T₀ = c by construction.
That definition fails. If light is defined as motion at c, it cannot also be the stuff of matter at rest. And “the motion of the net” presupposes a net that moves. Then light is again something other than the net.
The correction is this: light is a location.
Dewey Larson stated the same idea in 1959. In his natural reference system the photon has no motion of its own. It stays at its absolute location. The reference system itself progresses at unit speed. Larson’s dimensions are already part of the Foundational Paper. Source; assumed as postulate.
Relativity offers an echo but not a ground. Along a light path the interval is zero: ds² = 0. This follows from Einstein’s second postulate, so it restates an assumption. A photon also has no rest frame, so “what light experiences” lies outside that theory. Moreover relativity takes c as a universal constant, while N4 treats c as a state function. Relativity is therefore at most a correspondence here. Correspondence.
In the net, Theorem T7 gives every state exactly one address, a balanced-ternary string. Proved. If light is a location, light is an address. The 45.6 nm photon is the address at depth 0. Hδ is the same address at depth 2.
This reverses the earlier picture.
Light stays at its location and is carried along by the progression.
Matter is a location that holds itself against that progression. That is closure: what returns in phase stays put while everything else moves on. Larson calls it rotation opposing progression.
c is the measure of the difference between the two: one mesh per layer time. That is why matter always measures light at c.
The old question “can light rest?” disappears. Light always rests, in its own terms. What we call motion is matter holding on while the locations progress. Reading.
4. The point and the stroke
The character 道 makes the next step visible. It has 12 strokes in the standard count. It consists of 首, “head”, with 9 strokes, and 辶, “going”, with 3 strokes. The Kangxi tradition counts 辶 as 4 strokes, giving 13. Source.
The stroke order of 道 is fixed. A character drawn in another order counts as wrong. The order belongs to the character. Source.
The order runs as follows. Strokes 1 to 9 write 首: 丶 丿 一, then 自. Strokes 10 to 12 write 辶: a dot, a folded stroke and a long flat sweep. The component 辶 stands on the left and encloses the whole from below. Yet it is written last. First the place, then the going that carries the place.
Both parts begin with a dot. Stroke 1 is 丶. Stroke 10 is 丶. Source.
A dot in calligraphy is not a static speck. It is a short stroke: set down, a small direction, lift. The dot is movement of the smallest extent.
So there is one ground act, the stroke, in two states.
Point: movement that closes at once. This is location. This is light.
Stroke: movement that continues.
Every movement in 道 starts from a point. Never the reverse. A form is not a configuration but a fixed sequence of acts. Its identity lies in the order of making. This is N3, memory as shape, read in writing. It is also T7: the same trits in another order give another address. Reading.
No second actor is needed to draw. The order is the thing itself, as the stroke order is the character.
5. Random, but bound to the point
The movement that leaves a point need not be directed. It may be random. The question is whether it stays bound to its point.
Model the movement as a random string of trits. Each step is +1, 0 or −1 with equal chance. “Bound to the point” means the running sum comes back to 0.
Return after three steps. Count the strings of length 3.
Exact return, sum = 0: the string 000 and the six permutations of (1, 0, −1). That is 7 of 27.
Return modulo 3, the window rule of T9: 9 of 27. The two extra strings are 111 and TTT, with sum +3 and −3.
Computed. The numbers 1, 1, 3, 7, 19, 51, 141, … are the central trinomial coefficients. They count exact returns after 0, 1, 2, 3, … steps.
The difference between 7 and 9 matters. Exact return means coming back along the line. The strings 111 and TTT do not come back on the line. They go once around a ring of three. The register counts both as closure. There are two kinds of binding to the point: coming back and going round.
One dimension is always bound. The chance of exact return after n steps behaves as √3 / (2√(πn)). Each step has variance 2/3. The local limit gives 1/√(2πn·2/3) = √3/(2√(πn)).
That chance goes to zero. But the sum over all n diverges. A random movement along one line therefore returns to its point with certainty, and does so infinitely often. Proved: Pólya, 1921. It holds for every step law with mean zero and finite variance.
In three dimensions this fails. There the return probability is P₃ = 0.3405 (Watson, 1939; Theorem T18 of the Foundational Paper). Source.
The consequence is direct. “Random, but bound to the point” needs no force and no second actor. A movement with only one direction, forward and back, stays tied to its point by itself. That is N1, one strand, read without a strand.
Light is the point.
The movement is random but one-dimensional, and therefore bound.
Matter arises where the movement does not merely return, but returns in phase.
Reading, on top of the proved recurrence.
6. The act combined with itself
Write the ground act as “. >”: leaving a place. Its mirror is “< .”: arriving at a place.
The mirror need not be given separately. Give “. >” the value +1 and repeat it, counting modulo 3.
once: +1;
twice: 1 + 1 = 2 ≡ −1, the mirror “< .”;
three times: 1 + 1 + 1 = 3 ≡ 0.
Proved: the group ℤ₃ is cyclic and generated by one element. From one act all three trit values arise. This works only on the ring of three. On the plain line, 1 + 1 stays 2.
There are two roads to 0.
Meeting: “. > < .”, the act meeting its mirror: +1 + (−1) = 0.
Going round: the same act three times: 111 or TTT.
These are exactly the two kinds of closure in Theorem T10. A window closes when its three trits are all different (heterogeneous) or all equal (homogeneous). Proved.
So the 0 is neither emptiness nor rest. It is leaving and arriving cancelling in the sum while both exist. In terms of light it is a movement between two places, there and back: a standing wave with the points as nodes. That is the simplest bound light. The mesh is the space between a leaving and an arriving. Reading.
Order still counts. “> <” (towards each other) and “< >” (away from each other) both have sum 0. They differ. The first closes. The second opens. The sum does not see the difference. The order does, as in the stroke order of 道. Open: the register so far measures sums only. It needs a quantity that records order.
7. The oldest Dao: a crossing with someone in it
The modern 道 is a contraction. Older forms are written 衟 or 衜. Lexicographers read them as a person standing at a crossroads 行. The head 首 sits inside the crossroads. At the bottom stands 止, the foot. Source.
The component 行 carries the meanings walking, passage, birth canal and crossroads. The pairing of 首 with 止 can stand for a person. Source.
The oldest 道 thus holds three things in one sign.
行, the crossing: the self-crossing of N1.
首, the place in the crossing: the point, the location, light.
止, the foot: the step, the movement bound to the place.
There is no carrier besides these. The crossing, the place and the step are given together. Source for the graph; reading for the mapping.
8. The oldest account of origin: arising by return
In 1993 a tomb at Guodian in Hubei yielded bamboo slips from around 300 BCE. They hold the oldest known Laozi material. Beside it lies a short text of 14 slips, the Taiyi sheng shui, “The Great One Gave Birth to Water”. It has not been found anywhere else. Source.
The Great One gave birth to water. Water returned and assisted the Great One; thus heaven was completed. Heaven returned and assisted the Great One; thus earth was completed. Heaven and earth assisted each other again; thus the spirits above and below were completed. These assisted each other again; thus yin and yang were completed. Yin and yang assisted each other again; thus the four seasons were completed. The four seasons assisted each other again; thus cold and heat were completed. Cold and heat assisted each other again; thus moist and dry were completed. Moist and dry assisted each other again; the year was completed, and it stopped there.
The text then reads the chain backwards. The year is born of moist and dry, moist and dry of cold and heat, and so on back to the Great One. Source.
Then come three sentences that hold the whole mechanism:
是故太一藏於水，行於時，周而又始，以己為萬物母；一缺一盈，以己為萬物經。
Therefore the Great One hides in water and moves through the seasons. It goes round and begins again, and makes itself the mother of the ten thousand things. Once lacking, once full, it makes itself the warp of the ten thousand things.
Source. The word 經 means the warp: the fixed threads of a loom through which everything else is woven.
Chapter 42 of the received Laozi, “One gives birth to two, two to three, three to the ten thousand things”, does not occur in the Guodian Laozi. The older cosmogony does not count. It lets things arise by return and by mutual support. Source.
9. Three rules instead of five axioms
The material of sections 2 to 8 yields three rules.
R1 — Crossing, place and step (道). There is one movement that leaves a place through a crossing. The place is light. The movement is bound to it.
R1 replaces N1. It needs no strand as a separate carrier. The crossing, the place and the step are given together, as in 衟.
R2 — Return gives birth (反輔). What arises returns to its source, and from that return the next thing arises.
R2 replaces N2. In N2 closure only selects. In R2 closure also produces. Water returns to the One, and heaven is completed. In register terms this is “. >” coming back to its point.
R3 — Mutual support (相輔). Two things that have arisen support each other, and from that support the next layer arises.
R3 is the meeting “. > < .”, with sum 0, but productive. Repeated, it gives the ladder of N5: each layer arises from the pair of the layer before.
What N3 and N4 did is already present.
Memory (N3). The text reads the chain backwards, each thing with its source. The order of arising is the form, as the stroke order is the character.
Alternation. “Once lacking, once full” is +1/−1. “Goes round and begins again” is the cycle of R. The year is the first complete closure.
Context (N4). The rest of the text says that the way of heaven values the weak, cuts what is complete to increase what is growing. The response depends on the state, not on a fixed law. Correspondence.
The image is the same as before. The One is the warp of all things and hides in water. That is the net spanned in water, in the oldest source. Correspondence.
Why this is older and simpler: R1–R3 need no strand, no external rewrite and no selector. Return presupposes a place to return to. That place is light. Arising presupposes a movement bound to that place. That movement is the step of 止. Reading.
10. What changes in the theory
Several statements of the Foundational Paper shift.
N1 speaks of an oriented, continuous strand. It becomes a sequence of locations with a fixed order of arising.
N2 says what does not close does not last. Free light lasts billions of years. Under R2, closure means return in phase, and a cycle of R also returns. Light lasts because it is periodic. Matter lasts because it holds.
N4 makes c a state function. If light is the unit, c is 1 in native units. That c varies means the step from address to address per rewrite varies locally.
C6 reads 0 as rest. That reading lapses. Rest is now the property of light, not a trit value. The 0 is the meeting of an act with its mirror.
R is no longer applied from outside. It is the form reading itself. Its fixed points (T14) are locations that stay equal under self-reference. Kauffman’s eigenform O(A) = A states the same.
Reading for all five.
11. Open questions
Order. The register records sums. “> <” and “< >” have the same sum but differ. A quantity that records order is needed.
Dimension. A random movement along one line is bound with certainty. In three dimensions the return chance is 0.3405. If space consists of one-dimensional bound movements, where do the three dimensions come from? This is probably where binding and closure part ways.
The nine products. From water to the year, the Taiyi sheng shui names nine products: water, heaven, earth, the spirits, yin-yang, the four seasons, cold-heat, moist-dry and the year. The register has nine closing windows. Whether these belong together must be computed first. It must not be read in.
The third value. Varela (1975) extended Spencer-Brown’s calculus with a third value beside marked and unmarked: the autonomous value of re-entry. That gives three values from distinction alone. How they map onto +1, 0 and −1 is open.
12. Summary
The only thing there is, is light. The units of the net are already one photon: 45.6 nm, 1.52 × 10⁻¹⁶ s, 27.2 eV.
Light is a location. Matter is a location that holds against the progression. c measures the difference.
A location is a point with a bound movement. Along one line, a random movement returns to its point with certainty.
The act combined with itself gives all three trit values. Meeting and going round are the two kinds of closure.
The oldest 道 shows a person at a crossing with a foot. The oldest Daoist cosmogony lets everything arise by return and by mutual support, and calls the One the warp hidden in water.
Three rules follow. One movement leaves a place through a crossing. Return gives birth. Mutual support builds the ladder. The net is not threads and meshes. It is locations referring to locations.
Annotated references
Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper. Edition 8, version 8.1. Constable Research, Leiden.Why read? The five axioms, the window register (T9–T14), the unit derivation (VII.2) and the free-path lever (T18) that this essay rewrites. Reading advice: Part 0, then Parts III.4 and VII.2.
Konstapel, J. (2026). “The Possibilities and Impossibilities of the Net.” constable.blog, 19 October 2026.Why read? The fifty-cell map of what the net can do, with the power ceiling and the depth-ladder landings used in section 2.
Larson, D. B. (1959). The Structure of the Physical Universe. North Pacific Publishers, Portland.Why read? Space and time as two aspects of one motion. The photon at rest in the natural reference system; matter as rotation against the progression. Reading advice: the opening chapters on the natural reference system.
Spencer-Brown, G. (1969). Laws of Form. Allen & Unwin, London.Why read? The mark of distinction and re-entry. A location as a distinction needs no carrier. Reading advice: chapters 1–2 and 11.
Kauffman, L. H. (2005). “EigenForm.” Kybernetes 34(1/2), 129–150.Why read? The eigenform O(A) = A as a form that reproduces itself. The basis for reading R as self-reading rather than an external rewrite.
Varela, F. J. (1975). “A calculus for self-reference.” International Journal of General Systems 2, 5–24.Why read? Adds a third, autonomous value to Spencer-Brown’s two. A ternary ground from distinction alone. Varela was also a practitioner of meditation, not only a formalist.
Pólya, G. (1921). “Über eine Aufgabe der Wahrscheinlichkeitsrechnung betreffend die Irrfahrt im Straßennetz.” Mathematische Annalen 84, 149–160.Why read? The proof that a random walk returns with certainty in one and two dimensions, and not in three. The ground of section 5.
Watson, G. N. (1939). “Three triple integrals.” Quarterly Journal of Mathematics 10, 266–276.Why read? The exact return probability P₃ = 0.3405 of the three-dimensional walk.
OEIS A002426. Central trinomial coefficients.Why read? The sequence 1, 1, 3, 7, 19, 51, 141, … that counts exact returns of the trit walk, with its asymptotics.
Knuth, D. E. (1997). The Art of Computer Programming, Vol. 2, 3rd ed. Addison-Wesley.Why read? Balanced ternary and its symmetries, including truncation as rounding. Reading advice: pp. 195–213.
Jingmen Municipal Museum (1998). Guodian Chu mu zhujian 郭店楚墓竹简. Wenwu, Beijing.Why read? The photographic edition and transcription of the Guodian slips, including the Laozi material and the Taiyi sheng shui.
Allan, S. & Williams, C. (eds.) (2000). The Guodian Laozi: Proceedings of the International Conference, Dartmouth College, May 1998. Society for the Study of Early China, Berkeley.Why read? Scholarly discussion of the Guodian Laozi and the first wide treatment of the Taiyi sheng shui in English.
Henricks, R. G. (2000). Lao Tzu’s Tao Te Ching: A Translation of the Startling New Documents Found at Guodian. Columbia University Press, New York.Why read? Chapter-by-chapter translation of the Guodian Laozi, with the ancient and modern graphs. Shows which received chapters are absent, including chapter 42.
Wenlin Institute. Dictionary entry 道 (dào).Why read? The older graphs 衟/衜, read as a person at a crossroads 行 with the foot 止 below. The paleographic basis of section 7.
Williamson, J. G. & van der Mark, M. B. (1997). “Is the electron a photon with toroidal topology?” Annales de la Fondation Louis de Broglie 22(2), 133–160.Why read? The electron as light closed on itself in a double loop. Matter as held light, in a physical model.
Laozi, Daodejing, chapters 11, 25 and 42.Why read? The useful empty centre of the hub (11), the way that goes round (25) and the counting cosmogony (42), to set beside the older Taiyi sheng shui.
The Way Before the Net: Light, Location, and the Three Rules of Arising
1. Why Go Back
The Vacuum.Net theory rests on five axioms. There is one strand (N1). Only what closes persists (N2). A closed configuration carries memory (N3). The response of the medium is a state function (N4). Each step deeper multiplies the states by three (N5).
These axioms work. They give a ternary register, a closure rule, a depth ladder, and numbers that land on hydrogen, room heat, and the Earth’s cavity.
Yet they carry a hidden second actor. A strand that crosses, winds, and closes is something that travels over places. Then there are two things: the places and the carrier that runs along them. The rewrite rule R has the same problem. It is applied to a ring, but by what?
This essay removes the second actor. It starts from one statement: the only thing there is, is light. It then follows that statement through four steps. Light is a location. A location is a point with a bound movement. That movement combines with itself into all three trit values. And the oldest Chinese form of the character 道, together with the oldest Daoist account of origin, already contains the result.
The outcome is a theory with three rules instead of five axioms. It is simpler. It is also older.
Every step carries a status. Proved means a mathematical result that holds under its hypotheses. Computed means a number checked by arithmetic. Source means a published text or inscription. Reading means an interpretation offered here. Open means not yet settled.
2. The Units Are Already One Photon of Light
The Foundational Paper (Edition 8.1) fixes two conversion units from hydrogen.
The length unit is S₀ = 1/(2R·H). With the reduced-mass Rydberg constant R·H = 10,967,758.340 m⁻¹, this gives S₀ = 4.55882 × 10⁻⁸ m, or 45.588 nm.
The time unit is T₀ = S₀/c. With c = 299,792,458 m/s, this gives T₀ = 1.52066 × 10⁻¹⁶ s.
The paper names E₀ the Hartree energy of hydrogen. It is equally the energy of one photon of wavelength S₀. Length, time, and energy of the net are together one quantum of light. Computed.
Two consequences follow at once.
First, the power ceiling of the net—one closure per mesh per layer time—is E₀/T₀ = 0.0287 W. That is one photon per mesh per period. Computed.
Second, the depth ladder 3ⁿ·S₀ becomes a ladder of one photon read at other depths. At depth 2 the length is 9·S₀ = 410.29 nm, the Balmer Hδ line. At depth 5 it is 243·S₀ = 11.08 μm, close to the peak of room-temperature radiation. At depth 31 it is 3³¹·S₀ = 2.816 × 10⁷ m. The lowest ideal mode of the Earth-ionosphere cavity has wavelength 2πR/√2 = 2.831 × 10⁷ m, for R = 6,371 km. The two differ by 0.5%. Computed.
All three landings are light. The ladder does not pass through different kinds of phenomena. It passes through one spectrum.
3. Light Is a Location, Not a Motion
The first impulse is to call light the unit motion of the net. One mesh per layer time gives S₀/T₀ = c by construction.
That definition fails. If light is defined as motion at c, it cannot also be the stuff of matter at rest. And “the motion of the net” presupposes a net that moves. Then light is again something other than the net.
The correction is this: light is a location.
Dewey Larson stated the same idea in 1959. In his natural reference system the photon has no motion of its own. It stays at its absolute location. The reference system itself progresses at unit speed. Larson’s dimensions are already part of the Foundational Paper. Source; assumed as postulate.
Relativity offers an echo but not a ground. Along a light path the interval is zero: ds² = 0. This follows from Einstein’s second postulate, so it restates an assumption. A photon also has no rest frame, so “what light experiences” lies outside that theory. Moreover relativity takes c as a universal constant, while N4 treats c as a state function. Relativity is therefore at most a correspondence here. Correspondence.
In the net, Theorem T7 gives every state exactly one address, a balanced-ternary string. Proved. If light is a location, light is an address. The 45.6 nm photon is the address at depth 0. Hδ is the same address at depth 2.
This reverses the earlier picture.
Light stays at its location and is carried along by the progression. Matter is a location that holds itself against that progression. That is closure: what returns in phase stays put while everything else moves on. Larson calls it rotation opposing progression. c is the measure of the difference between the two: one mesh per layer time. That is why matter always measures light at c.
The old question “can light rest?” disappears. Light always rests, in its own terms. What we call motion is matter holding on while the locations progress. Reading.
4. The Point and the Stroke
The character 道 makes the next step visible. It has 12 strokes in the standard count. It consists of 首, “head,” with 9 strokes, and 辶, “going,” with 3 strokes. The Kangxi tradition counts 辶 as 4 strokes, giving 13. Source.
The stroke order of 道 is fixed. A character drawn in another order counts as wrong. The order belongs to the character. Source.
The order runs as follows. Strokes 1 to 9 write 首: a dot, then a horizontal, then 自. Strokes 10 to 12 write 辶: a dot, a folded stroke, and a long flat sweep. The component 辶 stands on the left and encloses the whole from below. Yet it is written last. First the place, then the going that carries the place.
Both parts begin with a dot. Stroke 1 is a dot. Stroke 10 is a dot. Source.
A dot in calligraphy is not a static speck. It is a short stroke: set down, a small direction, lift. The dot is movement of the smallest extent.
So there is one ground act, the stroke, in two states.
Point: movement that closes at once. This is location. This is light.
Stroke: movement that continues.
Every movement in 道 starts from a point. Never the reverse. A form is not a configuration but a fixed sequence of acts. Its identity lies in the order of making. This is N3, memory as shape, read in writing. It is also T7: the same tries in another order give another address. Reading.
No second actor is needed to draw. The order is the thing itself, as the stroke order is the character.
5. Random, but Bound to the Point
The movement that leaves a point need not be directed. It may be random. The question is whether it stays bound to its point.
Model the movement as a random string of tries. Each step is +1, 0, or −1 with equal chance. “Bound to the point” means the running sum comes back to 0.
Return after three steps. Count the strings of length 3.
Exact return, sum = 0: the string 000 and the six permutations of (1, 0, −1). That is 7 of 27.
Return modulo 3, the window rule of T9: 9 of 27. The two extra strings are 111 and TTT, with sum +3 and −3.
Computed. The numbers 1, 1, 3, 7, 19, 51, 141, … are the central trinomial coefficients. They count exact returns after 0, 1, 2, 3, … steps.
The difference between 7 and 9 matters. Exact return means coming back along the line. The strings 111 and TTT do not come back on the line. They go once around a ring of three. The register counts both as closure. There are two kinds of binding to the point: coming back and going round.
One dimension is always bound. The chance of exact return after n steps behaves as √3/(2√(πn)). Each step has variance 2/3. The local limit gives 1/√(2πn·2/3) = √3/(2√(πn)).
That chance goes to zero. But the sum over all n diverges. A random movement along one line therefore returns to its point with certainty, and does so infinitely often. Proved: Pólya, 1921. It holds for every step law with mean zero and finite variance.
In three dimensions this fails. There the return probability is P₃ = 0.3405 (Watson, 1939; Theorem T18 of the Foundational Paper). Source.
The consequence is direct. “Random, but bound to the point” needs no force and no second actor. A movement with only one direction, forward and back, stays tied to its point by itself. That is N1, one strand, read without a strand.
Light is the point. The movement is random but one-dimensional, and therefore bound. Matter arises where the movement does not merely return, but returns in phase.
Reading, on top of the proved recurrence.
6. The Act Combined with Itself
Write the ground act as “>” : leaving a place. Its mirror is “<” : arriving at a place.
The mirror need not be given separately. Give “>” the value +1 and repeat it, counting modulo 3.
Once: +1
Twice: 1 + 1 = 2 ≡ −1, the mirror “<“
Three times: 1 + 1 + 1 = 3 ≡ 0
Proved: the group ℤ₃ is cyclic and generated by one element. From one act all three trit values arise. This works only on the ring of three. On the plain line, 1 + 1 stays 2.
There are two roads to 0.
Meeting: “><“, the act meeting its mirror: +1 + (−1) = 0
Going round: the same act three times: 111 or TTT.
These are exactly the two kinds of closure in Theorem T10. A window closes when its three trits are all different (heterogeneous) or all equal (homogeneous). Proved.
So the 0 is neither emptiness nor rest. It is leaving and arriving cancelling in the sum while both exist. In terms of light it is a movement between two places, there and back: a standing wave with the points as nodes. That is the simplest bound light. The mesh is the space between a leaving and an arriving. Reading.
Order still counts. “><” (towards each other) and “<>” (away from each other) both have sum 0. They differ. The first closes. The second opens. The sum does not see the difference. The order does, as in the stroke order of 道. Open: the register so far measures sums only. It needs a quantity that records order.
7. The Oldest Dao: A Crossing with Someone in It
The modern 道 is a contraction. Older forms are written 行 or 十字 with 首 inside. Lexicographers read them as a person standing at a crossroads 行. The head 首 sits inside the crossroads. At the bottom stands 止, the foot. Source.
The component 行 carries the meanings walking, passage, birth canal, and crossroads. The pairing of 行 with 首 can stand for a person. Source.
The oldest 道 thus holds three things in one sign.
行, the crossing: the self-crossing of N1.
首, the place in the crossing: the point, the location, light.
止, the foot: the step, the movement bound to the place.
There is no carrier besides these. The crossing, the place, and the step are given together. Source for the graph; reading for the mapping.
8. The Oldest Account of Origin: Arising by Return
In 1993 a tomb at Guodian in Hubei yielded bamboo slips from around 300 BCE. They hold the oldest known Laozi material. Beside it lies a short text of 14 slips, the Taiyi sheng shui, “The Great One Gave Birth to Water.” It has not been found anywhere else. Source.
The Great One gave birth to water. Water returned and assisted the Great One; thus heaven was completed. Heaven returned and assisted the Great One; thus earth was completed. Heaven and earth assisted each other again; thus the spirits above and below were completed. These assisted each other again; thus yin and yang were completed. Yin and yang assisted each other again; thus the four seasons were completed. The four seasons assisted each other again; thus cold and heat were completed. Cold and heat assisted each other again; thus moist and dry were completed. Moist and dry assisted each other again; the year was completed, and it stopped there.
The text then reads the chain backwards. The year is born of moist and dry, moist and dry of cold and heat, and so on back to the Great One. Source.
Then come three sentences that hold the whole mechanism:
是故太一藏於水，行於時，周而又，以已為萬物母；一缺一盈，以已為萬物經。
Therefore the Great One hides in water and moves through the seasons. It goes round and begins again, and makes itself the mother of the ten thousand things. Once lacking, once full, it makes itself the warp of the ten thousand things.
Source. The word 經 means the warp: the fixed threads of a loom through which everything else is woven.
Chapter 42 of the received Laozi, “One gives birth to two, two to three, three to the ten thousand things,” does not occur in the Guodian Laozi. The older cosmogony does not count. It lets things arise by return and by mutual support. Source.
9. Three Rules Instead of Five Axioms
The material of sections 2 to 8 yields three rules.
R1 — Crossing, place, and step (道). There is one movement that leaves a place through a crossing. The place is light. The movement is bound to it.
R1 replaces N1. It needs no strand as a separate carrier. The crossing, the place, and the step are given together, as in 道.
R2 — Return gives birth (反). What arises returns to its source, and from that return the next thing arises.
R2 replaces N2. In N2 closure only selects. In R2 closure also produces. Water returns to the One, and heaven is completed. In register terms this is “>” coming back to its point.
R3 — Mutual support (相). Two things that have arisen support each other, and from that support the next layer arises.
R3 is the meeting “><” with sum 0, but productive. Repeated, it gives the ladder of N5: each layer arises from the pair of the layer before.
What N3 and N4 did is already present.
Memory (N3). The text reads the chain backwards, each thing with its source. The order of arising is the form, as the stroke order is the character. Alternation. “Once lacking, once full” is +1/−1. “Goes round and begins again” is the cycle of R. The year is the first complete closure. Context (N4). The rest of the text says that the way of heaven values the weak, cuts what is complete to increase what is growing. The response depends on the state, not on a fixed law. Correspondence.
The image is the same as before. The One is the warp of all things and hides in water. That is the net spanned in water, in the oldest source. Correspondence.
Why this is older and simpler: R1–R3 need no strand, no external rewrite, and no selector. Return presupposes a place to return to. That place is light. Arising presupposes a movement bound to that place. That movement is the step of it. Reading.
10. What Changes in the Theory
Several statements of the Foundational Paper shift.
N1 speaks of an oriented, continuous strand. It becomes a sequence of locations with a fixed order of arising. N2 says what does not close does not last. Free light lasts billions of years. Under R2, closure means return in phase, and a cycle of R also returns. Light lasts because it is periodic. Matter lasts because it holds. N4 makes c a state function. If light is the unit, c is 1 in native units. That c varies means the step from address to address per rewrite varies locally. C6 reads 0 as rest. That reading lapses. Rest is now the property of light, not a trit value. The 0 is the meeting of an act with its mirror. R is no longer applied from outside. It is the form reading itself. Its fixed points (T14) are locations that stay equal under self-reference. Kauffman’s eigenform O(A) = A states the same.
Reading for all five.
11. Open Questions
Order. The register records sums. “><” and “<>” have the same sum but differ. A quantity that records order is needed.
Dimension. A random movement along one line is bound with certainty. In three dimensions the return chance is 0.3405. If space consists of one-dimensional bound movements, where do the three dimensions come from? This is probably where binding and closure part ways.
The nine products. From water to the year, the Taiyi sheng shui names nine products: water, heaven, earth, the spirits, yin-yang, the four seasons, cold-heat, moist-dry, and the year. The register has nine closing windows. Whether these belong together must be computed first. It must not be read in. This is open.
12. Summary
The only thing there is, is light. The units of the net are already one photon: 45.6 nm, 1.52 × 10⁻¹⁶ s, 27.2 eV.
Light is a location. Matter is a location that holds against the progression. c measures the difference.
A location is a point with a bound movement. Along one line, a random movement returns to its point with certainty.
The act combined with itself gives all three trit values. Meeting and going round are the two kinds of closure.
The oldest 道 shows a person at a crossing with a foot. The oldest Daoist cosmogony lets everything arise by return and by mutual support, and calls the One the warp hidden in water.
Three rules follow. One movement leaves a place through a crossing. Return gives birth. Mutual support builds the ladder. The net is not threads and meshes. It is locations referring to locations.
Annotated References
Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper. Edition 8, version 8.1. Constable Research, Leiden. Why read? The five axioms, the window register (T9–T14), the unit derivation (VII.2), and the free-path lever (T18) that this essay rewrites. Reading advice: Part 0, then Parts III.4 and VII.2.
Konstapel, J. (2026). “The Possibilities and Impossibilities of the Net.” constable.blog, 19 October 2026. Why read? The fifty-cell map of what the net can do, with the power ceiling and the depth-ladder landings used in section 2.
Larson, D. B. (1959). The Structure of the Physical Universe. North Pacific Publishers, Portland. Why read? Space and time as two aspects of one motion. The photon at rest in the natural reference system; matter as rotation against the progression. Reading advice: the opening chapters on the natural reference system.
Spencer-Brown, G. (1969). Laws of Form. Allen & Unwin, London. Why read? The mark of distinction and re-entry. A location as a distinction needs no carrier. Reading advice: chapters 1–2 and 11.
Kauffman, L. H. (2005). “EigenForm.” Kybernetes 34(1/2), 129–150. Why read? The eigenform O(A) = A as a form that reproduces itself. The basis for reading R as self-reading rather than an external rewrite.
Varela, F. J. (1975). “A calculus for self-reference.” International Journal of General Systems 2, 5–24. Why read? Adds a third, autonomous value to Spencer-Brown’s two. A ternary ground from distinction alone. Varela was also a practitioner of meditation, not only a formalist.
Pólya, G. (1921). “Über eine Aufgabe der Wahrscheinlichkeitsrechnung betreffend die Irrfahrt im Straßennetz.” Mathematische Annalen 84, 149–160. Why read? The proof that a random walk returns with certainty in one and two dimensions, and not in three. The ground of section 5.
Watson, G. N. (1939). “Three triple integrals.” The Quarterly Journal of Mathematics 10, 266–276. Why read? The computation of 0.3405 for the three-dimensional walk.
OEIS A002426. Central trinomial coefficients. Why read? The sequence 1, 1, 3, 7, 19, 51, 141, … that counts exact returns of the trit walk, with its asymptotics.
Knuth, D. E. (1997). The Art of Computer Programming, Vol. 2, 3rd ed. Addison-Wesley. Why read? Balanced ternary and its symmetries, including truncation as rounding. Reading advice: pp. 195–213.
Jingmen Municipal Museum (1998). Guodian Chu mu zhujian 郭店楚墓竹简. Wenwu, Beijing. Why read? The photographic edition and transcription of the Guodian slips, including the Laozi material and the Taiyi sheng shui.
Allan, S. & Williams, C. (eds.) (2000). The Guodian Laozi: Proceedings of the International Conference, Dartmouth College, May 1998. Society for the Study of Early China, Berkeley. Why read? Scholarly discussion of the Guodian Laozi and the first wide treatment of the Taiyi sheng shui in English.
Henricks, R. G. (2000). Lao Tzu’s Tao Te Ching: A Translation of the Startling New Documents Found at Guodian. Columbia University Press, New York. Why read? Chapter-by-chapter translation of the Guodian Laozi, with the ancient and modern graphs. Shows which received chapters are absent, including chapter 42.
Wenlin Institute. Dictionary entry 道 (dao). Why read? The older graphs 行/十字, read as a person at a crossroads 行 with the foot 止 below. The paleographic basis of section 7.
Williamson, J. G. & van der Mark, M. B. (1997). “Is the electron a photon with toroidal topology?” Annales de la Fondation Louis de Broglie 22(2), 133–160. Why read? The electron as light closed on itself in a double loop. Matter as held light, in a physical model.
Laozi, Daodejing, chapters 11, 25 and 42. Why read? The useful empty centre of the hub (11), the way that goes round (25), and the counting cosmogony (42), to set beside the older Taiyi sheng shui.
The same pathogen, the same toxin and the same strain make one person ill and leave another untouched. Medicine explains part of this difference with genes, the microbiome and prior exposure. A remainder is left unexplained.
This essay reads that remainder through the Vacuum.Net theory. The theory pictures the world as one strand, a fishnet under tension. Only what closes in phase persists; what does not close carries tension to the meshes beside it.
In this frame a person is a closure. The personal blueprint is the address of that closure at the moment it first held. Illness is read as the place where the body’s reading layer fails. It reads either too weakly, so misfits pass, or too sharply, so healthy tissue is discharged.
The central proposition is simple. The blueprint fixes where tension lands. The state of the person fixes when. Illness needs both.
The blueprint is computed along three lines that the SWARP implementation has developed:
the address in the net;
the failure class;
the electromagnetic profile at birth.
Each line is set out with its arithmetic, and one full chart is computed from raw birth data.
The essay then sets the frame beside what has been measured. The birth season changes the risk of dozens of conditions. For multiple sclerosis the effect mirrors between hemispheres and follows maternal light exposure. Traditional medicine tied birth season to constitution long before it was measured. Sun signs show no effect after correction for multiple testing.
The essay ends with five tests. In each, the prediction is fixed before the count.
1. The question
Two people sit in the same room with the same virus. One has a cold for three days. The other has pneumonia.
Two sisters share half their genes and most of their childhood. One develops lupus at thirty. The other never does.
A quarter of the world carries tuberculosis. Between 5 and 10 per cent of those carriers ever become ill.
The standing answers are genes, microbiome, nutrition, age and prior exposure. Each explains a share. None closes the account. Twins with identical genomes diverge in autoimmune disease. Identical exposures give opposite outcomes.
Traditional medicine never treated this as noise. Ayurveda assigns each person a prakriti, a constitution set at conception and birth. Chinese medicine reads the season and hour of birth into the balance of the five phases. Both systems have been in continuous use for more than two thousand years. Both hold that illness depends on who the person is, not only on what reaches them.
The personal blueprint takes up that claim in a formal way. This essay asks what the blueprint says about illness, where it says it, and how that can be counted.
2. The frame: one strand, one test
2.1 The net
The Vacuum.Net theory starts from one strand. The standing image is a fishnet under tension.
Where the strand crosses itself, there is a crossing.
Between crossings lie the meshes.
A winding is a stretch of strand that returns on itself in step.
A knot is windings that hold each other in three directions.
One rule decides what stays. A winding persists only if it returns on itself in phase. What does not close does not stay as a form; its tension passes to the meshes beside it.
2.2 The register
At a crossing the strand passes over, passes under, or does not cross. These states are written +1, −1 and 0. One such state is a trit.
Three crossings in a row form a window. A window is closed when its trits add to a multiple of three.
Of the 3 × 3 × 3 = 27 possible windows, 9 close:
three are uniform: (+1, +1, +1), (0, 0, 0) and (−1, −1, −1);
six are mixed: the six orders of (−1, 0, +1).
All 18 windows with two equal trits and one different stay open. Their sum is never a multiple of three.
The two closed kinds behave differently:
A uniform closure holds and does not turn. It is the form of a crystal.
A mixed closure keeps turning while it stays closed. Every crossing changes at every step, and every window stays closed. It is the form of living matter.
Where a uniform stretch meets a mixed stretch, the window on the seam stands open. It carries tension.
2.3 The test
When one form reads another, there are three outcomes:
+1, it fits: the form stays;
−1, it does not fit: the form is taken apart;
0, it is not read: the form remains.
A living body carries a reading layer that runs this test at cell scale, continuously, from before birth to death. On an ordinary day it removes thousands of cells whose surface has drifted.
In some places the body must not discharge what it reads:
the gut wall at every meal;
the eye;
the child in the womb, which is half foreign by signature.
There the body uses lawful zero-tokens that turn the local reading to 0. Clinicians call them checkpoints.
This gives one dial with two directions of failure:
Reading too weakly lets misfits through. Read in the net, a tumour is the remainder of years in which every variant without the zero-token was discharged and every variant with it was not.
Reading too sharply discharges healthy tissue. That is autoimmunity.
Ageing is read as lost phase-lock. Heart-rate variability falls, deep sleep flattens, hormone tides drift out of step. The misfit rate barely changes; the discharge rate falls.
These are readings, not mechanisms. In a one-strand model there are no two separate things between which to draw a causal arrow. The frame uses four relations instead of cause:
selection by fit;
identity over scale;
constitutive relation;
correspondence.
3. The blueprint as an address
3.1 Closure at birth
A newborn becomes an autonomous closure in minutes. At the first breath:
the lungs open;
pulmonary resistance falls;
the foramen ovale and the ductus arteriosus close.
Circulation that ran through the mother now returns on itself.
The blueprint is the address of the person in the net at that moment. It has two parts:
The invariant is discrete. It changes only by a jump, never by drift. It says where in the person’s ring the windows stand open.
The state is continuous: a point q on the three-sphere S³, together with the person’s tension χ̄ and retention time τ. It moves daily. It says when tension builds.
From this split follows the proposition of this essay:
The blueprint fixes where tension lands. The state fixes when. Illness needs both.
A blueprint alone predicts no illness. A state alone does not say where an illness will show.
The four axes follow the four realities of Paths of Change: Blue (unitary), Red (sensory), Green (social) and Yellow (mythic). The dominant axis sets the failure class.
The state is updated after each observed outcome r(t):
q(t+1) = norm( q(t) + η · ( r(t) − q(t) ) ), with η = 0.15.
The invariant is excluded from this update.
3.3 Line one: the address in the net
The first line reads the birth moment as a position of the largest nearby sub-knots: the Sun, the Moon and the planets on the ecliptic ring.
The standard computation is the Human Design chart. It uses:
the positions at birth;
the positions at the moment when the Sun stood 88° of arc earlier.
The ecliptic is divided into 64 gates of 5.625° each, with gate 41 starting at 302°. Each gate has six lines of 0.9375°. Pairs of activated gates form channels. Channels define centres.
The chart is converted to the state vector with fixed weights:
Chart field
Weight
Type
0.30
Profile
0.20
Defined centres
0.25
Active channels
0.15
Incarnation cross
0.10
Channel circuits map to axes:
Individual → Yellow;
Collective → Blue;
Tribal → Green;
Integration → Red.
3.4 Line two: the failure class
The second line reads where a person’s learning cycle breaks. Each dominant axis has its own break:
Blue. Keeps a script too long. A deviation is not acknowledged.
Red. Does not register that a script has broken. The signal arrives late, often as a bodily crisis.
Green. Cannot retrieve relevant experience, mostly in relations.
Yellow. Does not revise the underlying model. Vision is followed by structural collapse.
In the SWARP implementation the failure class sets the blockage. The blockage is the state the person reaches with the greatest difficulty, among four tension states:
loading;
holding;
discharge;
rest.
3.5 Line three: the electromagnetic profile
The third line reads the physical field at birth, computed from the same date, time and place:
The solar-cycle phase: sunspot number and position in the 11-year cycle. The record goes back to 1749.
The geomagnetic load: the Ap or aa index on the birth day, and integrated over gestation. The aa record goes back to 1868, Ap to 1932.
The local field at the birthplace from the International Geomagnetic Reference Field: strength, declination and inclination.
This line ties the address to quantities that instruments record without interpretation.
3.6 A worked example
The example from The Narrative Signature Engine is a birth on 22 April 1951, 01:02, Leiden.
Time. The Netherlands kept Central European Time without summer time in 1951. The birth moment is therefore 00:02 UT. The design moment, with the Sun 88° earlier, falls on 23 January 1951, 21:13 UT, which is 88.12 days before birth.
Ephemeris. Positions are computed with the Swiss Ephemeris and the true lunar node.
Body
Birth
Design
Sun
3.6
41.2
Earth
50.6
31.2
Moon
28.1
33.4
North node
22.2
22.3
South node
47.2
47.3
Mercury
27.5
58.6
Venus
16.3
49.2
Mars
24.2
55.1
Jupiter
25.2
37.4
Saturn
6.5
46.5
Uranus
52.3
52.3
Neptune
57.3
57.5
Pluto
7.5
4.1
Entries are gate.line.
Result.
Two channels are defined: 7–31 (G centre to Throat) and 27–50 (Sacral to Spleen).
Defined centres: G, Throat, Sacral and Spleen.
Open centres: Head, Ajna, Heart, Solar Plexus and Root.
The Sacral centre is defined, and no motor connects to the Throat. The type is therefore Generator, with Sacral authority.
The profile is 6/2.
Stability.
The channels and centres are the same at every half hour of the birth day. They do not depend on the birth time.
The profile does. The Sun crosses from line 5 to line 6 of gate 3 at 00:57 local time. A birth five minutes earlier gives profile 5/2.
The state vector. The published vector for this example is (0.22, 0.07, 0.37, 0.82).
Its sum of squares is 0.0484 + 0.0049 + 0.1369 + 0.6724 = 0.8626.
Normalised, it becomes (0.237, 0.075, 0.398, 0.883).
The Yellow axis dominates, so the failure class is 𝕆.
The channels themselves are Collective (7–31, Blue) and Tribal (27–50, Green). The Yellow weight therefore comes from the type, profile and cross terms.
The example shows the order of the computation. Every step is arithmetic on public data. Two implementations given the same input return the same chart.
4. Where the blueprint meets illness
4.1 Two failure lists that belong together
The SWARP line describes how a person’s learning cycle breaks. Coherence medicine describes how a body’s regulation breaks. It distinguishes five regulatory faults:
loss of coherence;
pathological coherence, as in Parkinsonian beta synchrony or epilepsy;
reference drift, as in essential hypertension;
structural destruction, as in autoimmune loss of beta cells;
element failure, where one sensor or effector fails inside an intact loop.
Read in the net, the five faults are five configurations of one strand:
Regulatory fault
In the net
Loss of coherence
lost phase-lock
Pathological coherence
uniform closure where mixed closure belongs: it holds and no longer turns
Reference drift
closure in the wrong direction
Structural destruction
the dial reads too sharply
Element failure
one crossing breaks
The two lists describe one closure at two depths: the person and the tissue. If identity over scale holds, they should correspond:
Failure class
Corresponding regulatory fault
Blue (script kept too long)
reference drift
Yellow (model not revised)
pathological coherence, followed by collapse
Green (coupling lost)
loss of coherence
Red (signal registered late)
structural damage found late
This correspondence is the bridge between the blueprint and the illness. It is stated here as a prediction. Section 7 gives the count that tests it.
4.2 Conditioning, counted
Earlier work on the personal biofield named conditioning as a source of immune disturbance. Conditioning is a life lived in the pattern of another person or environment, against one’s own.
The net turns this into arithmetic. A mixed strand closes only if it keeps one direction.
Join a strand running (−1, 0, +1, −1, 0, +1) to a strand running (+1, 0, −1, +1, 0, −1). The windows on the seam are:
(0, +1, +1), sum 2, open;
(+1, +1, 0), sum 2, open.
Every other window closes.
So a life in the wrong direction costs exactly two open windows on the seam. They do not go away. They carry tension as long as the two directions are joined.
Read in the net, autoimmunity is the dial reading too sharply on that seam. The body discharges tissue that fits, because it lies on a seam that does not.
4.3 The blockage as the place of the complaint
The SWARP support for women in menopause uses the blueprint in the one way the frame allows.
The rhythm change is the same for everyone: the monthly closure becomes erratic and then stops.
Where it shows depends on the person’s blockage: in sleep, in mood, or in the body.
This is the proposition of Section 3.1 in practice. The state sets when; the address sets where.
5. What has been measured
5.1 The birth season
Many conditions. Boland and colleagues studied 1,749,400 patients born between 1900 and 2000 and tested 1,688 conditions. Fifty-five conditions depended on birth month after false-discovery-rate correction at 5 per cent. At that rate about 0.05 × 55 = 2.75 of the 55 are expected false. The signal is real.
Multiple sclerosis is the best measured case. A meta-analysis of 151,978 patients gives observed-to-expected birth ratios in its conservative set:
April 1.07;
May 1.11;
October 0.94;
November 0.88.
The hemisphere mirror. In Australia the pattern is mirrored. People born in November–December carried 1.32 times the risk of people born in May–June. After adjustment for maternal ultraviolet exposure in the first trimester, no association with birth month remained.
This mirror decides between season and sky. The Sun’s place among the stars is the same in both hemispheres. The season is not. The signal follows the season, and through it the light the mother received.
The scale for one person. MS prevalence in the Netherlands is 210.4 per 100,000 (2019). An excess of 11 per cent for May births adds 0.002104 × 0.11 = 0.00023. That is about one extra case per 4,300 people born in May.
The effect is real for the population and small for the person. In the frame of Section 3 it belongs to line three, the field at birth. It is part of the address.
5.2 Solar activity
One study of historical Norwegian records found that people born near solar maximum lived 5.2 years less than people born near solar minimum. The effect was stronger in girls. A study of national cohorts from ten countries, births 1751–1915, found no such relation. The question stays open, and line three of the blueprint is built to settle it.
5.3 The sign of the Sun
Austin and colleagues studied 10,674,945 residents of Ontario and searched 223 diagnoses for an association with astrological sign. They tested 24 associations in a second cohort. Two reached p < 0.05: Leo with gastrointestinal bleeding, and Sagittarius with fracture of the upper arm.
At α = 0.05 over 24 tests, the chance of at least two such results by chance alone is 0.34. After correction, the threshold is 0.05 / 24 = 0.0021. Neither association passed it.
This study does not test the blueprint. The blueprint is not a sun sign. It does show the condition every count must meet: the prediction must be fixed before the data are seen.
5.4 Tested by use
Traditional medicine is a long experiment with a fixed recipe. The recipe was set before each new patient, and generations saw the outcome. What visibly harmed or did nothing fell away. This is selection by fit at the scale of a culture.
The record shows how often that selection was right:
Ge Hong described cold extraction of Artemisia against fevers around 340 CE. From it came artemisinin, now the core of malaria treatment.
Foxglove became digitalis.
Willow bark became aspirin.
Bloodletting, long dismissed, is now the treatment of haemochromatosis and polycythaemia vera. In the CYTO-PV trial, cardiovascular death or major thrombosis fell from 9.8 to 2.7 per cent.
Arsenic, from Chinese medicine, cures more than 90 per cent of acute promyelocytic leukaemia.
Tai chi reduces falls, with high certainty.
Long use shows that a practice holds. It does not show which part of it works, or that its explanation is right. Bloodletting helps where there is too much iron, not for the reasons Galen gave.
The tie between birth season and constitution belongs to this class. Ayurveda and Chinese medicine both make it. It is now measured.
6. The human layer
A blueprint is not read by an instrument alone. It is read by a person who is ill, or afraid of becoming ill.
Three consequences follow for that person.
First, illness is not a verdict on the address. The address does not age and does not sicken. What changes is the state. A person with an open window on a given seam is not destined for the illness that lives there. They know where to look when tension rises.
Second, the remedy for conditioning is not effort. Two open windows on a seam close only when the directions agree again. That means a change of work, of house, of relationship or of rhythm. More effort in the wrong direction adds load to the seam. The person who recognises this gains something no medicine gives: a reason for the exhaustion that is not their own weakness.
Third, the ordinary means carry the most weight. The reference work Where Illness Comes From, Completed lists the measures with the largest effects:
social connection, with an odds ratio for survival of 1.50;
movement;
light at the right hour;
sleep;
the mother’s body for the newborn.
Each restores phase-lock. The blueprint does not replace them. It says which of them a given person needs first.
7. Five tests
Each test fixes its prediction before the data are seen. In each, the count runs against a stated chance level. Claude Code runs each count as a separate judge, with its own code. The SWARP data enter only with the consent of their owners.
Test 1. The blockage hit rate.
Claude Code computes each person’s predicted blockage from the frozen address.
A chest strap records heart-rate variability.
Claude Code fixes the thresholds of the four tension states on a first group, then freezes them.
On a second group, it counts how often the predicted blockage is the state the person stays in longest.
The chance level is the frequency of that state in the second group, about 1/4.
Prediction: a hit rate above chance.
Test 2. Failure class and regulatory fault.
The correspondence table of Section 4.1 is frozen.
In persons with a known diagnosis, Claude Code classes the diagnosis by regulatory fault and the blueprint by failure class, and counts the matches.
The chance level is 1/4.
Prediction: matches above chance.
Test 3. Season against sky.
Claude Code codes each birth twice: once from season and latitude only, once from planetary positions only.
It uses the published monthly MS tables from both hemispheres.
Prediction: the season code carries the signal and the planetary code does not.
Test 4. The field at birth.
Claude Code computes the solar-cycle phase, the geomagnetic load in gestation and the local field for each birth.
It sets these against the same outcome as Test 1.
Prediction: line three adds predictive power to lines one and two. If it adds none, it is removed.
Test 5. The conditioning seam.
Claude Code computes the direction of each person’s ring and the direction of their declared main environment: work and household.
Where the directions oppose, the seam has two open windows.
Prediction: these persons spend more time in the holding state than persons whose directions agree.
A test that succeeds moves its link from READING to MEASURED. A test that fails removes or rewrites the link.
8. Status
MEASURED
The birth-season effect on 55 of 1,688 conditions.
The MS month ratios, their hemisphere mirror, and the maternal UV adjustment that removes the month effect.
No sun-sign effect after correction for multiple testing.
MS prevalence in the Netherlands of 210.4 per 100,000.
COMPUTED
The count of 9 closing windows out of 27.
The two open windows of the conditioning seam.
The worked chart of Section 3.6: type, centres, channels, profile, and the profile boundary at 00:57.
The normalisation of the published vector.
The expected 2.75 false findings in Boland; one extra MS case per about 4,300 May births; the 0.34 chance level and 0.0021 threshold in Austin.
TESTED BY USE
The tie between birth season and constitution in Ayurveda and Chinese medicine.
The practices listed in Section 5.4.
READING
The blueprint fixes where; the state fixes when.
The five regulatory faults as five configurations of the strand.
Autoimmunity as the dial reading too sharply on the conditioning seam.
PREDICTION (Tests 1–5)
The blockage hit rate.
The correspondence of failure class and regulatory fault.
Season against sky.
The added value of the field at birth.
The conditioning seam.
Annotated references
Konstapel, J. (2026). Where Illness Comes From, Completed. First edition, version 1.0, October 2026. Constable Research, Leiden. constable.blog.Why read? Every chapter of ICD-11, one entry per condition, eight fixed sections: what it is, measured, where it starts, cure status, prevention, treatment, simple means, gaps. Part 0 sets out the rules of the net used in Section 2 of this essay. Reading advice: Read Part 0 and Part 1 first. Then use the entries one condition at a time.
Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper. Eighth edition, version 8.1, 17 October 2026. constable.blog.Why read? The five starting rules, the window count and the register on which the whole frame rests. Reading advice: Read the five rules and the window count before anything else.
Konstapel, J. (2026). The Personal Blueprint of the Net. 4 August 2026. constable.blog.Why read? The split of the blueprint into a discrete invariant and a continuous state, with the update rule and the closure moment.
Konstapel, J. (2026). The Narrative Signature Engine. 20 May 2026. constable.blog.Why read? The coordinate of the blueprint, its inputs and the worked example of Section 3.6. Reading advice: Recompute the example from the birth data, as Section 3.6 does.
Konstapel, J. (2026). Born into Your Failures: Expectation Failure as Fractal Karma in the SWARP Model. constable.blog.Why read? The failure class of each axis, the weights from chart to quaternion, and the mapping from circuits to axes.
Konstapel, J. (2026). The Personal Biofield and Immune Coherence. 27 May 2026. constable.blog.Why read? The four propositions that first tied the blueprint to immunity, including conditioning.
Konstapel, J. (2026). Het Coherentieprincipe. 20 July 2026. constable.blog.Why read? The five regulatory faults of Section 4.1.
Konstapel, J. (2026). The Chart Is Silent, the Field Is Not. 2 September 2026. constable.blog.Why read? The electromagnetic profile at birth, and the blockage as the place where a shared rhythm change shows.
Konstapel, J. (2026). The Body Is Also a Net. 1 September 2026. constable.blog.Why read? The design of Test 1: the frozen address, the four tension states and the hit rate.
Boland, M. R., Shahn, Z., Madigan, D., Hripcsak, G. & Tatonetti, N. P. (2015). Birth month affects lifetime disease risk: a phenome-wide method. Journal of the American Medical Informatics Association. doi:10.1093/jamia/ocv046.Why read? The widest search for birth-month effects: 1,688 conditions in 1,749,400 patients. Reading advice: Note that the correction is a false-discovery rate. That is why 55 findings carry only about three false ones.
Dobson, R., Giovannoni, G. & Ramagopalan, S. (2013). The month of birth effect in multiple sclerosis: systematic review, meta-analysis and effect of latitude. Journal of Neurology, Neurosurgery & Psychiatry 84(4):427.Why read? 151,978 patients, and the dependence of the effect on latitude.
Staples, J., Ponsonby, A.-L. & Lim, L. (2010). Low maternal exposure to ultraviolet radiation in pregnancy, month of birth, and risk of multiple sclerosis in offspring. BMJ 340:c1640.Why read? The southern-hemisphere mirror, and the adjustment that removes the month effect. Reading advice: This is the decisive study between season and sky.
Willer, C. J. et al. (2005). Timing of birth and risk of multiple sclerosis: population based study. BMJ 330:120.Why read? The first large pooled analysis: 42,045 patients from four northern countries.
Lemmens, C. M. C. et al. (2024). Prevalence and incidence of multiple sclerosis in the Netherlands. Multiple Sclerosis and Related Disorders 93:106207.Why read? The Dutch prevalence used in Section 5.1.
Austin, P. C., Mamdani, M. M., Juurlink, D. N. & Hux, J. E. (2006). Testing multiple statistical hypotheses resulted in spurious associations: a study of astrological signs and health. Journal of Clinical Epidemiology 59(9):964–969.Why read? Ten million people, 223 diagnoses, and a clean demonstration of why a prediction must be fixed first. Reading advice: Use it as the template for how a judge corrects for the number of tests.
Skjærvø, G. R., Fossøy, F. & Røskaft, E. (2015). Solar activity at birth predicted infant survival and women’s fertility in historical Norway. Proceedings of the Royal Society B.Why read? The positive study behind line three of the blueprint. Read it beside the null study over ten countries in Biogerontology (2009).
Tu, Y. (2011). The discovery of artemisinin (qinghaosu) and gifts from Chinese medicine. Nature Medicine 17:1217–1220.Why read? The clearest case of a practice tested by use, traced from a fourth-century text to a modern cure.
Marchioli, R. et al. (2013). Cardiovascular events and intensity of treatment in polycythemia vera. New England Journal of Medicine 368:22–33.Why read? The CYTO-PV trial. Bloodletting, measured.
Sherrington, C. et al. (2019). Exercise for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews.Why read? 108 trials. Tai chi needs no equipment and reduces falls with high certainty.
Holt-Lunstad, J., Smith, T. B. & Layton, J. B. (2010). Social relationships and mortality risk: a meta-analytic review. PLOS Medicine 7(7):e1000316.Why read? An odds ratio for survival of 1.50 from social connection. It is larger than many treatment effects.
McWhinney, W. (1992). Paths of Change: Strategic Choices for Organizations and Society. Sage.Why read? The four realities behind the four axes of the state vector.
Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience 11:127–138.Why read? The account of a living system as one that reduces surprise. It underlies the update rule of the state.
Where Illness Comes From, Completed is a free reference work of 1,253 pages. It covers every chapter of the International Classification of Diseases, eleventh revision (ICD-11), without exception. One entry treats one condition.
Every entry has the same eight sections, in the same order: what it is, measured, where it starts, cure status, prevention, treatment, simple means, gaps.
The uniformity is the instrument. The same eight questions are put to tuberculosis, cataract, diverticulitis and depression. The answers can then be laid side by side.
Three of the eight do the real work. Can it be cured? What prevents it? And what can be done where there is no industry, no cold chain and no money?
The book says of itself that the third question “turned out to be the one that reorganises the book”. This essay follows that thread. It concentrates on what is unusual in the book: the findings, the cases and the method that a reader will not find assembled anywhere else.
2. Four words, used narrowly
The spine of the book is a cure status with four values.
SELF-LIMITING. The body completes the discharge itself. Intervention adds nothing.
CURABLE. A treatment exists that removes the condition.
CONTROLLABLE. The state can be held. The deviation remains.
RELIEF ONLY. Nothing removes it.
The words are kept narrow on purpose. A treatment that must continue for life does not cure. Phenylketonuria on a lifelong diet is therefore CONTROLLABLE. So is haemochromatosis held by repeated phlebotomy. Gastric MALT lymphoma, cleared by one course of antibiotics, is CURABLE. The book names it as the type case.
Where the honest answer differs by stage, cause or strain, the label is split. Colorectal cancer in the bowel wall has a five-year relative survival of 91.3 per cent. With liver and lung deposits it has 16.9 per cent. One word for both would be false at one end.
The narrowness has consequences that the book enforces against itself. A penile prosthesis was once called “a mechanical cure” in one entry. The native tissue is destroyed and its function substituted, so the deviation remains. The label was corrected. Trachoma appeared twice. One entry split it three ways and reserved RELIEF ONLY for established corneal blindness. The other simply said CURABLE. For a book whose subject is what can and cannot be cured, the book notes, “that omission is the whole of the matter”.
Prevention carries a second scale: STRONG, MODERATE, WEAK, CONTESTED. A compound grade takes the lower value. CONTESTED overrides the others, because disagreement among sources is a statement about the state of the evidence, not its strength. The checking resolved 239 compound grades this way.
3. The simple means, audited
A simple means, in the book’s definition, prevents, cures or controls a condition without a long industrial supply chain behind it. A sachet of salt and glucose. A bar of soap. A bed net. A lamp at the right hour. Thirty minutes of walking. A mother’s chest.
Part 3 audits this class. Its dossier holds forty-six interventions, graded as fifty-nine rows. About half of what is commonly claimed for the class does not survive the grading. Both halves are printed, and the order is fixed: what holds first, what does not hold second.
What holds. Reduced-osmolarity oral rehydration solution, at a UNICEF indicative price of US$0.05 a sachet. Against intravenous rehydration, across seventeen trials and 1,811 children, the treatment-failure risk difference is 4 per cent. Roughly one child in twenty-five on oral therapy still needs a drip. Insecticide-treated nets: all-cause child mortality rate ratio 0.83, high certainty. Immediate kangaroo mother care: 28-day mortality 12.0 against 15.7 per cent in 3,211 infants. That is a number needed to treat of 27, and no equipment beyond the mother’s body. Magnesium sulfate for eclampsia, an inorganic salt solution at about US$0.33 an ampoule. Toothbrushing in intensive care: hospital-acquired pneumonia risk ratio 0.67 across fifteen trials. Exercise for knee osteoarthritis: twelve points less pain on a scale of 100, high certainty.
What does not hold, or holds less. The kitchen version of rehydration solution has no mortality estimate at all. Its documented risk is hypernatraemia from mismeasured salt. “Measurement is exactly the step that packaging removes.” Vitamin A supplementation is CONTESTED: a mortality risk ratio of 0.76 without the two-million-child DEVTA trial, 0.88 with it, and 0.96 for DEVTA alone. The 72 per cent protection of folic acid is recurrence prevention, in women with a previous affected pregnancy. For first occurrence there is one trial, not significant. Handwashing reduces diarrhoea episodes by about 30 per cent; no trial establishes fewer deaths. Building latrines without changing their use has not reduced diarrhoea in cluster-randomised trials. Social connection carries one of the largest associations in epidemiology, an odds ratio of 1.50 for survival. The widely repeated comparison with “fifteen cigarettes a day” has no numeric basis in the papers.
Light therapy shows both sides in one entry. As treatment for seasonal affective disorder its effect size is 0.84. In the single head-to-head trial against fluoxetine, light reached 0.80 and the drug 0.24. The same review records hypomania at a number needed to harm of 8. That figure, the book notes, “is rarely mentioned when light is described as side-effect-free”.
Two counter-cases. Vaccination is STRONG for effect by a wide margin: 154 million lives saved since 1974. Yet it sits outside simple means by construction, because every vaccine but one needs an unbroken 2 to 8 °C chain. Amoxicillin is the second case. It is off patent and costs about five US cents a tablet. Its active ingredient runs through more than ten industrial steps and up to about 70 kg of waste per kilogram. There were eight active US drug master files, held by seven companies. “Off-patent describes a legal status. It says nothing about how many steps, how much waste, or how few producers.”
From this the book draws its central reading: “a simple means is not a cheap means. It is a means whose chain closes at the point of need.”
The short list. No validated catalogue of simply producible essential medicines exists. What can be asserted, item by item, is a list of eleven: oral rehydration salts, zinc sulfate, iodised salt, folic acid, ferrous salts, vitamin A capsules, chlorhexidine 7.1 per cent gel, magnesium sulfate injection, sodium chloride and glucose solutions, paracetamol, oral vitamin D.
One WHO production manual exists for the whole class, for oral rehydration salts (2006). For three to four million sachets a year it specifies a mixing room of about 30 m², a packing room of about 45 m² and a quality-control laboratory of about 15 m². Part 3 closes on one sentence: “Simple does not mean improvised. The shortest chain in medicine still has a laboratory in it.”
4. Where the cheapest cures sit
Across the chapters the third question produces a map. Some of its points are well known to specialists and almost unknown outside them.
Cancer cured by antibiotics. Early gastric MALT lymphoma is cleared in roughly 60 to 80 per cent of cases by one to two weeks of generic antibiotics costing a few dollars. Helicobacter pylori eradication more broadly clears the organism in 70 to 94 per cent of courses. Ulcer recurrence falls from about 60 to 70 per cent a year to under 5 per cent. The organism is a Group 1 carcinogen behind 75 to 90 per cent of non-cardia gastric cancer.
Metastatic cancer cured with off-patent medicines. Four cases stand out. Widely metastatic testicular cancer: roughly 90 to 95 per cent of good-risk patients cured by bleomycin, etoposide and cisplatin. Stage IV Hodgkin lymphoma: 82.8 per cent five-year relative survival on four listed medicines. Acute promyelocytic leukaemia: over 90 per cent cured by all-trans retinoic acid and arsenic trioxide, without chemotherapy in non-high-risk disease. Childhood acute lymphoblastic leukaemia: roughly 90 to 94 per cent cured in high-income settings on a protocol built entirely from off-patent listed drugs.
Treatment that is the removal of blood. In haemochromatosis each 500 mL unit removed takes 200 to 250 mg of iron with it. A homozygote de-ironed before cirrhosis has normal life expectancy. The consumables are a needle, a bag and a nurse. Where the patient donates, the cost to the system is negative. In polycythaemia vera the CYTO-PV trial compared haematocrit targets: cardiovascular death or major thrombosis fell from 9.8 to 2.7 per cent.
Food as cure, and food that is not food. Severe wasting is cured by ready-to-use therapeutic food, with recovery around 80 to 90 per cent in outpatient programmes. The instrument that finds the cases is a paper tape around the upper arm, “and a mother can use it on her own child”. Phenylketonuria shows the limit. The heel-prick test costs a few dollars. The treatment is diet. But the phenylalanine-free amino-acid mixture is a manufactured medical food costing roughly US$5,000 to 25,000 a patient-year. “There is no generic pathway for medical foods.”
The eye, the ear and the skin. Cataract surgery at the Aravind Eye Care System cost US$29.02 per operation in 2013. Its own manufacturing arm brought the intraocular lens to US$3. The complication rate was 1.5 per cent, reported at about half that of the British National Health Service. In the PROSPER trial, 750 tea pickers in Assam received reading glasses. Productivity rose 21.7 per cent overall and 31.6 per cent among those over fifty. Cerumen impaction, a WHO-listed cause of hearing loss, is removed with water and a rubber bulb. The Epley manoeuvre cures roughly 80 per cent of positional vertigo in one session of about two minutes.
The heart and the lungs. Five low-priced interventions carry measured mortality benefits. Benzathine benzylpenicillin cuts rheumatic fever recurrence by 80 to 90 per cent, at about US$0.59 a vial. It suffers repeated global shortages “because the margin is too small to hold manufacturers in the market”. One chewed aspirin at the onset of infarction cuts 35-day vascular mortality by 23 per cent. Oral amoxicillin from a community health worker cuts child pneumonia mortality by 35 to 42 per cent. Dexamethasone and spironolactone complete the five. Beside them stand measures that are not drugs at all. Bystander resuscitation doubles to triples survival from cardiac arrest. Prone positioning in severe respiratory distress needs four or five people, pillows and a protocol, for a number needed to treat of 6.
Oxygen, the constraint that returns. In the infectious diseases, where nothing simple works, oxygen is the recurring limit. It is the rate-limiting treatment in childhood pneumonia and the intervention that turns severe RSV bronchiolitis from often fatal to usually survivable. “It cannot be improvised.” The minimum chain is a concentrator with uninterrupted electricity, or cylinder logistics.
Three measures that are not clinical. In the chapter on reproduction, the perinatal period, congenital anomalies and injury, three of the largest reported effects come from outside the clinic.
The first is Sri Lanka’s pesticide regulation. The suicide rate peaked at 47 per 100,000 in 1995 and then fell about 70 per cent. The reported result is about 93,000 lives over twenty years, at under US$50 per life saved. The bans removed specific agents; total pesticide use still rose.
The second is folic acid in flour. A neural tube closes by the twenty-eighth day after conception, before most pregnancies are recognised. Fortification removes the timing problem. Neural tube defects fell 25 to 50 per cent wherever it was done, at one to ten cents per person per year. “The intervention is a decision taken at a mill.”
The third is mass dog vaccination. Seventy per cent coverage interrupts rabies transmission, at about a dollar per dog.
None of the three falls inside a health-service budget. The book names this as a recurring reason why they are underfunded relative to their effect.
5. The two ends of one list
The WHO Model List of Essential Medicines, 24th edition, now holds both ends of the price range. Oral morphine costs cents a dose. Pembrolizumab costs about US$53,000 a year at the United States cash price. Blinatumomab costs about US$180,000 a course at United States list.
The morphine end is the more telling. About 56.8 million people need palliative care each year; roughly 14 per cent receive it. In 2018, 79 per cent of the world’s population consumed 13 per cent of the morphine used for pain. That morphine was 1 per cent of the 388 tonnes manufactured worldwide: about 3.9 tonnes. Oral morphine is the only intervention that appears in the simple-means section of every cancer entry. It is also the one most consistently unavailable. The barrier is legislation, licensing and fear of diversion, not price, patent or technology.
The same pattern appears in insulin. Human insulin is priced at 1.2 to 1.8 times its estimated production cost. Glargine is priced at 5.6 to 7.8 times its own. “The economic problem is in the analogues, not in insulin.”
6. The chain
Part 4 asks the question that no single entry can answer. A medicine exists, it is cheap to make, and it is not available. How short can the production chain be, and who may legally make what? The answer is given as law and institution, never as a recipe.
Three Dutch cases. Each has a published price before and after registration.
Chenodeoxycholic acid for cerebrotendinous xanthomatosis: from €308 to about €153,300 per patient-year after orphan registration in 2017. That is a factor of 498.
Cysteamine eye drops for cystinosis: the pharmacy preparation at €1,839.24 per patient-year against €53,300 for the registered product. A factor of 29.
Mexiletine for myotonia: €6.95 against €23.39 a capsule. A factor of 3.4.
The multiples are the headline. The book states that they are not the most important finding.
What actually binds. The Amsterdam preparation of chenodeoxycholic acid did not stop on law. The minister wanted it to continue. The competition authority later fined the rights holder €19.6 million for abuse of a dominant position. The preparation still stopped for sixteen months, because the national institute found unknown impurities in the starting material above the pharmacopoeial limit. “The binding constraint on the short chain is qualification of the active substance, not the legal exemption.”
Registration closes the short chain. The cysteamine preparation existed first and worked. Registration did not add a treatment. It moved an existing one inside the authorisation regime. After that the preparing pharmacies could no longer supply it. The Dutch counter-move, from June 2021, is for academic hospitals and a compounding pharmacy to register their own preparations. The book calls it “the short chain accepting the long chain’s paperwork in order to keep the short chain’s cost base”.
A precedent. In January 2021 the Dutch national health care institute advised against reimbursing registered mexiletine. It advised bringing the pharmacy preparations into the basic package instead. The regulation followed in July 2021. The compounded route became the reimbursed standard of care at roughly a third of the registered price.
Two legal architectures. The European exemption is relational. Directive 2001/83/EC exempts the magistral and the officinal formula, tied to the pharmacy’s own patients, with no volume ceiling. The American Drug Quality and Security Act of 2013 does the opposite. Its section 503B creates a licensed industrial tier with no patient relationship at all, regulated as manufacture. On 30 September 2026 there were 98 registered outsourcing facilities.
Off-patent is not secure supply. Across 554 generic active ingredients, 56.3 per cent have five or fewer certified manufacturers worldwide. In 2025 the Netherlands had 1,134 medicines in shortage, affecting 3.5 million people. The medicines named are levothyroxine, rifampicin, salbutamol, prednisolone, quetiapine, lithium and oxazepam: old, cheap, off-patent and preparable. Pharmacy preparation appears once in the entire national shortage report. The two datasets describe the same medicines from opposite ends and have never been joined. Appendix B calls this the only question in Part 4 with no published answer at all.
Where the chain cannot be shortened. Monoclonal antibodies: the process defines the molecule. Cell therapy: the product is the patient’s own cells. India’s point-of-care CAR-T is expected at about US$50,000 against about US$400,000 in the United States. That is a factor of 8, not the factor of hundreds that compounding achieves for a small molecule. Radiotherapy: the constraint is the workforce, not the machine count. Dialysis: a permanent consumables line. Transplantation shortens that chain. By the second year after transplant the annual cost is €15,018 against €92,616 for in-centre haemodialysis, about a sixth.
The asymmetry of mRNA. An mRNA vaccine cannot be made in a pharmacy. It does not need a global industry either. One production line can make 100 million doses a year with seventy-five to eighty staff. The chain is “short in headcount and footprint and long in technical specificity”. The binding constraint is technology transfer, not scale.
7. What no longer holds
Most reference works record what is currently believed. Several chapters of this one also record what was believed and has been retired. Readers will meet these claims in older textbooks, in clinical habit and in marketing.
Some of the retired certainties are striking. In idiopathic pulmonary fibrosis, the standard immunosuppression of decades raised mortality from 1 to 8 per cent in PANTHER-IPF. Adaptive servo-ventilation in heart failure increased cardiovascular mortality in SERVE-HF. Arthroscopic partial meniscectomy did no better than sham surgery in FIDELITY. Paracetamol was identical to placebo for acute low back pain in PACE, with seventeen days to recovery in both arms. In head injury, the CRASH trial found increased mortality with corticosteroids. In neonatal encephalopathy, cooling helps in high-income settings and raised mortality in the HELIX trial in South Asia. It is a treatment “that works where it was first tested and does not travel”.
Diverticular disease, the entry that prompted the book, contains two reversals. Nuts, seeds and popcorn were forbidden for decades. In a cohort of 47,228 men, popcorn carried a diverticulitis hazard ratio of 0.72: significantly protective. And antibiotics for uncomplicated diverticulitis, standard for seventy years, showed no demonstrated benefit in the DIABOLO and AVOD trials. Hospital stay was shorter without them. The book notes that the failure also puts the infective model of the disease in doubt, and that this question is open.
One observation runs through the chapter on the gut, the frame and the waterworks. Structural findings on imaging are largely age-normal. Disc degeneration is present in about 37 per cent of people without symptoms at twenty and about 96 per cent at eighty. Full-thickness rotator cuff tears are present in over half of people above eighty, many of them painless. “A finding that is present in nearly everyone of a given age is a property of the age, not of the complaint.”
8. The reservoir
The cancer chapter collects the autopsy evidence in one place. It comes from people who died of something else and whose organs were then examined systematically.
Incidental prostate cancer is present in 5 per cent of men under thirty and 59 per cent of men over seventy-nine. Occult differentiated thyroid cancer is found in 4.1 per cent on partial examination of the gland and 11.2 per cent on whole-gland examination. The difference is the point: “how much cancer is found depends on how hard the pathologist looks.” Occult invasive breast cancer is found in 0.85 per cent; in-situ cancer in 8.9 per cent.
The size of the reservoir tracks the size of the overdiagnosis problem. Prostate and thyroid have large reservoirs and large overdiagnosis. Breast has a small invasive reservoir and a moderate, contested figure concentrated in in-situ disease. South Korea added thyroid ultrasound to health checks; incidence rose roughly fifteenfold between 1993 and 2011 while mortality stayed flat. Nasopharyngeal carcinoma is the control case. No occult reservoir is found at autopsy, and screening by plasma Epstein-Barr virus DNA gives a large stage shift without a measurable overdiagnosis penalty.
The book adds a second list that is rarely made: where overdiagnosis has never been quantified. Gastric, liver, kidney and bladder cancer. Cervical intraepithelial neoplasia, colonic adenoma, uterine fibroids, incidental meningioma, monoclonal gammopathy and monoclonal B-cell lymphocytosis. Several of these have reservoirs as large as the measured ones, or larger.
Childhood cancer carries its own quantity. Paediatric tumours have about 0.13 mutations per megabase against 1.8 in adults, roughly fourteen times fewer. There is almost nothing for checkpoint inhibitors to work on, and checkpoint blockade has broadly failed in paediatric oncology. The immunotherapies that do work in children target a lineage antigen. Survival diverges threefold between rich and poor countries. The book states that the divergence is not biological. A leading cause of death is abandonment of treatment, driven by travel cost, lost income and accommodation.
9. Gaps that are content
The book treats the absence of a number as a finding in its own right.
A whole ICD-11 chapter sits outside the global burden accounts. No sleep-wake disorder is a level-three cause in the Global Burden of Disease study. That includes obstructive sleep apnoea, which on the only global estimate affects 936 million adults aged 30 to 69. The funding statement of that study is the single word ResMed.
Chapter 17 of ICD-11, conditions related to sexual health, has existed in released form for seven years. Not one of its categories has a prevalence estimate measured against its own definition. Erectile dysfunction carries a published global prevalence range of 3 to 76.5 per cent. “A 25-fold spread is not a prevalence estimate. It is a measurement failure.”
Seventeen circulatory and respiratory conditions carry no global burden figure in the tables retrieved. Heart failure alone affects about 64 million people; allergic rhinitis about 400 million.
Four WHO fact sheets give their headline figure without a data year. Two expected fact-sheet addresses return pages with no Key Facts and no body. “Any work citing a ‘WHO fact sheet’ for either is citing a document that does not exist.”
10. The frame underneath
Part 0 sets out the rules the book applies. They come from the Vacuum.Net theory, governed by its Foundational Paper, eighth edition, version 8.1. The theory was built at the bottom, with one strand and one test, and no constants given in advance. The book applies it upward and does not repeat the derivation.
The fishnet. The standing image is a fishnet under tension. The net is one strand. Where it crosses itself there is a crossing; between crossings lie openings, the meshes. A winding is a stretch of strand that returns on itself in step. A knot is windings that hold each other in three directions. One rule decides what stays: a winding persists only if it returns on itself in step. What does not close passes its tension to the meshes beside it.
The register. At a crossing the strand passes over, passes under, or does not cross. These are written +1, −1 and 0: a trit. Counting in balanced ternary needs no separate minus sign; negation is mirroring. Three is also the cheapest base. The cost of writing numbers in base r goes as r / ln r. That gives 2.73 for base three against 2.89 for base two.
Windows. Three crossings form a window. It is closed when its three trits add to a multiple of three. Of the 27 possible windows, 9 close. Three are uniform, with three equal trits: the form of a crystal, holding and not turning. Six are mixed, with three different trits: a form that keeps turning while it stays closed. Every content changes; the closure holds. The book calls this the form of living matter. Where a uniform stretch meets a mixed stretch, the window on the seam stands open and carries tension. “That seam is where this book finds most of its deviations.”
The test. When one form reads another there are three outcomes. +1: it fits, and stays. −1: it does not fit, and is taken apart. 0: it is not read, and simply remains. The third outcome carries most of the book. “Health is not the absence of misfits; misfits arise every day, in everyone, in thousands. Health is a test that reads fast enough.”
Illness read through the test. The body carries a reading layer that runs the test at cell scale. There are places where it must not discharge what it reads: the gut wall at every meal, the eye, the child in the womb, which is by signature half foreign. So the body carries official zero-tokens that turn the local reading to zero. Clinicians call them checkpoints. Without them the body destroys itself; with them it has a blind spot. A test that reads too weakly lets misfits through. A test that reads too sharply discharges healthy tissue, which is autoimmunity. “One layer, one dial, two directions of failure.”
Cancer then sits in one sentence: a tumour is the remainder of a selection. For years every variant without the zero-token was discharged and every variant with it was not. The autopsy reservoir of Section 8 is the measured form of that remainder: altered tissue that the test around it held. Ageing is read as lost phase-lock — heart-rate variability, deep sleep, hormone tides drifting out of step — so lifespan is read as coherence-span. Tuberculosis in a quarter of the world, carried and mostly never expressed, is the zero outcome at the scale of a population.
A check at the bottom. Part 0 also prints the theory’s measured anchor. The threshold for gravitational anomalies, a₀, comes out as c·H₀/2π. With H₀ = 70 km s⁻¹ Mpc⁻¹ this gives 1.084 × 10⁻¹⁰ m s⁻², against about 1.2 × 10⁻¹⁰ measured from galaxy rotation curves. With H₀ = 67 it gives 1.050 × 10⁻¹⁰. Both are printed, because the two measurements of the Hubble constant disagree. For the book the point is one only: the frame was not built for biology and bent to fit.
Readings, not mechanisms. The book uses the frame to order, never to supply numbers. Every number comes from the standing literature. Where a measured fact is set beside the rules, the sentence is marked “Read in the net” and closes with “This is a reading, not a mechanism.” There are twelve such readings in the book. In place of cause the book uses four relations: selection by fit, identity over scale, constitutive relation, and correspondence. A dictionary maps every word in the entries — thread, winding, the test, tension, depth, the legitimate zero — onto exactly one element of Part 0.
Two readings show how the frame earns its place in a medical book. On diverticula: “the wall is a mesh under tension, and it yields where it is already open.” The vascular penetration points of the colon wall are the pre-existing openings. Nothing new has to be made for a pouch to form. And on simple means, quoted above: a means whose chain closes at the point of need. Price and patent are read at the wrong depth. What decides is whether the thing can be completed where it is used.
11. A book that shows its own accounting
The last distinctive feature is the apparatus. It is unusual for any reference work, and it makes this one usable at a known level of confidence.
The dagger. Every numeric claim was checked against a named source during compilation. A figure read off its source, or resting on a named source not opened again, is printed plainly. A figure whose source could not be opened at all carries a dagger (†). The book counts 2,361 daggers against 6,352 figures resting on a named source. A figure with no source at all was struck, and the sentence says in words what is known without it. “No figure in this book is an estimate.”
The Register of Contradictions. Appendix A records thirty-one places where a single WHO fact sheet contradicts itself. Hepatitis C appears as about 47 million chronic infections in the key facts and 50 million in the sum of the body’s own regional figures. Dengue risk appears as about four billion people and as 5.6 billion. A further table lists pairs of sources of comparable standing that disagree, printed both ways with no winner picked. Four of these differ in method or vintage rather than in the world, and the book says which.
The book against itself. Three conditions turned out to have a full entry in two chapters, written by different hands from different dossiers. Merging the two fibroid entries found one visible disagreement and eight that nobody had seen. “The lesson is not about fibroids.” Two competent writers produced nine disagreements, and only one concerned the headline answer.
An invented number. A number needed to treat of about 50 for aspirin in secondary prevention circulated through the preparation, plausible and correctly cited. The cited paper publishes event rates of 6.7 and 8.2 per cent a year, and no number needed to treat. Recomputed, the figure is 67 a year and 33 over two years. “The figure had no origin.”
The drift. Appendix C records 3,388 corrections over four rounds of checking by readers who had not produced what they read. The dominant finding is that the book “systematically presented itself as better sourced than it was” — everywhere, by every hand, in the same direction, and never by a large step. About 1,480 provenance tags were lowered. “Nothing was discovered to be false by this. What was discovered is that the book did not know what it knew.” The checks themselves were checked. A comparison of entry titles had silently skipped all forty-two infectious-disease entries because they were headed one level higher. An assembly tool had dropped 541 daggers at line breaks. Both faults were found and corrected, and both are recorded.
12. What the book makes visible
Read across its 1,253 pages, the book produces two lists that no other single source assembles.
The first is a list of interventions that are cheap, old and effective, with their measured effects and their honest limits. Many sit outside the clinic: at a mill, in a ministry of agriculture, in a veterinary campaign. Others are in the clinic but not in reach, because a chain does not close where they are needed.
The second is a list of conditions where nothing simple works, named as plainly as the successes. Glaucomatous field loss has never been recovered by any treatment. Cochlear hair cells do not regenerate. Multimorbidity remains CONTROLLABLE after thirty-eight trials and 18,073 participants showed no effect of structured deprescribing on admissions or quality of life.
Between the two lists runs the fishnet of Part 0. It does not explain the numbers. It orders them by one question, asked at every depth: does it close where it is used?
Annotated references
The references below are those that carry the most weight in the book and in this essay. Each has two parts: why it is worth reading, and where useful, how to read it. The book itself cites several hundred sources; its appendices name the sources it could not open.
A. The book and its frame
1. Konstapel, J. (2026). Where Illness Comes From, Completed: A Reference Work on Every Chapter of Human Illness, Its Prevention, Its Cure, and What Can Be Done With Simple Means. First edition, version 1.0, October 2026. Constable Research, Leiden.Why read? The subject of this essay. 1,253 pages, every ICD-11 chapter, one eight-section entry per condition. Reading advice: Start with “How to use this book” (pp. 12–13). Then read Part 3 (pp. 1221–1230) and Part 4 (pp. 1231–1240) before any entry. The appendices (pp. 1241–1253) explain how much weight each figure carries. Use the entries as a reference, one condition at a time.
2. Konstapel, J. (2026). Where Illness Comes From. Ninth edition, September 2026. constable.blog.Why read? The predecessor. It walks one frame from first chemistry to the human being and takes its cases where they illustrate a layer. It is not a catalogue. Diverticulitis does not appear in it; that absence prompted the completed book. Reading advice: Read for the frame and the layer-by-layer argument. Use the completed book for any particular condition.
3. Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper. Eighth edition, version 8.1, 17 October 2026. constable.blog.Why read? The governing statement of the rules in Part 0: one strand, closure as the only selection rule, memory, context, scale, and the ternary register. Reading advice: Read the five starting rules and the window count first. The a₀ derivation shows how the frame was made testable at the bottom before it was walked upward.
B. Burden and classification
4. GBD 2023 Collaborators (2025). Global Burden of Disease Study 2023. The Lancet, published 12 October 2025.Why read? The book’s burden basis: 292 causes of death, 375 diseases and injuries, 204 countries, 1990–2023. Reading advice: Note what is not in it. No sleep-wake disorder is a level-three cause, and heart failure is not a cause line. The per-condition Results Tool is the single source that would retire the most daggers in the book.
5. World Health Organization (2019–). International Classification of Diseases, 11th Revision (ICD-11 MMS).Why read? The scaffold of the book. Its chapters 07 (sleep-wake) and 17 (sexual health) are new in ICD-11, and both lack measured epidemiology. Reading advice: Read the chapter introductions and the extension codes (Chapter X). The HA40 aetiology axis shows how ICD-11 moves causation off the diagnosis and onto a separate coordinate.
6. Ferlay, J. et al. / International Agency for Research on Cancer. GLOBOCAN 2022 and GLOBOCAN 2024.Why read? Cancer incidence and mortality by site. The book prints both editions side by side. Reading advice: Cervical cancer moves from 662,301 to 604,196 cases between editions. The book reads this as a methodological revision, not a trend. Do not compare editions as if they were a time series.
C. Simple means and their audit
7. World Health Organization (2006). Oral Rehydration Salts: Production of the New ORS. Geneva: WHO.Why read? The only operational WHO production manual for a simple means. Formulation, raw-material specifications, quality-control tests, premises and shelf life. Reading advice: Read the premises specification. It is the concrete form of the book’s sentence that the shortest chain still has a laboratory in it.
8. Imdad, A. et al. (2022). Vitamin A supplementation for preventing morbidity and mortality in children from six months to five years of age. Cochrane Database of Systematic Reviews.Why read? Nineteen trials, 1,202,382 children, all-cause mortality risk ratio 0.88. The central source for one of the book’s CONTESTED grades. Reading advice: Read the sensitivity analysis with and without the DEVTA trial. The familiar 24 per cent rests mainly on trials before 1994.
9. Magpie Trial Collaborative Group (2002). Do women with pre-eclampsia, and their babies, benefit from magnesium sulphate? The Lancet 359:1877–1890.Why read? 10,141 women in 33 countries. Eclampsia 1.9 against 0.8 per cent, a number needed to treat of 91, from an inorganic salt solution. Reading advice: Read it beside the UNICEF price of about US$0.33 an ampoule. Few interventions combine this effect with this simplicity of manufacture.
10. WHO Immediate KMC Study Group (2021). Immediate “kangaroo mother care” and survival of infants with low birth weight. New England Journal of Medicine 384:2028–2038.Why read? 3,211 infants in five countries. 28-day mortality 12.0 against 15.7 per cent, a number needed to treat of 27, with no equipment beyond the mother’s body. Reading advice: The trial stopped early for benefit, which overstates effect size. The book prints that qualifier with the figure.
11. Clasen, T. F. et al., for the HAPIN Investigators (2022). Liquefied petroleum gas or biomass fuel for cooking and effects on birth weight. New England Journal of Medicine.Why read? 3,200 households. Personal fine-particle exposure fell 66 per cent, verified. Birth weight did not change. Reading advice: Read it with RESPIRE (Guatemala) and CAPS (Malawi). Large attributable fractions and null trials stand side by side, and the book leaves the question open.
12. Holt-Lunstad, J., Smith, T. B. & Layton, J. B. (2010). Social relationships and mortality risk: a meta-analytic review. PLOS Medicine 7(7):e1000316.Why read? 148 studies, 308,849 participants, survival odds ratio 1.50. Larger than many treatment effects in the book. Reading advice: Look at the sub-measures. Complex social integration is strongest and most confounded. Living alone, the hardest measure, is weakest and not significant. The “fifteen cigarettes” comparison is not in the paper as a number.
13. Sherrington, C. et al. (2019). Exercise for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews.Why read? 108 trials, 23,407 participants. Rate of falls 0.77 at high certainty: the only high-certainty result in the book’s chapter on sexual health, the non-disease chapters and the cross-cutting states. Reading advice: Balance and functional training carry the effect. Tai Chi needs nothing at all.
14. Leong, D. P. et al. (2015). Prognostic value of grip strength: findings from the Prospective Urban Rural Epidemiology (PURE) study. The Lancet 386:266–273.Why read? 139,691 people in 17 countries. Each 5 kg less grip strength raises all-cause mortality by a hazard ratio of 1.16. Reading advice: Read it as a measurement that takes seconds and reports on the whole person.
D. Cancer: reservoir, mutation burden, prevention
15. Bell, K. J. L. et al. (2015). Prevalence of incidental prostate cancer: a systematic review of autopsy studies. International Journal of Cancer 137:1749–1757.Why read? 29 autopsy studies. 5 per cent under thirty, 59 per cent over seventy-nine. Reading advice: Set the 59 per cent beside a lifetime risk of dying of prostate cancer of roughly 2 to 3 per cent.
16. Furuya-Kanamori, L. et al. (2016). Prevalence of differentiated thyroid cancer in autopsy studies over six decades: a meta-analysis. Journal of Clinical Oncology 34:3672–3679.Why read? 12,834 autopsies. 4.1 per cent on partial and 11.2 per cent on whole-gland examination. Reading advice: The difference between the two figures is the finding: the amount of cancer found depends on how hard one looks.
17. Thomas, E. T. et al. (2017). Prevalence of incidental breast cancer and precursor lesions in autopsy studies. BMC Cancer 17:808.Why read? 2,363 autopsies. Invasive cancer 0.85 per cent, in-situ 8.9 per cent. Reading advice: Explains why breast overdiagnosis is moderate and concentrated in in-situ disease.
18. Vaccarella, S. et al. (2016). Worldwide thyroid-cancer epidemic? The increasing impact of overdiagnosis. New England Journal of Medicine 375:614–617.Why read? About 470,000 women and 90,000 men overdiagnosed in twelve countries over two decades. Reading advice: Read with the South Korean incidence curve: a fifteenfold rise with flat mortality.
19. Gröbner, S. N. et al. (2018). The landscape of genomic alterations across childhood cancers. Nature 555:321–327.Why read? 961 tumours. 0.13 mutations per megabase against 1.8 in adults. Reading advice: Read it with the clinical consequence the book draws: checkpoint blockade has broadly failed in children, while lineage-antigen therapies work.
20. de Martel, C. et al. (2020). Global burden of cancer attributable to infections in 2018. The Lancet Global Health 8:e180–e190.Why read? 2.2 million infection-attributable cancers, about 13 per cent. H. pylori alone about 810,000. Reading advice: Only four agents are counted. Epstein-Barr virus, HHV-8, HTLV-1 and Schistosoma haematobium are not, which affects several of the book’s cancer entries.
21. World Health Organization and International Agency for Research on Cancer (2026). Analysis of 2022 cancer cases attributable to modifiable risk factors. Nature Medicine.Why read? 37.8 per cent of 2022 cancer cases, about seven million, with a published derivation. Reading advice: The book prefers this figure to the widely quoted but unsourced 30 to 50 per cent. Attributable is not the same as preventable now.
E. Trials that retired standing practice
22. Daniels, L. et al. (2017). Randomized clinical trial of observational versus antibiotic treatment for a first episode of CT-proven uncomplicated acute diverticulitis (DIABOLO). British Journal of Surgery 104:52–61.Why read? 528 patients in 22 Dutch centres. No difference in recovery; hospital stay shorter without antibiotics. Reading advice: Read with AVOD (Chabok, British Journal of Surgery 2012). The trials were not powered for rare outcomes and excluded the immunosuppressed. The supported conclusion is “no demonstrated benefit”.
23. Strate, L. L. et al. (2008). Nut, corn, and popcorn consumption and the incidence of diverticular disease. JAMA 300:907–914.Why read? 47,228 men over eighteen years. Popcorn hazard ratio 0.72 for diverticulitis. Decades of dietary restriction pointed the wrong way.
24. Crowe, F. L. et al. (2011). Diet and risk of diverticular disease in Oxford cohort of EPIC. BMJ 343:d4131.Why read? 47,033 participants. Highest against lowest fibre fifth: hazard ratio 0.59 in the abstract and 0.58 in the results section of the same paper. Reading advice: Read with Peery et al., Gastroenterology 2012, where higher fibre went with more diverticulosis. The outcomes differ: clinical disease against pouches seen at colonoscopy. No trial has randomised fibre.
25. Idiopathic Pulmonary Fibrosis Clinical Research Network (2012). Prednisone, azathioprine, and N-acetylcysteine for pulmonary fibrosis (PANTHER-IPF). New England Journal of Medicine 366:1968–1977.Why read? Stopped early. Mortality 8 against 1 per cent. The heaviest reversal in the respiratory chapter.
26. Sihvonen, R. et al. (2013). Arthroscopic partial meniscectomy versus sham surgery for a degenerative meniscal tear (FIDELITY). New England Journal of Medicine 369:2515–2524.Why read? No benefit over sham surgery. The operation continues at scale. Reading advice: Read with the book’s section on age-normal imaging findings; the two explain each other.
27. Thayyil, S. et al. (2021). Hypothermia for moderate or severe neonatal encephalopathy in low-income and middle-income countries (HELIX). The Lancet Global Health 9:e1273–e1285.Why read? No reduction in death or disability; increased mortality. Reading advice: The clearest case in the book of a treatment that works where it was tested and does not travel.
F. Price, supply and the chain
28. World Health Organization (2025). WHO Model List of Essential Medicines, 24th List. Technical Report Series 1064.Why read? The governing edition for the book. It holds oral morphine at cents a dose and pembrolizumab at tens of thousands of dollars a year. Reading advice: Distinguish a selected medicine from a therapeutic equivalent. The book treats the distinction as load-bearing.
29. UNICEF Supply Division. UNICEF Supply Catalogue, read 19 October 2026.Why read? The working substitute for the retired MSH price guide. Indicative US-dollar prices with material numbers that can be rechecked. Reading advice: The catalogue returns different prices for the same material number on different pages. Treat its figures as indicative and dated, never as a price.
30. Management Sciences for Health. International Medical Products Price Guide. Last edition 2017; website retired after 30 June 2024.Why read? For its history. Four of the book’s research lines named it as the source that would settle the weakest data. It no longer exists as a live product. Reading advice: Any price traceable to it is a 2017 price.
31. Gotham, D., Barber, M. J. & Hill, A. (2018). Production costs and potential prices for biosimilars of human insulin and insulin analogues. BMJ Global Health 3:e000850.Why read? Production cost estimates per phial for human insulin and three analogues. Human insulin is priced near cost; glargine at five to eight times its own. Reading advice: The estimates are for phials only, while current prices cover pens and cartridges too. A second, lower estimate of about US$1 a vial exists; the book prints both.
32. European Parliament and Council (2001). Directive 2001/83/EC on the Community code relating to medicinal products for human use, Articles 3(1), 3(2) and 5(1).Why read? The two sentences on which the whole European short chain rests: the magistral and the officinal exemption. Reading advice: Note what is absent. There is no volume ceiling; the limit is the pharmacy’s own patient relationship.
33. United States Congress (2013). Drug Quality and Security Act, sections 503A and 503B of the Federal Food, Drug, and Cosmetic Act.Why read? The opposite architecture: a licensed industrial compounding tier with no patient relationship, regulated as manufacture.
34. Autoriteit Consument & Markt (2021). Decision on Leadiant Biosciences: abuse of dominant position in the pricing of CDCA-Leadiant. The Hague, 19 July 2021.Why read? The €19.6 million fine in the chenodeoxycholic acid case. The authority held magistral preparation to be an exception to orphan exclusivity and a room for lower prices. Reading advice: Read it with the book’s finding that the preparation stopped on impurities in the starting material, not on law.
35. Zorginstituut Nederland (2021). Advice on reimbursement of registered mexiletine, 14 January 2021.Why read? A national authority prefers the pharmacy preparation to the registered orphan product and brings the preparation into the basic package.
36. Inspectie Gezondheidszorg en Jeugd (2023). Circular on supply of pharmacy preparations beyond the pharmacy’s own patients.Why read? Six conditions that define where the short chain may serve a population in the Netherlands. Reading advice: The circular says openly that this route rests on enforcement discretion, not yet on statute.
G. Palliative care and multimorbidity
37. World Health Organization (2020). Palliative Care. Fact sheet, 5 August 2020, citing the International Narcotics Control Board.Why read? 56.8 million people in need, about 14 per cent reached. In 2018, 79 per cent of the world’s population used 13 per cent of the morphine for pain. Reading advice: Compute the tonnage: 1 per cent of 388 tonnes is about 3.9 tonnes for most of humanity.
38. Temel, J. S. et al. (2010). Early palliative care for patients with metastatic non-small-cell lung cancer. New England Journal of Medicine 363:733–742.Why read? Better quality of life, less depression, less aggressive end-of-life care, and a median survival 2.7 months longer. Reading advice: Read with the 2017 Cochrane review. The quality-of-life benefit survives pooling; the survival benefit does not reach significance.
39. Barnett, K. et al. (2012). Epidemiology of multimorbidity and implications for health care, research, and medical education. The Lancet 380:37–43.Why read? 1.75 million patients in Scotland. More multimorbid people are under sixty-five than over it. Deprivation brings onset forward by ten to fifteen years.
H. Cardiovascular risk
40. Yusuf, S. et al. (2004). INTERHEART. The Lancet 364:937–952; and O’Donnell, M. J. et al. (2016). INTERSTROKE. The Lancet 388:761–775.Why read? Nine and ten modifiable factors with combined attributable risks of 90.4 and 90.7 per cent. Reading advice: Both are case-control studies. An attributable risk from that design is not a demonstrated preventable fraction. The phrase “90 per cent of heart attacks are preventable” is not what either study measured.
41. ISIS-2 Collaborative Group (1988). Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases of suspected acute myocardial infarction. The Lancet 2:349–360.Why read? One chewed aspirin at the onset of infarction: 23 per cent less 35-day vascular mortality, for under one cent.
De meeste vragen over het vacuüm worden gesteld in de taal van het vuur: energie, brandstof, arbeid en rendement. Dat zijn begrippen die beschrijven hoe een keten van omzettingen verloopt, waarbij op elke schakel warmte verloren gaat. Het essay “Reading the Net” betoogt dat die taal haar eigen antwoord beperkt: wie naar energie vraagt, vindt alleen kranen, en ziet over het hoofd wat geen energiehoeveelheid is. Beyond Energy (derde editie) draait de vraag daarom om. Het vraagt niet wat de mens uit het vacuüm kan halen, maar wat het net zelf kan doen, uitgedrukt in zijn eigen grootheden. Dat omslaan van perspectief levert een kaart op die groter is dan de energievraag: het lezen van de toestand van een plaats, het in één toestand houden van twee verre structuren, het selecteren van materie op adres en het in permanente circulatie houden van een configuratie.
De kern van de methode is opvallend systematisch. Het net kent tien beïnvloedbare grootheden — spanning, dichtheidsafwijking, fase, koppeling, adres, diepte, rust, geheugen, handedheid en de tritwet — die elk op vijf manieren kunnen worden benaderd: lezen, instellen, selecteren, koppelen en openlaten. Tien maal vijf geeft vijftig cellen, en elke cel is uitgewerkt, waar mogelijk getoetst en voorzien van een expliciete status: aangenomen, bewezen, berekend, gemeten of open. De technologie van het vuren-tijdperk vult er naar schatting slechts vier van: zij leest en stelt stroom, en zij leest en stelt warmte. De overige cellen zijn leeg, niet omdat ze onmogelijk zijn, maar omdat er nog geen woord voor bestaat.
Het beeld en de rekenregels
De theorie stelt zich het vacuüm voor als een visnet dat in water is gespannen, met draden die overal doorheen lopen en altijd onder spanning staan. Waar draden om elkaar heen slingeren, ontstaan knopen — dat is materie — en tussen de draden liggen mazen met een eigen vorm en grootte. Vijf axioma’s dragen het geheel: één draad die alleen lokaal over zichzelf kan kruisen (met de drie toestanden +1, −1 en 0), sluiting als enige selectiecriterium, geheugen dat het verdere verloop begrenst, context waardoor natuurconstanten functies worden van de lokale spanning, en schaal waarbij elke trede dieper een drievoud aan toestanden toevoegt. Twee eenheden, beide aan waterstof ontleend, fixeren de maat: een lengte-eenheid van 45,6 nm en een tijdeenheid van 1,52 × 10⁻¹⁶ s.
Het cruciale is dat deze regels op een register kunnen worden nagespeeld. Een venster van drie aangrenzende kruisingen sluit wanneer de som van zijn tritsen 0 modulo 3 is; negen van de zevenentwintig mogelijke vensters sluiten. Een ring van tritsen wordt herschreven door één regel, R, die elke trit vervangt door de negatieve som van zijn buren. De methode is daarmee bijzonder streng voor zichzelf: negen voorspellingen werden vooraf opgeschreven, zes hielden stand en drie faalden — en juist de mislukkingen markeren de grenzen van de kaart.
Zes hoogtepunten
Het plafond en de kloof. Eén winding op diepte nul draagt 4,357 × 10⁻¹⁸ J. Onder de aanname van één volledige sluiting per maas per laagtijd volgt een maximaal vermogen van 0,0287 W per maas, oftewel een plafond van 1,38 × 10¹³ W per vierkante meter per laag. De enige daadwerkelijk gemeten kraan — een diode met een holte van 33 nm, die 1,4 pW leverde zonder aangelegde spanning — zit daar op een factor vijf biljoen (5,1 × 10⁻¹²) onder. Dat getal herkadert het veld: het toont niet dat het net zwak is, maar hoe ver het huidige ontwerp van wat het net vervoert verwijderd staat. Hetzelfde rekenwerk voorspelt 70 W/m² bij uitsmering over oppervlak, en een kruisproef op diepte vijf landt binnen een factor 9,5 van de warmtestraling van een kamer van 20 °C.
Drie stappen beneden zijn één stap boven. Groepeer een ring in blokken van drie, beschouw elk blok als één trit van een ouderring, en de regel R werkt op de ouder exact zoals op het kind — mits men drie stappen beneden gelijkstelt aan één stap boven. Het bewijs is kort en elegant: over het lichaam van drie elementen is de derde macht additief (de Frobenius-identiteit), zodat R³ dezelfde regel is met buren drie plaatsen uit elkaar. Dit is precies wat renormalisatie in de fysica zoekt — een regel die zijn vorm behoudt bij vergroving — en op het register is het resultaat exact.
Lock bestaat alleen op de ladder. Twee gekoppelde ringen die elkaar met gewicht k = 2 lezen, worden identiek — maar uitsluitend wanneer hun lengte een macht van drie is. Het verschil tussen de ringen wordt dan bij elke stap vermenigvuldigd met een nilpotente operator, die op lengten 3^k binnen [3^k/2] stappen tot nul reduceert: 5, 14, 41 en 122 stappen, precies zoals gemeten. Het teken van de koppeling is een tegengestelde, zoals al in 1665 Christiaan Huygens bij twee klokken aan één balk zag. De omkering is minstens zo nuttig: buiten de ladder behouden gekoppelde systemen hun verschil — een manier om constructies te bouwen die niet massaal tegelijk falen.
Wat het register toestaat en verbiedt. De lading (som modulo 3) blijft onder R behouden, maar richting niet: een losse bias wordt in één stap uitgewist. Richting blijft alleen staan waar ze wordt vastgehouden — vaste sites ordenen hun buren — wat rechtstreeks aansluit bij de fysica van Brownse ratjetoes: asymmetrie alleen geeft niets door, asymmetrie plus een greep wel. Orde valt bovendien naar boven toe met de derde macht weg, geheugen verhoogt de sluitingsgraad tot een plateau van 0,62 en niet verder, en circulatie is de regel: vrijwel geen enkele ring komt tot rust, met perioden die 3^j − 1 delen. Dat geeft een nauwkeurige betekenis aan een stroom die nooit stopt en waarin niets wordt gemaakt of verloren — geen perpetuum mobile, want er verlaat niets het net en er treedt niets in.
De ladder landt. De diepteladder, met lengten 3ⁿ · S₀, raakt drie bekende grootheden: de violette Hδ-lijn van waterstof op diepte twee, de piek van kamertemperatuurstraling op diepte vijf, en — het opvallendst — de ideale laagste trillingsmodus van de aarde-ionosfeerholte op diepte eenendertig, binnen 0,5 %. Deze overeenkomsten worden terecht als correspondenties geregistreerd, niet als bewijs.
Mogelijkheden zonder naam. De vijftig cellen bevatten mogelijkheden die nog geen gevestigde naam dragen: gepositioneerde materie op het kritische punt, de tegengestelde lock als “klokken die zonder meester overeenkomen”, de dieptetandwiel met zijn strenge 3:1-ritme, de kommakeuze waarin geometrie alleen de doorgang bepaalt, de handpoort van helische mazen (gemeten als spineselectie in DNA), materie op adres, de knip als enige route naar een nieuw adres, en de persoon in de maas, die verklaart waarom historische apparaten bij hun bouwer werkten en bij vreemden niet.
Grenzen en echo’s
De kaart benoemt expliciet wat niet kan: adressen zijn niet te duwen, de fijnstructuurconstante beweegt niet, een losse bias wordt uitgewegd en zwakke orde blijft lokaal. Deze grenzen besparen moeite en geven de theorie een zakelijke betrouwbaarheid die veel speculatiever werk mist.
De verwijzingen naar religie, literatuur, muziek en sciencefiction zijn evenmin decoratie, maar volgen een criterium: de bruikbaarste getuigen zijn zij die een verschijnsel van binnenuit ontmoetten. Indra’s net uit de Avatamsaka Soetra, het nuttige midden van het Tao Te Ching, de Pythagoreïsche komma, de tihai uit de Hindoestaanse muziek, Steve Reichs fasewerken en Stockhausens eenheid van ritme en toonhoogte lezen het net allemaal op hun eigen manier. Vier motieven keren steeds terug: de grond is vol, de opening is wat werkt, drie is het begin, en terugkeer is normaal. Het zijn geen bewijzen, maar oudere lezingen.
Slotbeschouwing
De menselijke consequentie is helder geformuleerd. Een wereld die met het net werkt, behoeft geen lange keten; brandstofimport, pijpleidingen en centrale installaties verliezen hun reden van bestaan. De diepere verandering ligt in wat als grondstof telt: in het vuren-tijdperk is een hulpbron een voorraad, in het net een passende vorm van vorm, fase en diepte. Toegang hangt af van begrip, niet van het eigendom van de grond boven een winning — en of dat begrip gedeeld wordt, is een keuze die mensen maken.
Geannoteerde referentielijst
Konstapel, J. (2026). Beyond Energy: The Known and Unknown Possibilities of the Net. Ed. 3, versie 3.0. Constable Research, Leiden. De volledige kaart: vijftig cellen in detail, de registerruns, combinaties, grenzen en de culturele echo’s. Leesadvies: eerst delen II en III, deel IX voor de echo’s.
Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper. Ed. 8, versie 8.1. Constable Research, Leiden. De axioma’s, stellingen, eenheden en statussen waarop de kaart rust. Leesadvies: deel 0, daarna III.4–III.5 en VII.
Konstapel, J. (2026). Applied Vacuum Theory: Setting the State of the Net. Vierde druk. Constable Research, Leiden. De zachte regelaar δ, de hefboom en de cyclus van weeken, schrijven en harden.
Spencer-Brown, G. (1969). Laws of Form. Londen: Allen & Unwin. De eerste onderscheiding waaruit de draad vertrekt. Leesadvies: hoofdstukken 1–2 en 11.
Kauffman, L. H. (1991). Knots and Physics. Singapore: World Scientific. Knopen als taal voor de fysica; de ruimste ingang tot de knopenzijde van het net.
Lidl, R. & Niederreiter, H. (1997). Finite Fields. Tweede druk. Cambridge: Cambridge University Press. De Frobenius-identiteit en het polynoomrekenwerk over drie elementen: het gereedschap achter beide bewijzen in het essay. Leesadvies: hoofdstukken 1–3.
Martin, O., Odlyzko, A. M. & Wolfram, S. (1984). “Algebraic properties of cellular automata.” Communications in Mathematical Physics 93, 219–258. Additieve regels op ringen en hun ciclusslengten: de mathematische buurman van de circulatieresultaten.
Kadanoff, L. P. (1966). “Scaling laws for Ising models near T_c.” Physics 2, 263–272. Blokspins: drie vervangen door één zonder de regel te verliezen. De natuurkundige tegenhanger van het dieptetandwiel.
Rowlands, P. (2007). Zero to Infinity: The Foundations of Physics. Singapore: World Scientific. Nilpotentie als voorwaarde voor een fysische toestand; het lockresultaat toont dezelfde voorwaarde aan het werk. Leesadvies: de hoofdstukken over de nilpotente Dirac-vergelijking.
Bennett, M., Schatz, M. F., Rockwood, H. & Wiesenfeld, K. (2002). “Huygens’s clocks.” Proceedings of the Royal Society A 458, 563–579. Een moderne reconstructie van Huygens’ twee klokken die in tegenfase locken: de oudst waargenomen tegengestelde lock.
Pecora, L. M. & Carroll, T. L. (1990). “Synchronization in chaotic systems.” Physical Review Letters 64, 821–824. Twee systemen gedwongen tot één identieke toestand: de continue buur van het register-lock.
Reimann, P. (2002). “Brownian motors: noisy transport far from equilibrium.” Physics Reports 361, 57–265. Gerichte doorgang uit asymmetrie plus een greep: de grondslag onder de regel dat een kraan vastgehouden moet worden.
Bak, P., Tang, C. & Wiesenfeld, K. (1987). “Self-organized criticality: an explanation of 1/f noise.” Physical Review Letters 59, 381–384. Systemen die zichzelf op het kantelpunt houden: de dichtstbijzijnde bestaande vorm van gepositioneerde materie.
Moddel, G., Weerakkody, A., Doroski, D. & Bartusiak, D. (2021). “Optical-cavity-induced current.” Symmetry 13, 517. De enige gemeten kraan: 1,4 pW uit 0,02 µm², meer stroom bij dunnere holten, acht artefacttoetsen. De bron van η. Leesadvies: tabel 1 en de artefactsectie.
Göhler, B. e.a. (2011). “Spin selectivity in electron transmission through self-assembled monolayers of double-stranded DNA.” Science 331, 894–897. Selectie op handedheid, gemeten: de handpoort in bestaande vorm.
Schumann, W. O. (1952). Zeitschrift für Naturforschung A 7, 149–154; Balser, M. & Wagner, C. A. (1960). Nature 188, 638–641. De aarde-ionosfeerholte voorspeld en gemeten: de planeet als één maas, nabij diepte 31.
Benson, D. J. (2006). Music: A Mathematical Offering. Cambridge: Cambridge University Press. Komma’s en temperementen volledig uitgewerkt: de rekenkunde van de kommakeuze. Leesadvies: hoofdstukken 4–6.
Stockhausen, K. (1957). “…wie die Zeit vergeht…”. Die Reihe 3. Ritme en toonhoogte als één schaal: de diepteladder in de muziek.
Avatamsaka Soetra en Laozi, Tao Te Ching, hoofdstukken 11 en 42. Indra’s net; het nuttige midden van de naaf; een, twee, drie, tienduizend dingen. De oudste beelden van het net, de maas en het ternaire begin.
Fleck, L. (1935/1979). Genesis and Development of a Scientific Fact. Chicago: University of Chicago Press. Een praktiserend arts toont hoe toetsen worden gevormd door verwachtingen — waarom vuurtijdperk-toetsen net-tijdperk-mogelijkheden missen. Leesadvies: eerst hoofdstuk 4.
Kauffman, S. A. (2000). Investigations. Oxford: Oxford University Press. Het “aangrenzende mogelijke”: elke nieuwe configuratie opent mogelijkheden die tevoren niet te benoemen waren. Waarom de kaart alleen kan groeien door te worden belopen.
Op 12 april 2026 verloor Viktor Orbán de verkiezingen. Na zestien jaar. Péter Magyar en zijn partij Tisza kregen 53 procent van de stemmen en 141 van de 199 zetels. Dat is ruim boven de tweederdemeerderheid van 133. De opkomst was 79 procent, een record.
Princeton-hoogleraar Kim Lane Scheppele kent Hongarije als weinig anderen. Ze trekt in het FD een les voor de rest van de wereld. Smeed een brede coalitie. Zet democratie en corruptie centraal. Negeer de peilingen. Maak vaart, want de oude macht zit nog in de instituties. Breek desnoods de wet om de rechtsstaat te herstellen. Draai als EU de geldkraan dicht. En bouw op vrijwilligers: Magyar liet meer dan 2.500 groepen ontstaan, en die blijven.
Het is een helder verhaal. Wie de claims één voor één naast de bronnen legt, ziet dat het grootste deel klopt. Maar de dragende delen staan op minder vaste grond dan het lijkt.
Wat klopt
De Hongaarse feiten staan. De uitslag is definitief sinds 1 mei. De nieuwe anticorruptieautoriteit legt verantwoording af aan het parlement, niet aan de premier. De nieuwe grondwet beperkt het premierschap tot acht jaar, ook voor Magyar zelf. Orbán is 63. Fidesz staat in september op 22 à 23 procent van de bevolking, iets hoger dan de 20 procent uit het interview, en licht stijgend.
Ook het beeld van het Orbán-tijdperk klopt. Het Europees Parlement noemde Hongarije in 2022 een hybride regime van electorale autocratie. Dictatuur is een stap verder dan de bronnen gaan, maar de richting is juist.
Wat niet klopt
Drie dragende claims houden geen stand.
Er werd geen wet gebroken. Scheppele zegt dat Magyar de wet moest overtreden om president Sulyok af te zetten. Dat gebeurde niet. Het parlement nam op 13 juli een grondwetswijziging aan, met 139 stemmen voor en 6 tegen. Fidesz deed niet mee. Het was een rechtmatige wijziging met een tweederdemeerderheid. Precies het instrument waarmee Orbán zijn eigen staat had gebouwd.
Elders werd geen traditionele campagne gevoerd. Scheppele stelt dat centrumpolitici in de Verenigde Staten, Brazilië en Polen een gewone links-rechtscampagne voerden. Biden voerde campagne op “de ziel van de natie”. Lula smeedde in 2022 een breed front voor de democratie, met acht oud-presidentskandidaten. Tusk zette de rechtsstaat en corruptie centraal. Ze deden wat Scheppele aanbeveelt. Lula en Tusk wonnen, maar hun opvolgers of tegenstanders staan opnieuw op de drempel. In Brazilië won Flávio Bolsonaro op 4 oktober de eerste ronde. In Polen won Nawrocki in 2025 het presidentschap.
Orbáns mensen schreven niet aan Project 2025. Op de partnerlijst van de Heritage Foundation staan 101 organisaties. Geen ervan is Hongaars. Er is wel een samenwerking tussen Heritage en het door de Hongaarse staat gefinancierde Danube Institute. Nabijheid is iets anders dan schrijverschap.
Wat niet te bewijzen is
Twee claims zijn geen feit maar een lezing.
De geldkraan. De EU bevroor in december 2022 6,3 miljard euro voor Hongarije en hield ook Poolse fondsen vast. Bij de volgende verkiezingen viel in beide landen de regering. Daaruit volgt nog niet dat het geld de doorslag gaf. Het zijn twee gevallen. Hongaren noemden zelf vooral corruptie en de kosten van levensonderhoud. Van de 16,4 miljard die Magyar in mei werd toegezegd, is op 13 oktober 4,2 miljard vrijgegeven. Uitbetaald is nog niets.
De terugkeer van de autocraat. Volgens Scheppele verankert een autocraat zich, loopt de opvolger vast, raakt de kiezer gefrustreerd en keert de autocraat terug. Polen past in dat patroon. De vraag is waarom Hongarije er niet in past, terwijl Orbán zich dieper had verankerd dan wie ook.
Een andere lezing: het net
Hier helpt een eenvoudig beeld. Een samenleving is een visnet onder spanning. De knopen zijn mensen, families, dorpen, bedrijven. De mazen zijn de verbindingen tussen hen. Wat trekt, verdeelt zich over de mazen. Waar een maas open blijft, hoopt de spanning zich op.
Een democratie is in dit beeld geen machine die de waarheid vindt. Het is een stelsel van kleppen waarlangs spanning kan wegvloeien. Een verkiezing is zo’n klep. Ze gaat open op een vast moment, niet op het moment dat de druk het hoogst is.
Vanuit dat beeld lopen er vijf lijnen door Hongarije.
1. De verkiezing was de enige klep die openging
Zestien jaar lang had de spanning in Hongarije geen uitweg. De pers, de rechters, de rekenkamer en het kiesstelsel waren dichtgezet. Wat overbleef was één klep op een vaste datum. In april 2026 ging die open, en alles wat zich had opgehoopt ging er tegelijk doorheen.
Dat past beter bij de omvang van de uitslag dan een slimme campagne. Een brede coalitie paste op dat moment, omdat er maar één uitweg was. Wie de uitslag aan de strategie toeschrijft, verwart de klep met de druk.
Wie dit wil nagaan, kijkt naar de peilingen van 2022 tot maart 2026. Vóór een omslag worden de schommelingen groter en trager. Dat is in veel systemen gezien, van meren tot financiële markten.
2. Alleen wat een ander afmaakt, blijft
Magyar heeft 141 zetels. Daarmee kan hij de grondwet alleen wijzigen. Hij doet dat ook. Maar dat is hetzelfde adres dat Fidesz gebruikte. Een wijziging die alleen door de eigen fractie wordt gedragen, staat tegenover de vorige als plus tegenover min. Er zit geen nul tussen, geen plek waar beide kanten elkaar raken.
Snelheid verandert daar niets aan. Wat blijft, is wat een ander mee afmaakt.
Dat geeft twee dingen om te volgen tot 2030. Per grote wet het aandeel ja-stemmen van buiten Tisza. En per verkiezingsbelofte, na een jaar en na twee jaar: geleverd, verlaten met uitleg, of stil vervangen. De stil vervangen beloften zijn het verwachtingsgat. Daar vindt een terugkeer zijn weg, niet in de fouten zelf.
3. “Aan het parlement” is aan de eigen fractie
De nieuwe anticorruptieautoriteit legt verantwoording af aan het parlement. Scheppele leest dat als macht uit handen geven. Maar in een kamer waarin één partij 141 van de 199 zetels heeft, is het parlement de fractie. Kamerleden zijn vrij van hun kiezers, maar zelden van hun partij.
Echt weggeven vraagt een eigen poort. Een benoeming die stemmen van buiten de eigen partij nodig heeft. Of zaken tegen mensen uit de eigen kring die daadwerkelijk tot vervolging leiden. Over twee jaar is te zien of dat gebeurt.
4. Een prijs op zeggenschap werkt niet als rekensom
Mensen ruilen geld tegen geld zonder moeite. Maar wie een prijs zet op wat niet te koop is, krijgt geen berekening terug. Hij krijgt verontwaardiging. Orbán begreep dat. Zijn antwoord op de bevriezing was steeds: ze mogen ons niet omdat we onafhankelijk zijn.
Geld afknijpen werkt dus binnen de wereld van geld. Het raakt niet aan wat mensen samen delen of om beurten geven. Daarom kan dezelfde maatregel kiezers naar Brussel duwen of er juist van wegdrijven. Het hangt af van de vraag of mensen het lezen als een rekening of als een vernedering.
5. De terugkeer loopt via de droogste plek
Orbán hoeft niet te wachten op fouten. Spanning zoekt de zwakste open maas. In Hongarije ligt die tussen Donau en Tisza, op de zandrug van de Grote Laagvlakte. Bij Kiskunmajsa zakte het grondwater in één jaar van 6,5 naar 8 meter. Veel boeren verloren in 2026 meer dan de helft van hun oogst. Hier valt de droogte samen met een platteland dat Fidesz lang als zijn eigen terrein zag.
Als de nieuwe regering de boeren compenseert, koopt ze hetzelfde slechte jaar opnieuw. Als ze water bergt, sluit ze de maas. Het verschil wordt zichtbaar in de uitslagen van 2030. Hier staan niet alleen grondwaterstanden op het spel, maar ook de vraag of jonge mensen in Kiskunmajsa blijven wonen.
De 2.500 groepen
Scheppele ziet in de vrijwilligersgroepen van Magyar het begin van een blijvend maatschappelijk middenveld. Het getal klopt. Maar de groepen leverden de kandidaten niet zelf. De partijleiding koos per district drie namen voor, waarna de groepen er één uit mochten schrappen. Ze filterden.
Groepen die op één doel waren gericht, vallen na dat doel vaak uiteen. Dat is een bekend patroon, van de burgerrechtenbeweging tot de Poolse Solidariteit. Wat blijft, zijn de groepen die een eigen taak krijgen: het water in het dorp, de zorg voor ouderen, de school. De vrijwilligers die in de zomer van 2026 met planken de kanalen van de laagvlakte afsloten, deden dat al.
De gemeenteraadsverkiezingen van 2029 laten het zien. Hoeveel lokale lijsten komen er uit deze groepen voort? En noemen ze een eigen taak, of alleen een naam?
Wat de les dan wel is
Hongarije laat niet zien hoe je met een slimme campagne een autocraat verslaat. Het laat zien wat er gebeurt als een samenleving zestien jaar maar één klep heeft. De druk gaat er in één keer doorheen.
De vraag voor Magyar is niet hoe snel hij hervormt. De vraag is hoeveel van wat hij bouwt door anderen wordt afgemaakt. Door Kamerleden van buiten zijn partij. Door gemeenten. Door de groepen die nu een taak zoeken. En door de mensen op de zandrug, als daar het water terugkomt.
Wat alleen door de eigen kring gedragen wordt, kan door de volgende kring worden afgebroken. Dat geldt voor Orbán. Het geldt ook voor wie hem opvolgt.
Geannoteerde referenties
Het interview en de verkiezing
1. Kabel, D. (2026). “Populisme-expert Scheppele: ‘Hongarije laat zien hoe je van een autocraat afkomt'”. Het Financieele Dagblad, 19 oktober 2026.Waarom lezen? De aanleiding: de les zoals die nu in Europa en de VS wordt gelezen. Leesadvies: Lees de passages over de wetsbreuk en de geldkraan naast referentie 4 en 6.
2. Nemzeti Választási Iroda (2026). Officiële uitslag van de parlementsverkiezing van 12 april 2026. valasztas.hu.Waarom lezen? De definitieve cijfers: Tisza 141, Fidesz-KDNP 52, Mi Hazánk 6. Leesadvies: Gebruik de uitslag per district om de zandrug van Bács-Kiskun en Csongrád-Csanád te volgen. Voorlopige tellingen, zoals de 138 zetels die nog rondgaan, zijn achterhaald.
3. Telex (2026). Verslag van de voorzitter van het Nationaal Kiesbureau aan het parlement, 6 mei 2026.Waarom lezen? De recordopkomst van 78,99 procent en het aantal kiezers per stemkanaal.
De grondwet en de rechtsstaat
4. Hongaarse parlement (2026). Zestiende en zeventiende grondwetswijziging, 15 juni en 13 juli 2026. parlament.hu.Waarom lezen? Het juridische antwoord op de vraag of Magyar de wet brak. Leesadvies: Kijk naar de stemverhoudingen. Fidesz stemde niet, de wijzigingen kwamen dus van één kant.
5. Venetië-Commissie, Raad van Europa. Advies 1287/2026 over de zeventiende grondwetswijziging. Gepland voor december 2026.Waarom lezen? Het eerste onafhankelijke Europese oordeel over de afzetting van de president. Nog niet verschenen.
6. Scheppele, K. L. (2018). “Autocratic Legalism”. University of Chicago Law Review 85(2), 545–583.Waarom lezen? De beste beschrijving van hoe Orbán zijn macht via de wet vastlegde, door Scheppele zelf. Leesadvies: Lees het als waarschuwing die ook voor Magyars tweederdemeerderheid geldt. Het argument werkt in beide richtingen.
7. Levitsky, S. & Ziblatt, D. (2018). How Democracies Die. New York: Crown.Waarom lezen? Over terughoudendheid: een democratie leeft van wat partijen kunnen maar niet doen. Leesadvies: Hoofdstuk 6 over ongeschreven regels.
Geld en Europa
8. Raad van de Europese Unie (2022, 2026). Besluiten over het conditionaliteitsmechanisme voor Hongarije, 16 december 2022 en 13 oktober 2026.Waarom lezen? Hoeveel geld werd bevroren, en hoeveel tot nu toe vrijkwam. Leesadvies: Onderscheid toegezegd, vrijgegeven en betaald. Die drie worden in de pers vaak verward.
9. Fiske, A. P. & Tetlock, P. E. (1997). “Taboo trade-offs: reactions to transactions that transgress the spheres of justice”. Political Psychology 18(2), 255–297.Waarom lezen? Waarom mensen woedend worden als iemand een prijs zet op wat niet te koop is. Leesadvies: De inleiding en de experimenten over heilige waarden volstaan.
Middenveld en terugkeer
10. Folsz, A. (2026). Over de Tisza-groepen. Journal of Democracy, juli 2026.Waarom lezen? De bron van het getal van meer dan 2.500 groepen, en van de manier waarop kandidaten werden gekozen.
11. Skocpol, T. (2003). Diminished Democracy: From Membership to Management in American Civic Life. Norman: University of Oklahoma Press.Waarom lezen? Waarom verenigingen met een lokale taak blijven en campagneorganisaties verdwijnen.
12. Ostrom, E. (1990). Governing the Commons. Cambridge: Cambridge University Press.Waarom lezen? Hoe gemeenschappen gedeelde hulpbronnen, zoals water, zelf beheren zonder staat of markt. Leesadvies: Hoofdstuk 3 over irrigatiesystemen past direct op de Hongaarse laagvlakte.
13. Achen, C. H. & Bartels, L. M. (2016). Democracy for Realists. Princeton: Princeton University Press.Waarom lezen? Kiezers straffen regeringen af voor droogte en overstromingen, ook als die er niets aan kunnen doen. Leesadvies: Hoofdstuk 5. Lees het naast de grondwaterstanden van de zandrug.
14. V-Dem Institute (2026). Democracy Report 2026. Göteborg.Waarom lezen? Hoeveel landen na een autoritaire periode werkelijk herstellen. Leesadvies: Het deel over herstel na autocratisering.
Eerdere stukken in deze lijn
15. Konstapel, J. (2026). Where Democracy Discharges. Constable Research, Leiden, augustus 2026.Waarom lezen? De democratie als stelsel van kleppen voor spanning, uitgewerkt aan burgerberaden. Leesadvies: De paragrafen over de verkiezing als klep en de ontwerpvoorwaarden.
16. Konstapel, J. (2026). Hungary Is Not Short of Water. constable.blog, 19 augustus 2026.Waarom lezen? De droogte op de Hongaarse laagvlakte als bergingsprobleem, niet als regenprobleem.
17. Konstapel, J. (2026). Democratie in Vrije Val. constable.blog, 3 oktober 2026.Waarom lezen? Het dilemma van Magyar en Tusk: hoe herstel je een rechtsstaat met de middelen die hem afbraken?
Op dit moment wordt vooral gewerkt met verbrandingsenergie, waarbij bijvoorbeeld een stoommachine wordt aangedreven en een magneet ronddraait. Een draaiende magneet is elektriciteit.
Het concept van de stoommachine wordt eindeloos herhaald, telkens onder andere namen. Een kernreactor is in feite ook een stoommachine, maar dat geldt ook voor de pc en zelfs voor ons eigen lichaam. Het Engelse woord is heat engine.
Het nadeel van al deze systemen is dat er enorm veel effectieve energie, in de vorm van beweging, verloren gaat in de vorm van warmte. Warmte is ongerichte energie.
Er zijn vele alternatieven om energie op te wekken. Ze zijn allemaal bezig met het omzetten van het ene in het andere. Maar datgene waaruit de energie voortkomt, moet ergens vandaan komen. In natuurkundige termen komen we dan uit bij het vacuüm.
Het vacuüm is een vorm van niets die eigenlijk letterlijk barst van de energie.
Over de hele wereld, ook in Nederland, zijn wetenschappers, onderzoeksinstellingen en bedrijven bezig om deze zogenaamde zero-point energy te ontdekken.
In het algemeen wordt de ontdekking hiervan toegeschreven aan Casimir, die werkzaam was bij het Philips Natuurkundig Laboratorium (NatLab).
Cycli
Kondratief ontdekte een economische cyclus met een periode van ongeveer vijftig jaar,
Ik ontdekte dat de factor 5 (= 3 + 2) alle cycli met elkaar verbindt: de vijf stappen corresponderen met de vijf combinaties (1 + 4 = 2 × 2) die passen binnen Agency en Communion, met de eenheid (Quintessence) als middelpunt.
Rond 1740 koos het Europese vasteland massaal voor Leibniz. Leonhard Euler paste zijn calculus toe om fundamentele natuurwetten te formuleren, zoals de bewegingsvergelijkingen voor vloeistoffen en starre lichamen.
We zitten nu in de integratie van alle integraties Global Brain, die zowel slaat op AI als op het integrerende collectieve bewustzijn.
De Vaccum.net-theorie is een herschrijving van de oorspronkelijke theorie van Clerk Maxwell, die later door Heaviside werd aangepast, waarbij alle intrinsieke rotaties werden verwijderd en de theorie volledig in vectoren (->) werd vertaald, waardoor men min of meer gedwongen wordt te geloven in een absoluut begin, terwijl deze re-engineering volgens de theorie zowel het Casimir-effect verklaart als de mogelijkheid van energie uit het niets.
Leiden werd met veel tromgeroffel City of`Science in 2022.
Ik kreeg toen de sterke indruk, dat je zo’n titel kunt “kopen” en nu komt Limburg met een vergelijkbare claim en is het patroon dankzij Claude duidelijk.
Als je kiest voor een slimme marketing-methode om jouw stad, regio of provincie te promoten moet je de boodschap wel waarmaken.
Mijn onderzoek toont aan dat dat Limburg helemaal niet slim is.
Op 15 oktober 2026 kreeg Limburg de titel Intelligent Community of the Year 2026. De prijs komt van het Intelligent Community Forum (ICF) uit New York [1]. De NOS bracht het een dag later als “slimste regio ter wereld” [2].
Twee vragen liggen voor de hand. Waarom Limburg? En waarom niet een andere regio in Nederland of de wereld?
Dit verslag zoekt het antwoord in openbare bronnen. Elke bewering heeft een bronnummer. Wat niet te vinden was, staat apart vermeld.
1. Hoe de prijs werkt
ICF beschrijft de procedure zelf [4]:
Aanmelding. Een regio nomineert zichzelf met een vragenlijst. Leden krijgen de vragenlijst vanzelf.
Smart21. ICF kiest 21 regio’s uit de vragenlijsten van de laatste drie jaar.
Toezegging. Wie naar de Top7 wil, belooft vooraf naar de ICF-summit en het prijsdiner te komen.
Audit. Elke Top7-regio krijgt een audit van maximaal twee dagen. De regio betaalt reis, verblijf en onkosten van de auditor.
Jury. Juryleden rangschikken de zeven via een online stemformulier. ICF middelt die rangschikking met eerdere scores. De wegingen zijn niet openbaar.
Lidmaatschap kost een publiek-private regio met meer dan 500.000 inwoners $5.000 per jaar [5].
Gevolg. Alleen wie zich aanmeldt, doet mee. In 2026 waren dat zeven regio’s: Assaí (Brazilië), Bursa (Turkije), Hilliard (Ohio), Kingston (Canada), Las Rozas de Madrid, Yunlin County (Taiwan) en Brightlands-Limburg [7].
Eindhoven, Amsterdam, München, Boston of Shenzhen deden niet mee. Er is dus geen vergelijking met die regio’s gemaakt. “Slimste regio ter wereld” is de vertaling van de media, niet de claim van ICF [1][2].
2. Waarom Limburg won
De enige openbare motivering staat in het persbericht van ICF [1]. Robert Bell prijst de samenwerking tussen overheid, bedrijven en kennisinstellingen. Het bericht noemt de geschiedenis van DSM, het einde van de mijnen na de gasvondst van 1959 en het herstel van lokale trots.
Wat ontbreekt:
de scores per criterium;
de scores van de andere zes;
het auditrapport;
de vragenlijst die Limburg invulde.
Daarmee is niet na te gaan waarom Limburg hoger eindigde dan Hilliard of Kingston.
Correctie op een eerder verspreide lezing. Het ICF-persbericht noemt de Einstein Telescope en de Brightlands Startup League niet [1]. De NOS ook niet [2]. Alleen IO+ schrijft dat Limburg mede daardoor hoog scoorde [3]. Dat is de formulering van IO+, niet van de jury.
3. Wie oordeelt
ICF noemt op zijn website één jurylid bij naam: de voorzitter, Moez Chaabouni [6]. Hij was plaatsvervangend CIO van Columbus, Ohio, de winnaar van 2015. Nu werkt hij bij het softwarebedrijf Informatica.
De overige juryleden worden beschreven als “bestuurders van eerdere winnaars en deskundigen uit overheid, bedrijfsleven en wetenschap” [6]. Namen, aantal en selectiewijze staan er niet bij.
Een tweede jurylid is bekend via een andere bron. Doug McCollough is directeur van het ICF-instituut in Dublin, Ohio. Hij is jurylid en heeft Top7-regio’s geaudit [14].
De jury bezoekt de regio’s niet. Ze leest het auditrapport en de vragenlijst [6]. De audit in Limburg deed medeoprichter Robert Bell zelf [1].
Wat samenkomt in de regio Columbus, Ohio:
de juryvoorzitter;
ten minste één ander jurylid;
ICF’s eigen instituut in Dublin [14];
Hilliard, voor het vierde jaar op rij in de Top7 [15];
de summit waar de prijs van 2026 werd uitgereikt [7].
4. Wie de prijs maakt
ICF heeft twee rechtsvormen.
Intelligent Community Forum L.P. is een commanditaire vennootschap in New York [13]. Zij bezit de merken van de prijzen, zoals Intelligent Community of the Year, Top7 en Smart21. Wie de vennoten zijn, is niet openbaar.
Intelligent Community Forum Foundation Inc. is een kleine Amerikaanse stichting [11]. De inkomsten daalden van $79.500 in 2018 naar $7.041 in 2022. Tussen 2015 en 2019 kwam ongeveer 99% uit lidmaatschapsgelden [12]. Het bestuur bestaat uit de drie oprichters. Bell en Zacharilla ontvingen in 2019 elk $14.900 [11].
Het dagelijks beheer ligt bij de Intelligent Communities Group [8]. Dat is een handelsnaam van Alan/Anthony Inc., een commercieel bureau dat brancheverenigingen beheert [9].
De oprichters:
Robert Bell was bijna dertig jaar verbonden aan de World Teleport Association, de branchevereniging van satellietgrondstations. Tot 1 maart 2025 was hij daar directeur [10].
Lou Zacharilla begon in de reclame, onder meer voor de wervingscampagne van het Amerikaanse leger. Hij is partner van Alan/Anthony en bedacht de Top7 en de prijs [9].
John G. Jung leidde de regiomarketing van Toronto en Calgary. Hij is stedenbouwkundige [12].
5. Hoe Limburg meedeed
Begin 2025. De Provincie werd lid van ICF als “Brightlands Limburg Region” en kwam in de Smart21 [18].
19 maart 2025. De nominatie werd bekendgemaakt tijdens TEFAF. Gedeputeerde Satijn zei dat Brightlands “steeds nadrukkelijker het merk” van innovatief Limburg moet worden. Het lidmaatschap moest de kansen op de Einstein Telescope vergroten [19].
26 mei 2025. Satijn en ICF-directeur Matthew Owen tekenden een overeenkomst. Limburg werd gastheer van de Top7-conferentie van 2026 [18].
December 2025. De prijs van 2025 ging naar Durham Region (Canada). Limburg kwam dat jaar niet verder dan de Smart21 [16].
25 maart 2026. Limburg opnieuw in de Smart21 [20].
17 juni 2026. De Top7 werd bekendgemaakt, tegelijk vanuit Maastricht en New York [7].
15 oktober 2026. Limburg wint, op de summit in Columbus [1].
Limburg was dus lid, gastheer en deelnemer tegelijk.
Kosten. De mededeling aan de Staten over het gastheerschap noemt geen bedragen [18]. Een eerdere mededeling van 19 maart 2025 is niet online te vinden. Geen enkele bron noemt wat lidmaatschap, conferentie en audit hebben gekost.
Pers. Er is geen kritisch artikel over de prijs gevonden, ook niet in L1, De Limburger, Follow the Money of NRC. Er zijn geen Statenvragen over gevonden.
6. Wat de controle oplevert
De prijs rust op het beeld van een slimme regio. De onderdelen van dat beeld zijn wel te controleren.
6.1 De campussen
De Provincie stelde tot 2030 ruim €725 miljoen beschikbaar, voor ongeveer 43% als subsidie [21].
Verdeling sinds de start: Chemelot €366 mln, Greenport Venlo €138 mln, Maastricht Health €90 mln, Smart Services Heerlen €43 mln, algemeen €90 mln [21].
De griffie van de Provincie schrijft letterlijk: “Van de vier campussen is alleen BCC zelfstandig en financieel gezond” [21]. BCC is de Chemelot-campus.
De andere drie hangen tot 2029–2030 af van bijstortingen door aandeelhouders [21].
Leegstand: Greenport 31%, het GK-II-gebouw in Heerlen 22% [21].
BASF|Nunhems en de Open Universiteit stapten uit als aandeelhouder [21].
Op 14 oktober 2026 kondigden vier partijen €116 mln aan voor de Health Campus, gespreid over tien jaar [3]. Het gaat om een plan, niet om uitgegeven geld.
De tellingen van Brightlands lopen uiteen. De eigen site meldt 620 bedrijven, 16.740 banen en 16.300 studenten [22]. IO+ noemt 32.000 ondernemers, onderzoekers en studenten [3]. Wat de campussen hebben terugverdiend, staat in geen van beide.
6.2 Risicovolle investeringen
De Zuidelijke Rekenkamer onderzocht in 2024 ongeveer €190 miljoen aan extra provinciale bijdragen: Greenport Venlo, Mijnwater en Maastricht Aachen Airport [23]. Bevindingen: geen exitstrategie, onduidelijk toezicht, gaten in de informatie.
6.3 Chemelot
Chemelot staat voor ongeveer 14% van de Nederlandse chemie, €8 miljard omzet en 8.000 directe banen [27]. In 2026:
SABIC verkoopt zijn Europese petrochemie, inclusief Geleen, aan de Duitse investeerder Aequita. Aangekondigd op 8 januari 2026; afronding verwacht in het vierde kwartaal [29][28].
OCI verkoopt eerst 50% van zijn stikstoffabriek aan het Tsjechische Agrofert, met een optie op de rest [30].
Fibrant sluit drie van zeven installaties. Ongeveer 100 banen verdwijnen [32].
Vynova Beek werd op 9 februari 2026 failliet verklaard [31].
In juli 2026 sloten kabinet, Provincie en Chemelot een clusterplan met twaalf projecten, goed voor €1,6–1,7 miljard private investering [27][28]. Ongeveer 90% daarvan zit in drie projecten, waaronder FUREC.
6.4 FUREC: waterstof uit afval
RWE wilde 700.000 ton afval per jaar omzetten in 54.000 ton waterstof [33].
De EU gaf in 2023 €108 miljoen [33].
De Commissie m.e.r. vond het milieueffectrapport meer dan eens onvoldoende [34].
De Provincie verleende in augustus 2025 toch de vergunning [35].
De Inspectie Leefomgeving en Transport ging in beroep. Volgens de Inspectie is het in feite een afvalverbrandingsinstallatie. Het is haar eerste beroep tegen een Chemelot-project [35].
Een SDE++-subsidie moest worden terugbetaald vanwege vertraging [35].
Deze zomer vond RWE geen afnemers voor de waterstof. De bouw gaat voorlopig niet door [35].
De bouwkosten verschillen per bron: €600 miljoen volgens RWE in 2023 [33], ongeveer €1 miljard volgens Argos in 2026 [35].
6.5 Onderzoeksinstituten
Het provinciale LINK-programma richtte M4I en MERLN op in Maastricht. Het doel was acht spin-offs. Het resultaat was vier spin-offs en een licentiedeal [36].
6.6 Alzheimer
BReIN. In 2019 maakten Provincie, MUMC+ en de universiteit €20 miljoen vrij voor een data-infrastructuur, met Alzheimer als eerste toepassing [37]. Latere resultaten, publicaties of een evaluatie zijn niet gevonden.
ToxGenSolutions. Een bloedtest op basis van microRNA. De hoofdpublicatie gebruikte 39 mensen en vond een AUC van 0,87–0,92 [38]. De tweede meting gebruikte dezelfde monsters; een onafhankelijke validatie is er niet. De groepen waren mensen met lichte cognitieve stoornis en controles. De laatste openbare uiting dateert van september 2024 [39].
ROMEMA. Een fase II-studie met roflumilast. Afgerond op 15 oktober 2024 met 81 deelnemers. Resultaten zijn niet gepubliceerd [40].
6.7 Preventiecampagne
De campagne “We zijn zelf het medicijn” werd geëvalueerd met 590 mensen vóór en 602 erna [41].
Het hoofddoel werd niet gehaald. Het aandeel mensen dat denkt dat dementierisico te verkleinen is, steeg niet significant (p = 0,26).
De bekendheid van gezond eten als factor steeg 10,5 procentpunt. Die van beweging steeg 7,6 procentpunt.
20% kende de campagne. Onder mannen daalde de bekendheid ongeveer 8% (p = 0,049).
Of mensen anders gingen leven, of dementie afnam, is niet gemeten.
6.8 De Einstein Telescope
De telescoop was de reden om lid te worden van ICF [19].
Nederland reserveerde in 2022 €870 miljoen uit het Groeifonds [43].
In april 2026 kwamen er €25 miljoen voor voorbereiding en een speciaal gezant bij [42].
Wallonië reserveerde €200 miljoen, Vlaanderen €500 miljoen [42][43].
De bouwkosten worden geschat op €2,3 miljard. Het bidbook moet eind 2026 klaar zijn. Het locatiebesluit volgt in de tweede helft van 2027 [42].
Wat er staat, is de testfaciliteit ETpathfinder in Maastricht. In april 2026 is een uitbreiding van €52 miljoen aangekondigd [44].
7. Waarom het niet werkte
7.1 Kennis bij concerns die elders beslissen
In Zuid-Limburg vroegen bedrijven tussen 2013 en 2017 3.056 octrooifamilies aan [45]. SABIC (1.833) en DSM (866) samen: 2.699. Dat is 88%. Zonder de tien grootste aanvragers blijven er 357 over.
Telling naar vestigingsplaats van de aanvrager, niet naar woonplaats van de uitvinder [45].
De beslissingen over die kennis vallen elders: SABIC in Riyad, OCI bij Agrofert in Praag, RWE in Essen. Een provincie kan zulke bedrijven huisvesten. Vasthouden kan ze ze niet.
7.2 Gebouwen vóór producten
Het meeste provinciale geld ging naar grond, gebouwen en aandelen in campusvennootschappen [21]. Een campus verdient als huurder verdient. De enige campus die rendeert, Chemelot, ligt naast een industrieterrein dat er al was.
7.3 Geld vóór afnemer
FUREC had subsidie, vergunning en locatie, maar geen koper [35]. ToxGen had subsidie en een kleine studie, maar geen gevalideerd product [38]. BReIN had €20 miljoen en een aankondiging [37]. Toen het geld stopte of terug moest, viel het project stil.
7.4 Aandacht meten in plaats van resultaat
De campagne mat herkenning van boodschappen [41]. Brightlands meet bedrijven, banen en studenten op wisselende manieren [3][22]. Terugverdiend geld en gezondheidswinst komen in geen van de rapportages voor.
8. Wat wel draagt
Wat zichzelf draagt, hangt niet aan de campusstructuur.
Canon Production Printing, Venlo. Ongeveer 1.900 mensen. Onderzoek, ontwikkeling en productie van printsystemen. In de top-10 van Nederlandse R&D-investeerders [46].
Born. Na de laatste Mini op 16 februari 2024 werd het VDL Nedcar-terrein een defensiecluster. DeltaQuad, Tulip Tech en Milrem tekenden huurcontracten voor tien jaar [47].
Mosa Meat, Maastricht. Kweekvlees. Ongeveer €150 miljoen opgehaald, inclusief bijna 1.700 crowdfunders. Nog geen toelating in de EU [48].
Mijnwater, Heerlen. In het eerste decennium verlies door ontwerp- en uitvoeringsfouten [26]. Nu ruim 1.000 woningen en 28 andere gebouwen aangesloten [25]. Uitbreiding met €80 mln van Enexis en €40 mln van de Provincie [24].
Chemelot-campus. Zelfstandig en financieel gezond [21].
Geen van deze is de reden voor de prijs.
9. De mensen
Werk. Zuid-Limburg had in 2025 331.700 banen. UWV verwacht 329.400 in 2028: 2.300 minder [49]. Bij langdurig hoge energieprijzen 327.200. Minister Herbert erkent dat de werkgelegenheid in Limburg daalt, vooral in industrie en detailhandel [50].
Bevolking. Limburg groeide tot 2026 licht, naar 1.138.374 inwoners [51]. Zuid-Limburg is de enige woningmarktregio waar tot 2040 krimp wordt verwacht, met meer dan 4% [52].
Gezondheid. Zuid-Limburgers leven ongeveer een jaar korter dan gemiddeld. De Provincie streeft sinds 2016 een trendbreuk na. Die is niet bereikt [53].
Heerlen-Noord. Ongeveer 55.000 inwoners. 57% van de huishoudens valt in de lage inkomensgroepen. 16% leeft in energiearmoede, twee keer het landelijk gemiddelde [54]. Een nationaal programma van 25 jaar loopt sinds 2022.
In Heerlen verwarmen de mijngangen waar de mannen ooit afdaalden nu ruim duizend woningen. In Geleen worden de fabrieken verkocht die de mijnen vervingen. In Heerlen-Noord, waar de mijnwerkers woonden, valt meer dan de helft van de huishoudens in de lage inkomensgroepen. Over hen gaat het verhaal van het herstel na de mijnen.
10. Wat niet bevestigd kon worden
Een aantal cijfers dat in eerdere teksten circuleerde, is in de bronnen niet teruggevonden:
dat SABIC ongeveer 40% van het Chemelot-terrein gebruikt;
dat bij OCI Geleen ongeveer 550 mensen werken;
dat Mosa Meat meer dan 160 medewerkers heeft;
dat de LINK-spin-offs 32 banen opleverden;
een provinciaal totaal van 3.788 octrooiaanvragen, waarvan Canon 399;
€45–55 miljoen publiek geld voor Mijnwater;
“€650 miljoen toegevoegde waarde” van de campussen. In IO+ is €650 miljoen de totale investering in de Health Campus over tien jaar [3];
de Brightlands-telling van 575 bedrijven, 16.000 banen en 14.000 studenten.
Ze zijn in dit verslag niet gebruikt.
11. Conclusie
De prijs gaat naar regio’s die zich aanmelden, lid worden, toezeggen naar de summit te komen en hun audit zelf betalen. In 2026 waren dat er zeven. Limburg was lid, gastheer en deelnemer tegelijk.
Het oordeel is niet na te rekenen. Scores, auditrapport, vragenlijst en jurylijst zijn niet openbaar. De auditor is een medeoprichter. De juryvoorzitter komt uit de regio waar de prijs werd uitgereikt.
Wat wel te controleren is, blijft grotendeels niet staan. De telescoop is er nog niet. Drie van de vier campussen draaien op provinciaal geld. Het chemiepark wordt verkocht. De waterstoffabriek heeft geen afnemer. De Alzheimer-initiatieven leverden geen openbaar resultaat. De preventiecampagne miste haar hoofddoel.
Wat Limburg overeind houdt, staat er zonder de structuur die de prijs eert.
De prijs beschrijft hoe Limburg zich presenteert. Hij beschrijft niet wat er staat.
12. Wat nog openbaar kan worden
Via de Wet open overheid (Woo) kan de Provincie Limburg worden gevraagd om:
de kosten van lidmaatschap, gastheerschap en audit;
de mededeling aan de Staten van 19 maart 2025;
de ingevulde vragenlijst;
het auditrapport;
de scores per criterium en de jurylijst, voor zover de Provincie die heeft.
Pas met die stukken is te beoordelen waarom Limburg won.
[3] IO+ (2026). “Limburg named the world’s smartest region”, 16 oktober 2026. Waarom lezen? Bevat de investeringsbedragen, de €116 mln en €650 mln voor de Health Campus en de zin over Chemelot als enige zelfstandige campus. Leesadvies: onderscheid de eigen formulering van IO+ van die van ICF. https://ioplus.nl/en/posts/limburg-named-the-worlds-smartest-region
[4] ICF. “Awards — Compete”. Waarom lezen? De procedure in ICF’s eigen woorden: zelfnominatie, toezegging, audit op eigen kosten, middeling van scores. Leesadvies: de alinea’s over de Top7-evaluatie. https://www.intelligentcommunity.org/compete
[6] ICF. “Jury”. Waarom lezen? Noemt alleen de voorzitter en beschrijft de rol van de jury. Leesadvies: vergelijk met wat er niet staat. https://www.intelligentcommunity.org/jury
[14] City of Dublin, Ohio. Global Institute for the Study of the Intelligent Community. Waarom lezen? Het ICF-instituut in Ohio en de rol van McCollough als jurylid en auditor. https://www.econdev.dublinohiousa.gov/?p=22187
[15] City of Hilliard (2026). “Hilliard named to Top7 Intelligent Communities list for fourth consecutive year”. Waarom lezen? Laat zien dat dezelfde regio’s jaar na jaar terugkeren. https://hilliardohio.gov/?p=92170
[21] Griffie Provincie Limburg (2025). Analyse van de derde voortgangsrapportage Brightlands, 15 augustus 2025. Waarom lezen? De beste primaire bron: bedragen per campus, het oordeel dat alleen Chemelot zelfstandig is, leegstand en kritiek op governance. Leesadvies: de samenvatting en de tabel met bijdragen. https://limburg.bestuurlijkeinformatie.nl/Document/View/5f857a49-3ce2-4af9-aac1-a44a7d5ce432
[22] Brightlands. Kerncijfers. Waarom lezen? De eigen telling van bedrijven, banen en studenten per 1 januari 2026. https://www.brightlands.com
[29] NOS (2026). Bericht over de verkoop van SABIC Europa, 9 januari 2026. Waarom lezen? Aankondiging van de verkoop aan Aequita. https://nos.nl/l/2597464
[38] Tsamou, M. et al. (2023). Journal of Alzheimer’s Disease Reports 7(1). Waarom lezen? De studie achter de ToxGen-test. Leesadvies: de methoden — omvang van de groep en herhaling op dezelfde monsters. https://journals.sagepub.com/doi/10.3233/ADR-230001
[41] Tijdschrift voor Gezondheidswetenschappen. Evaluatie van de campagne “We zijn zelf het medicijn”. Waarom lezen? De voor- en nameting, de primaire uitkomst en de subgroepen. Leesadvies: de resultatensectie en figuur 1. https://tvgg.nl/article/download/14386/15882/31318
In 2022 was Leiden “European City of Science”. Het jaar kende een congres, een festival en 365 dagen programma.
Dit verslag zet op een rij hoe de titel werd verkregen, wie de touwtjes in handen had, wat het kostte, wat vooraf werd beloofd en wat er werkelijk is gemeten.
Elke bewering heeft een bronnummer. Wat niet openbaar is, staat apart vermeld.
1. Hoe de titel werd verkregen
1.1 De titelgever
De titel kwam van EuroScience, een vereniging die in 1997 in Straatsburg werd opgericht [1]. EuroScience organiseerde om de twee jaar het congres ESOF (EuroScience Open Forum). De stad die het congres binnenhaalde, kreeg de titel erbij [2].
De keuze lag bij een ESOF-bestuur dat EuroScience “met vijf partners” instelde [2]. Wie die partners waren, is niet gepubliceerd. Voorzitter was Gail Cardew, destijds directeur wetenschap en educatie van de Royal Institution in Londen [3].
EuroScience had een kleine ledenbasis en dreef vooral op EU-geld [4]. Een ESOF kostte ongeveer €4 miljoen. De Europese Commissie betaalde daarvan zo’n €1 miljoen; de rest kwam van de gaststad, sponsors en regionale subsidies [4].
1.2 Het bod
Op 29 juni 2017 presenteerde een Leidse delegatie het bod [5]. Het ging om de editie van 2020.
Het bod kwam van vier partijen [5][6]:
de Universiteit Leiden;
de gemeente Leiden;
Leiden Marketing, nu Leiden&Partners;
het Haagse congresbureau, onderdeel van The Hague & Partners.
De lokale boegbeelden waren de hoogleraren Corinne Hofman en Ferry Breedveld [5].
1.3 De keuze
Het ESOF-bestuur ontving twee volledige biedingen, beide voor 2020: Trieste en Leiden/Den Haag [7].
Trieste kreeg 2020. In hetzelfde besluit kreeg Leiden/Den Haag de editie van 2022, “om beide biedingen te eren” [8]. Het bestuur noemde de kandidaten “gelijkwaardig” en wilde “beide steden erkennen als European City of Science” [5].
Leiden kreeg 2022 dus zonder nieuwe ronde en zonder concurrentie.
1.4 Den Haag verdwijnt
Den Haag zat in het bod. In de uitvoering komt de stad niet meer voor:
In maart 2019 bestond het kernteam alleen uit Leidse partijen [9].
De stichting die het jaar uitvoerde, werd in 2020 opgericht door vier Leidse instellingen [10].
Een Haagse financiële bijdrage is niet gevonden [10].
Het evaluatierapport van het congres noemt Den Haag niet. De nevenlocatie werd Katowice, via een contract tussen EuroScience en Katowice [11].
Waarom Den Haag verdween, staat in geen openbare bron.
Universiteit Leiden, gemeente Leiden, Leiden Marketing, Haags congresbureau
[5][6]
Boegbeelden
prof. Corinne Hofman, prof. Ferry Breedveld
[5]
Uitvoerder
Stichting Leiden European City of Science 2022, opgericht 3 april 2020
[10]
Oprichters stichting
gemeente Leiden, Universiteit Leiden, LUMC, Hogeschool Leiden
[10]
Raad van toezicht (2021)
burgemeester Henri Lenferink; Annetje Ottow (Universiteit, na Carel Stolker); Kim Smit (LUMC); Joeri van den Steenhoven (Hogeschool); Joost Scholten (bedrijfsleven)
[10]
Dagelijkse leiding
directeur Meta Knol; intendant Lucien Geelhoed
[10]
Inhoud programma
intendant Geelhoed, via de “Playbooks” I en II
[10][12]
Inhoud congres
Executive Board (Leiden2022, EuroScience, Hofman en Breedveld), voorzitter Breedveld
[11]
Marketing en ontvangst
Leiden Marketing, tegen kostprijs
[10]
Waar de beslissingen lagen. De raad van toezicht kon besluiten van de Executive Board terugdraaien op financiële gronden of om risico’s, maar had geen zeggenschap over de inhoud [11]. De inhoud van het jaarprogramma lag bij de intendant. De 365 dagthema’s werden gekozen uit ingezonden ideeën en lagen vast op 1 augustus 2021 [10].
Erkenning achteraf. De directeur van Leiden Marketing, Martijn Bulthuis, werd in december 2019 uitgeroepen tot “machtigste Leidenaar”. Het binnenhalen van ESOF werd daarbij uitdrukkelijk genoemd [13].
Contracten met EuroScience. Leiden2022 sloot drie contracten: een memorandum (21 juli 2020), een licentie op het merk (2 november 2020) en een congrescontract (5 juli 2021) [10]. De bedragen zijn niet openbaar.
3. Kosten
3.1 Begroting
De meerjarenbegroting 2020–2023 was €8,3 miljoen. Eind 2021 werd ze bijgesteld naar €7,5 miljoen [10].
3.2 Aantoonbare financiering
Bron
Bedrag
Bron
EU, project ECS/ESOF
€2.000.000 (totale projectkosten €3.709.149)
[14]
EU, project EUCYS 2022
€2.000.000
[15]
Oprichters samen
€1.200.000
[10]
Gemeente Leiden extra, Kennis door de Wijken
±€200.000 (“2 ton”)
[10]
Ministerie van OCW
±€500.000 (“5 ton”), plus €217.000, plus €210.000 via Naturalis
[10]
Provincie Zuid-Holland
maximaal €250.000
[16]
Rabobank, NWO, Energiepark Leiden
€58.650
[10]
Totaal aantoonbaar
±€6,6 miljoen
Controle. De provincie betaalde 3,8% van de subsidiabele kosten, met een maximum van €250.000 [16]. Dan bedroegen de subsidiabele kosten 250.000 / 0,038 = ongeveer €6,6 miljoen. Dat komt overeen met de optelling.
Ongeveer €4 miljoen van de financiering, zo’n 60%, kwam van de EU.
3.3 Niet in deze bedragen
Personeelsuren van gemeente, universiteit, LUMC, Hogeschool en Leiden Marketing.
Wat aan EuroScience is betaald voor merk en licentie.
Eigen kosten van musea en andere organisatoren.
De definitieve afrekening. Een jaarrekening over 2022 of 2023 is niet gevonden.
4. Beloofde effecten
Belofte
Wanneer, door wie
Bron
Meer dan 4.500 wetenschappers en technologieleiders naar Leiden
bod, 2017
[5][6]
Congresbezoekers besteden gemiddeld ongeveer €1.000 per verblijf
kernteam, 2019
[9]
Blijvende effecten, met Tallinn (Culturele Hoofdstad 2011) als voorbeeld
kernteam, 2019
[9]
Meer dan 500.000 bezoekers aan de activiteiten
doelstelling Leiden2022
[17]
Minstens 8 congressen
doelstelling Leiden2022
[17]
365 dagen activiteiten
doelstelling Leiden2022
[17]
200 wetenschappelijke lezingen
doelstelling Leiden2022
[17]
Congres: 1.500 deelnemers ter plaatse en 3.500 online
doelstelling ESOF2022
[11]
Een blijvend “European City of Science”-label
EU-projectdoel
[14]
Leiden vestigen als kennisstad
EU-projectdoel
[14]
5. Werkelijke effecten
5.1 Belofte tegenover meting
Belofte
Gemeten
Bron
>4.500 wetenschappers naar Leiden
ruim 1.500 deelnemers ter plaatse
[11]
1.500 ter plaatse + 3.500 online
ruim 1.500 ter plaatse, ±900 online, ±700 in Katowice: samen ±3.100
[11]
>500.000 bezoekers
400.000–500.000 “engagements”, geschat; het aantal unieke bezoekers is “aanzienlijk kleiner” en onbekend
[17]
Minstens 8 congressen
85 congressen in 2022, waarvan 56 gekoppeld aan Leiden2022
[17]
365 dagen activiteiten
gehaald, minstens één activiteit per dag
[17]
200 lezingen
niet geteld; het rapport neemt aan dat het gehaald is
[17]
Blijvend label
vervallen: EuroScience ging failliet, de EU stapte over op een ander model
[4][18]
Leiden als kennisstad
87% van de ondervraagde bezoekers vond dat Leiden wetenschap, kennis, kunst en vakmanschap verbindt
[17]
5.2 Hoe er werd gemeten
Het impactrapport werd gemaakt door het bureau Cigarbox [17]. De metingen bestonden uit:
vragenlijsten onder deelnemers, per onderdeel;
een enquête onder organisatoren;
schattingen van bezoekersaantallen;
tevredenheidsscores en Net Promoter Scores.
De respons was beperkt. Het congres leverde 244 ingevulde vragenlijsten op, de organisatoren 87, het festival 69 [11][17][19].
5.3 De “€20 miljoen economische spin-off”
Bij de afsluiting werd een economische spin-off van meer dan €20 miljoen gemeld [20]. Het rapport geeft een bandbreedte van €20–26 miljoen [17]:
Onderdeel
Laag
Hoog
Grondslag
Uitgaven van congresbezoekers
€7 mln
€9 mln
56 congressen, 13.000 extra bezoekers, 80% van buiten Leiden, gemiddeld 2,1 dag, gemiddelde bestedingen volgens NBTC
Uitgaven van bezoekers aan publieksactiviteiten
€12 mln
€16 mln
400.000–500.000 bezoeken, 70% van buiten de regio, €55 per dagbezoek
Uitgaven van Leiden2022 zelf
€0,7 mln
€0,7 mln
€2,4 mln lokaal uitgegeven, min €1,7 mln lokale sponsoring en subsidie
Het grootste deel, €12–16 miljoen, rust op geschatte bezoeken vermenigvuldigd met landelijke gemiddelde bestedingen. Dat is een modelberekening, geen meting van wat er werkelijk in Leiden werd uitgegeven. Het aantal unieke bezoekers is onbekend [17].
5.4 Wat het rapport zelf niet kon vaststellen
Het impactrapport zegt bij drie van de zeven hoofddoelen dat er “meer informatie nodig” is voor een oordeel [17]. Verder:
Of nieuwe contacten leidden tot onderzoek, publicaties of innovatie: onduidelijk.
Of de samenwerking op lange termijn verbeterde: onduidelijk.
Uitstoot: niet gemeten.
Hoe Leiden in het buitenland wordt gezien: niet gemeten.
Netwerken op het congres scoorde lager dan bij vergelijkbare congressen.
5.5 Musea
Leidse musea telden in 2022 1,3 miljoen bezoeken: 10% meer dan in 2019, terwijl landelijk een daling van 27% gold [17]. Het rapport schrijft 100.000–150.000 extra bezoeken toe aan Leiden2022. Niet alle veertien musea groeiden [17].
5.6 Congressen na 2022
In 2023 telde Leiden 16 internationale congressen volgens de ICCA-maatstaf: 6e van Nederland, 81e van Europa [21]. Dat getal is niet te vergelijken met de “85 congressen” van 2022, omdat ICCA strengere criteria hanteert. Een vergelijkbare telling over meerdere jaren is niet gevonden.
6. Wie werd bereikt
Kennis door de Wijken. Activiteiten in 101 buurten in en rond Leiden, met ongeveer 10.000 bezoeken. 45% kwam in de eigen buurt. 20% bezocht voor het eerst in vijf jaar een wetenschappelijke activiteit [17].
Festival op het Pieterskerkplein. De ondervraagde bezoekers waren gemiddeld 44 jaar oud. 78% had een academische opleiding. 62% kwam uit Leiden [19].
Inwoners vooraf. In 2021 wist 61% van de inwoners niet dat Leiden de titel zou dragen. 84% zei wel interesse te hebben om mee te doen [17].
Zorg voor iedereen. De laagste score in het merkonderzoek: 59% vond dat Leiden “zorgt voor alle leden van de samenleving” [17].
Het programma bereikte vooral mensen die al dicht bij wetenschap staan. De buurtactiviteiten bereikten een kleiner, breder publiek.
7. Wat bleef
Het label. EuroScience ging failliet en wordt geliquideerd [4]. De Commissie verving de City of Science door de open oproep “Science comes to town”. In 2026–2027 loopt die in Aarhus, Genua, Kaunas en Oeiras, met €6 miljoen EU-geld [18][22]. Leiden of Leidse organisaties zitten niet in dat consortium [22].
Een blijvende structuur. Het impactrapport noemt er geen [17].
De stichting. Directeur Knol en intendant Geelhoed schreven het eindrapport The Leiden2022 Model for Public Engagement with Science (15 januari 2023). Ze presenteren Leiden2022 daarin als model voor latere steden [23].
Leiden&Partners. De subsidieovereenkomst 2024–2027 verplicht Leiden&Partners om mee te doen aan “het vervolg op European City of Science” [24].
8. Wat niet openbaar is
De vijf partners in het ESOF-bestuur en de selectiecriteria.
De rol van Den Haag na 2017 en waarom die verviel.
De bedragen in de contracten met EuroScience.
De definitieve jaarrekening van de stichting over 2022 en 2023.
De eigen bijdragen per oprichter.
Deze stukken zijn op te vragen bij de gemeente Leiden en via de KvK-deponering van de stichting.
9. Conclusie
Leiden kreeg de titel zonder concurrentie. Er waren twee biedingen voor twee edities, en beide steden kregen er een. Den Haag zat in het bod en verdween uit de uitvoering.
Het geheel werd gedragen door een kleine kring: gemeente, universiteit, LUMC, Hogeschool en Leiden Marketing. Dezelfde instellingen dienden het bod in, richtten de stichting op en zaten in het toezicht.
Het kostte ongeveer €6,6 miljoen, waarvan zo’n €4 miljoen EU-geld. De eigen uren van de instellingen komen daar nog bij.
Het bod beloofde 4.500 wetenschappers. Er kwamen er ruim 1.500. De €20 miljoen economische spin-off is een modelberekening. Bij drie van de zeven hoofddoelen kon het eigen rapport geen oordeel geven. Het label waarvoor het jaar de basis moest leggen, bestaat niet meer.
Geannoteerde referenties
[1] EuroScience. “How did EuroScience start?” Waarom lezen? De oprichting in 1997, de eerste bestuurders en de vroege financiering door de Commissie en het Stifterverband. https://www.euroscience.org/about/history/
[6] WBC-RTI. “ESOF2022 Leiden/The Hague: Leiden will be the European city of science in 2022”. Waarom lezen? Somt de vier indieners op, inclusief het Haagse congresbureau. https://wbc-rti.info/object/news/20859.html
[23] Knol, M. en Geelhoed, L. (2023). The Leiden2022 Model for Public Engagement with Science. Zenodo. Waarom lezen? Het eigen eindoordeel van de leiding en het voorgestelde model. https://zenodo.org/records/18258655
de “wolframcrisis” is geen grondstoffenschaarste , maar een bewust gecreëerd knelpunt in de verwerkingsfase: China controleert de omzetting van erts naar ammoniumparatungstate (APT), waar de prijs sinds 2025 enorm door steeg.
De Verenigde Staten hebben daarop met eigen beperkingen gereageerd, waardoor een politiek bepaalde prijsvloer is ontstaan die vooral Europese recyclers raakt.
De oplossing ligt niet in grootschalige mijnbouw, maar in kleinschalige, regionale ketens: herslijpen, zinkrecycling en directe carbideproductie, die China’s knelpunt omzeilen.
Europa moet daarvoor zijn eigen wolfraamschroot in de eigen keten houden, in plaats van het te exporteren naar de Verenigde Staten, die hun schroot sinds augustus 2026 volledig voor de eigen markt reserveren.
NB: Tungsten en wolfram zijn exact hetzelfde chemische element: het symbool W.
Tungsten is everywhere and nowhere. It sits in the drill that cuts a printed-circuit board, the insert that machines a crankshaft, the tip of an armour-piercing shell and the wall of a fusion reactor. In 2026 the metal became a byword for scarcity. Its price rose roughly ninefold in under two years. Dart sets cost thirty percent more. A metalworking firm in Zoetermeer speaks simply of “the tungsten crisis.”
Yet the ground is not the problem. World reserves exceed 4.7 million tonnes. At the 2025 mining rate of 85,000 tonnes, that is more than fifty years of supply. The scarcity sits in one step of the chain: the conversion of ore into ammonium paratungstate (APT), the white salt from which every tungsten product begins. China controls about 85–86 percent of APT capacity, according to S&P Global and the Centre for Materials & Resilience — figures that remain single-sourced and could not be independently verified.
The crisis, in other words, is a node, not a resource. And the node has been closed deliberately, twice.
2. How Beijing closed the gate
China’s Announcement No. 10 of 4 February 2025 placed specific tungsten items under export licence: APT, tungsten oxides and carbide powder, heavy alloys above 17.5 g/cm³, massive tungsten pieces above set dimensions, and the production technology for all of these. The controls sit precisely on the node. Concentrate, metal powder and finished tools are absent from the list. China mines 67,000 of the world’s 85,000 tonnes (79 percent); the prefecture of Ganzhou says of itself: think of Ganzhou with tungsten as you think of the Middle East with oil.
Beijing then locked the gate shut. On 26 December 2025, the Ministry of Commerce published its state-trade list for 2026–2027: only fifteen companies in all of China may legally export tungsten. Unlike a negotiable quota, a state-designated exporter list is difficult to contest or unwind. Chinese APT exports collapsed from 782 tonnes in 2024 to 243 tonnes over January–November 2025, with 57 percent of that going to Japan alone; by early 2026 they approached zero, and total tungsten exports fell 34 percent year-on-year in the first two months of the year.
The market responded as a tree does when its trunk closes. Tungsten concentrate rose from $750–850 to $2,500–2,800 per metric ton unit (mtu), where it has held since late May 2026. APT in Rotterdam traded above $3,000 per mtu. The International Energy Agency called tungsten the fastest-rising strategic metal.
Here is what the price is not. It is not an energy price. Making tungsten carbide from concentrate takes about 10 kWh of electricity per kilogram of tungsten; the step from concentrate to APT takes about 1.9 kWh. At €0.15 per kWh, energy was four percent of the APT price at the start of 2025. By 15 October 2026 it was roughly half a percent, and a tenth of that for the APT step itself. Nothing became costlier to make. One gate closed, and the price followed it up.
3. Washington answers — and builds a second gate
The conventional reading of this crisis stops at Beijing: the price is geopolitical, and what one political decision raised, another can lower. That reading is now incomplete. In 2026 the United States built its own gate, on the other side of the Atlantic market.
On 27 August 2026, the Bureau of Industry and Security published a temporary final rule under the Defense Priorities and Allocations System. American sellers of tungsten waste and scrap (Schedule B 8101.97) and battery black mass must allocate one hundred percent of their monthly sales to US buyers. Exports are prohibited unless BIS grants an exception; a narrow pathway exists for material refined abroad and returned. The rule runs to 27 August 2027 and can be extended. Washington used the Defense Production Act, not standard export-control law, and its logic is unambiguous: tungsten has not been mined commercially in the United States since 2015, so scrap is one of the few domestic sources it can secure immediately.
A second measure follows on 1 January 2027, when US defence procurement rules will generally prohibit tungsten metal powder, heavy alloy and covered components sourced from China, Russia, North Korea or Iran — reaching back through the supply chain to ore and recycled material. A July 2026 executive order tightened the exemptions. Commentators have called this “scrap nationalism”: Washington adopting Beijing’s playbook. The label is fair, and it matters for one group above all — European recyclers.
The consequence is a policy-driven price floor with two supports. If Beijing reopens its gate, the price will not simply fall back, because the American system now excludes Chinese tungsten from defence supply and keeps American scrap at home. Any business case for Western tungsten built on today’s price rests not on one political decision but on two, in two capitals, pulling in the same direction. For mining investors, the old risk remains — a reopened Chinese tap — but it is now paired with a new certainty: a structurally protected Western demand base. The two do not cancel out. They make the price less informative and the commitment longer.
4. History’s warning about trunks
Western tungsten mining has been paid for in wartime, and punished in peacetime. The largest tungsten mine outside China, in Australia, opened in the First World War, reopened just before 1939, produced through the Korean and Vietnam wars, closed in 1990, and is open again now. Ten years ago China sold tungsten below cost and Western mines went under water.
Large projects are again under way. H.C. Starck, owned by Mitsubishi Materials since December 2024, invests about €50 million in Goslar, lifting capacity by nearly half to about 7,000 tonnes a year by 2028; some eighty percent of its feed is already recycled material. Mitsubishi Materials doubles APT capacity in Akita, Japan, by April 2029. Almonty began processing at Sangdong in Korea in July 2026 and has certified the plant for commercial production, with nearly half its output allocated to munitions production in Pennsylvania. Tungsten West is commissioning at Hemerdon in England. Kazakhstan’s Boguty began producing in 2025.
These plants make APT outside China. That is valuable. But they are still trunks: single, large, thirty-year commitments. They carry the same risk the Australian mine carried. Large capacity outside China needs a floor under the price — long-term offtake, strategic stock, or public ownership. Absent that floor, history repeats.
5. Europe pays for a mine and exports its own loop
The European Union has answered the crisis with geology. In its second batch of strategic projects under the Critical Raw Materials Act (9 October 2026), the Commission approved 46 new projects across 16 member states, including three on tungsten. The Borralha mine in Portugal — held by Mineralia, with Allied Critical Metals advancing the neighbouring project — is the only tungsten extraction project in the batch. A correction to earlier analysis of this batch: a tungsten processing plant is in fact included. “EU APT Tungsten Plant,” promoted by Abenójar Tungsten S.A., is planned in Puertollano, Spain — the first EU APT project in the programme. No tungsten recycling project was selected. That is the gap.
The gap is measurable. In 2025 the EU exported 4,911 tonnes of tungsten waste and scrap outside the Union, two-thirds of it to the United States. The Netherlands alone exported 928 tonnes, mostly to Germany, and is a net scrap exporter. Tungsten scrap holds far more tungsten per tonne than any ore. Half of Europe’s carbide already comes from recycling, and the European recycling sector says its capacity is “approaching its limits.”
Meanwhile, every major player except Europe is closing its loop. Japan’s machine-tool builders discourage carbide scrap from leaving the country. China draws on recycled feedstock to offset reduced mining quotas. And since 27 August 2026, the United States — destination of two-thirds of Europe’s scrap exports — requires that American scrap be sold only to American buyers. Europe’s richest secondary ore is flowing toward a market that no longer releases its own.
EU policy has noticed metal scrap in general, but not tungsten in particular. The RESourceEU plan proposes export limits on scrap of magnets, aluminium, possibly copper and battery black mass; tungsten is not named, even though the same programme calls for keeping metal stocks within the Union. Tungsten scrap remains on the “green list” of wastes (B1010), and from 21 May 2027 it may go outside the OECD only to listed countries — a restriction that does not touch the flow to the United States. On 18 September 2026 the Commission published its draft country list under the revised Waste Shipment Regulation; the broader metal-scrap framework is tightening, but tungsten scrap to the US remains unaffected. The Dutch list of critical raw materials (7 July 2026) includes tungsten and contains no measure on scrap.
The correction is obvious and cheap relative to a mine. Fund the loops and the processing, not only the extraction. Treat clean carbide scrap as a resource, not a waste. Keep tungsten scrap in the European loop, as is already proposed for magnets and aluminium. The European Parliament votes on the CRMA amendment on 20 October 2026 — tomorrow, as this is written.
6. Small loops beat the long chain
The routes that bypass the Chinese node are not large. They are small, proven, and they pay at today’s prices.
Regrinding. A worn solid-carbide drill can be reground several times. An independent study finds that more than sixty percent of original tool life returns; vendors claim 85–95 percent. A regrind costs about thirty percent of a new tool. Per unit of cutting work, that is 64–73 percent less new carbide, with no new technology required.
Keeping carbide as carbide: zinc recycling. Worn carbide is placed in molten zinc at 450–500 °C. The zinc swells the cobalt binder, the piece falls apart, the zinc is distilled off and reused, and the carbide is milled to powder. The grain is never rewritten. It costs about 3.3 kWh per kilogram of tungsten — a third of the primary route. Recycled powder is not on the Chinese control list.
Supplier take-back, now at scale. Sandvik reports 70 percent less energy and 64 percent less CO₂ for recycled carbide, aims to buy back ninety percent of its insert volume, and has now gone further: its PowerCarbide grades will carry up to 75 percent recycled cemented carbide as standard across global drill-bit production through 2026–2027, using closed-loop extraction units stationed on every continent. Its Austrian subsidiary Wolfram Bergbau already makes half its products from recycled material. Ceratizit sells a grade with more than 99 percent recycled content at 56 percent lower carbon footprint than its conventional equivalent.
Making carbide straight from ore. Polini and colleagues (2021) milled tungsten concentrate with carbon black, heated it for six hours at 1,150 °C, leached it, and obtained tungsten carbide of 99 percent purity with a 92 percent extraction yield. The route skips APT, oxide and metal. It uses half the energy, four reagents instead of fourteen, and less than half the waste. The carbide sintered into hard metal of industrial quality. The Italian firm F.I.L.M.S. holds the patent. Verification research finds no industrial-scale deployment yet; the route remains laboratory-proven — but so did every other route once.
The economics now support the loop. Carbide scrap fetched under €20 per kilogram in 2023, about €75 in January 2026, peaked at €130–140 in March 2026, and settled at €65–80 by mid-October 2026. European prices in 2026 ran at 80–100 €/kg. Recycling capacity, not scrap availability, is the binding constraint: processing runs at capacity, and new facilities take two to three years to build and qualify. The US Department of Energy has already put an $11.5 million grant behind domestic carbide recovery.
This is the shape of the answer. A zinc furnace serves a region. A direct carbide line can stand beside a mine. A regrind service needs a workshop and a spare-tools bin. Against one trunk, a net of small meshes — each mesh cheap, fast, and outside the reach of Announcement No. 10.
7. The substitutes are ready; the incumbents are not
For many functions, tungsten already has a working alternative. Silicon nitride ceramics machine cast iron three to four times faster, without tungsten. Hard turning with PCBN instead of grinding uses ten to fifteen times less energy per volume removed, and one lathe replaces several grinders. Ti(C,N) cermets finish steel at higher speeds. SiAlON ceramics cut nickel alloys at four times the conventional speed. Diamond tooling lasts four times longer in aluminium and composites. Lead and steel do most shielding work at a fraction of the cost.
In semiconductors, the substitution has crossed from possibility to plan. SK hynix has verified 375-layer NAND flash and will begin mass production by the end of 2026, replacing part of the tungsten in its word lines with molybdenum — chosen for lower resistance as wiring narrows. Samsung was first to introduce molybdenum in metal wiring; Micron has adopted it; future 480- and 604-layer generations will rely on it more heavily. Lam Research and Tokyo Electron supply the deposition equipment. A materials cycle that lasted a quarter-century is ending.
Two caveats keep this in proportion. First, the volume is small: industry expectations run at 25–80 tonnes of molybdenum per year for memory between 2027 and 2030. Technology routes around the problem; it does not relieve the tungsten market. Second, molybdenum is not a repeat of tungsten’s geography: China mines 37 percent of molybdenum, against 79 percent of tungsten, and its export control on molybdenum covers only fine powder for missile parts.
Yet the cutting-tool industry’s answer to the 2026 crisis has been price rises — up to sixty percent on some tool lines — and carbide grades with less tungsten, not a shift to the tungsten-free tools it already sells. Ceramics held 8.5 percent of the cutting-tool market in 2011; nothing in the reviewed sources shows that share moving now. The obstacle is not physics. It is habit, installed machines, and a supply chain built around one material. Firms that adopt the alternatives now buy independence from the node.
Two functions have no plentiful alternative: density combined with hardness, as in armour-piercing ammunition, and the wall of a fusion reactor. These are the places for strategic stock, not substitution.
8. No way out through the elements
Could tungsten be made from something else? No route works at acceptable cost, and the physics has now been computed in detail.
The state road: some stable atoms become unstable when stripped of their electrons — measured for dysprosium-163, rhenium-187 and thallium-205. Computed at every charge state, no natural nucleus opens a path to tungsten. The nearest candidate, tantalum-181, stays closed by 138 keV even when fully ionised.
Capture: tantalum-181 captures a neutron and becomes tungsten. In a strong reactor flux, a thin tantalum foil converts 4–9 percent per year. But every tungsten atom costs one neutron, which puts the energy at a minimum of 2.6 million kWh per kilogram — against roughly 10 kWh for the chemical route. And tantalum is 34 times scarcer than tungsten.
One reverse effect is worth a footnote. A fusion reactor’s tungsten wall slowly becomes rhenium and osmium: about 0.27 tonnes of rhenium per gigawatt-year, three-tenths of a percent of world output. Fusion consumes the very metal its wall is made of — one more reason the reactor wall belongs on the short list of functions with no substitute.
9. What it means, for whom
For governments and the EU. Fund loops and processing, not only mines. Add tungsten scrap to the scrap-export measures already drafted for magnets and aluminium — before the flow to the United States, now a closed American loop, becomes irreversible habit. Build strategic stock where no substitute exists: defence and fusion. Support the Puertollano APT plant and Borralha, and match them with the recycling capacity that no strategic project yet provides.
For investors. Mining shares have risen three- to tenfold. Their value rests on a price set by two export regimes, not by cost. History shows what a reopened Chinese gate does to uncovered capacity; the American floor shows what it does to prices for compliant, non-Chinese supply. Processing and recycling capacity outside China — with price protection and qualification — is the sounder exposure.
For entrepreneurs. The openings are small and regional, and the constraint is now capacity: regrind services; scrap pools that reach the half-tonne lots the market requires; zinc recycling near clusters of tool users; qualification partnerships for new recovery lines. Each is a two-to-three-year build with a price environment that pays for it.
For large manufacturers. Specify by function, not by habit. Ceramic, PCBN and cermet tools already do part of the work without tungsten. Return all carbide to the supplier. Buy recycled grades — they are now standard, certified and cheaper in carbon as well as in exposure.
For the people involved. In Borralha, opposition is growing, not fading: the environmental assessment drew 653 submissions, a residents’ collective — “Cávado Sem Minas” — has taken the case to protest and court, and the mine’s EU designation has intensified the contest between Lisbon, Brussels and the people who drink the water. In Rwanda, Trinity Metals employs more than 6,500 people, ninety-nine percent Rwandan. At the machine, a loop means separate bins, clean scrap, a spare stock of tools during regrinding, and one person who owns it. Anyone grinding carbide in-house faces the EU’s proposed cobalt exposure limit of 0.01 mg/m³. The circular economy has working conditions, too.
10. The shape of the answer
The tungsten crisis is often described as a shortage. It is better described as a design flaw, exposed. A century of supply chain design concentrated the world’s tungsten in one processing step, in one country — a trunk carrying every leaf. China closed the trunk; the United States is now closing its own; each closure raises the price of every alternative route.
The way out was never going to be a bigger trunk. It is the net: reground tools that never re-enter the chemical chain, zinc furnaces that keep carbide as carbide, suppliers who take their own products back at 75 percent recycled content, a one-step carbide line waiting for its first industrial owner, ceramics and cermets already cutting cast iron without tungsten, and molybdenum quietly replacing tungsten in memory chips. None of these routes needs thirty years. None of them passes through the node. All of them need the one thing that policy has not yet given: Europe’s own scrap, kept in Europe’s own loop.
The ground holds fifty years of tungsten. The gate holds the price. The net, small mesh by small mesh, is how the West walks around the gate.
Annotated references
The occasion and the market
Smid, R. (2025). “Chinese dominantie in metalen staat Europese wapenplannen in de weg.” Het Financieele Dagblad, 11 June 2025. Why read? Ganzhou’s self-description, the Govini count of Chinese dominance in 78% of 80,000 American weapon parts, and the EU Institute for Security Studies’ thirty-year refinery warning, in one piece.
FD editors (2026). “Aandeel wolfraamproducent EQ Resources knalt omhoog.” Het Financieele Dagblad, 20 July 2026. Why read? The IEA’s verdict on tungsten as fastest-rising strategic metal and the sixfold price rise. Reading advice: the 1,680 tonnes cited for EQ Resources are most likely tonnes of WO₃, mostly from Spain.
Brasser, P. (2026). “Zonder dit heetste metaal geen AI, defensie en ruimtevaart.” Het Financieele Dagblad, 16 October 2026. Why read? How the crisis reaches small firms, the darts, and the reopening of Minas da Borralha; the CMR’s warning that a hundred new mines solve the wrong problem.
Mining.com (2026). “Tungsten price breaks records as China export curbs, military demand boost investment.” Why read? APT above $3,000/mtu in Rotterdam, Project Blue’s market and defence-demand estimates (defence ~12% of demand, rising to ~15% by 2027–28), and the December 2025 fifteen-firm export decision in one market overview.
Fastmarkets (2026). “Tungsten markets fragmenting as domestic Chinese APT market diverges from exports” and “Tungsten 2026: Geopolitics sets global tone.” Why read? The split between Chinese domestic and export pricing, concentrate at $2,500–2,800/mtu since May 2026, and the trade’s own expectation of persistent volatility. Caveat: paywalled price data; figures cited here via secondary reporting.
Investing.com (2026). “Tungsten Price Shock Signals Deeper Supply Crisis.” Why read? The 310% January–July 2026 price increase, the ~40% fall in Chinese controlled-product shipments in 2025, and the US defence sourcing prohibition effective 1 January 2027.
CNBC (2026). “China, tungsten and a supply shock in metal critical for war that will last far beyond Iran conflict.” 3 June 2026. Why read? The demand side of the story: simultaneous conflicts depleting tungsten stocks, the $1.6 billion US Export-Import Bank commitment, and the Kazakh supply deal and its scrutiny.
The node and the controls
U.S. Geological Survey (2026). Mineral Commodity Summaries 2026: Tungsten, Molybdenum, Tantalum. Why read? Reserves, production by country, and the 2025 price path — the basis for the “node, not resource” argument.
MOFCOM and GAC (2025). Announcement No. 10 of 2025, 4 February 2025. Why read? The exact items under export licence. Reading advice: compare items 1C004, 1C117 and 1E004 with what is absent — concentrate, metal powder, finished tools.
MOFCOM (2025). Document 商贸函〔2025〕696号, 26 December 2025: state trading enterprise list for tungsten, antimony and silver, 2026–2027. Why read? The hardening of the gate: fifteen designated tungsten exporters, replacing quota administration with a state-designated structure that is difficult to contest or unwind.
Kamoa Capital (2026). “China’s Secret Tungsten Law — What Western Miners Need to Know.” Why read? The immediate effect of the whitelist: exports down 34% by volume in January–February 2026. Caveat: commercial blog; figures consistent with news reporting.
S&P Global (2026). Tungsten Market Report, 8 May 2026. Why read? China’s share of APT capacity. Caveat: single source; the figure could not be independently verified and remains the essay’s principal open point.
SMM (Shanghai Metals Market). APT and carbide scrap price reports, February–October 2026. Why read? The price path of APT and scrap through the crisis.
The second gate
Federal Register (2026). “DPAS Directive Allocation Order and Additional Requirements for Recoverable Critical Minerals and Materials,” 6 August 2026, effective 27 August 2026–27 August 2027. Why read? The primary text of the American scrap rule: 100% of monthly sales of tungsten waste and scrap (Schedule B 8101.97) and black mass to US persons, exceptions at BIS discretion, with a refinement-and-return pathway.
Baker McKenzie Global Sanctions Blog (2026). “BIS Restricts Exports of Black Mass and Tungsten Waste and Scrap without a License.” Why read? Clean legal analysis of the rule’s scope, the narrow “US person” definition, and the 14-day exception process.
Resource Recycling (2026). “New federal rule restricts black mass, tungsten scrap exports,” 5 August 2026. Why read? The rule’s reach beyond arm’s-length exports, including foreign-owned processing affiliates, and the return-of-material exception.
Eurasia Review (2026). “Scrap Nationalism: Washington Adopts Beijing’s Playbook On Tungsten And Black Mass,” 10 October 2026. Why read? The sharpest statement of the strategic meaning: the United States has reversed a decade of its own raw-materials doctrine.
Europe’s flows and policy
Eurostat (2026). Comext database DS-045409, product code 810197. Why read? The scrap trade flows: 4,911 t EU exports in 2025, two-thirds to the United States; the Netherlands a net exporter. Re-queried and matched exactly by an independent reviewer.
European Commission (2026). Selected strategic projects under the CRMA, second list, 9 October 2026. Why read? The 46 new projects, including three on tungsten. Correction incorporated: “EU APT Tungsten Plant” (Abenójar Tungsten S.A., Puertollano, Spain) is a processing project in the batch, alongside Borralha as the sole extraction project. No tungsten recycling project was selected.
Leapman (2026). “CRMA: 46 new Strategic Projects approved in Europe.” Why read? The Spanish project detail that corrects the earlier reading of the batch, and the observation that more than half of the recognised projects include recycling.
European Cutting Tools Association (2026). “Recycling Key to Europe’s Carbide Supply,” 8 May 2026. Why read? Half of Europe’s carbide from recycling, and capacity “approaching its limits.”
European Commission (2025–2026). RESourceEU; proposal COM(2025) 946 to amend the CRMA. Why read? What the EU protects (magnets, aluminium, possibly copper, black mass) and what it leaves out (tungsten scrap). The Parliament plenary vote is scheduled for 20 October 2026.
steelnews.biz (2026). “Metal Scrap: Export Restrictions in Green Disguise,” 23 September 2026. Why read? The 18 September draft country list under the revised Waste Shipment Regulation — the tightening framework that nonetheless leaves tungsten scrap to the United States untouched.
Ministerie van Economische Zaken (2026). Lijst kritieke grondstoffen, 7 July 2026. Why read? Tungsten on the Dutch list, without a scrap measure.
Loops and routes
Zeiler, B., Schubert, W.-D. and Bartl, A. (2018). “Recycling of tungsten (part 1).” ITIA Newsletter, May 2018. Why read? The world flow of tungsten and the recycling rate per product.
Furberg, A., Arvidsson, R. and Molander, S. (2019). “Environmental life cycle assessment of cemented tungsten carbide (WC-Co) production.” Journal of Cleaner Production 209, 1126–1138. Why read? The energy of every step of the chain — the basis of the energy-share calculation. Reading advice: use one table per route; do not mix primary and recycling tables.
Polini, R. et al. (2021). “Toward Greener Synthesis of WC Powders for Cemented Tungsten Carbides Manufacturing.” ACS Sustainable Chemistry & Engineering 9, 8458–8466. Why read? Carbide from ore in one step: 99% purity, 92% extraction yield, half the energy, industrial-quality hard metal. Verification note: still laboratory-proven; no industrial deployment of the F.I.L.M.S. patent (EP3098199B1) was found.
Conradie, P. J. T., Oosthuizen, G. A. and Dimitrov, D. (2017). International Journal of Advanced Manufacturing Technology 90. Why read? An independent measurement of what regrinding recovers — the conservative 60%-plus figure.
Sandvik (2023, and 2026 ReCarbide announcements). Why read? The supplier’s own numbers: 70% less energy, 64% less CO₂, a 90% buy-back ambition — and now up to 75% recycled carbide as standard in PowerCarbide drill bits through 2026–2027, with extraction units on every continent.
Ceratizit, upGRADE KLC20+ product information. Why read? More than 99% recycled content at 8.99 against 20.64 kg CO₂e per kg — 56% less than the conventional grade.
H.C. Starck Tungsten (2026). Press release on the Goslar expansion, 30 July 2026; Mining Weekly (2026). Mitsubishi Materials Akita expansion, 9 September 2026. Why read? The large APT investments outside China — the trunks, with their thirty-year horizons.
DiscoveryAlert (2026). “Tungsten Carbide Recycling: Strategic Material Recovery Solutions for Industry.” Why read? The $11.5 million DOE grant and the capacity-expansion logic; pairs with reporting that new recycling facilities take two to three years to build and qualify.
Ultra-Met (2026). “Zinc Furnace Technology for Carbide Recycling.” Why read? A clear practitioner’s description of the zinc process and why it preserves carbide grain integrity.
Substitutes
Siwiec, J. and Zębala, W. (2012). “Hard turning vs grinding of pinion shaft 62 HRC.” Cracow University of Technology. Why read? The measured energy gap between PCBN turning and grinding.
Cutting Tool Engineering (2000). Article on silicon nitride ceramics for cast iron. Why read? The speed advantage of tungsten-free ceramics on cast iron.
The Elec (2026). “SK hynix to Mass Produce 375-Layer NAND by Year-End, Introduce Molybdenum.” Why read? The most precise account of the chip substitution: molybdenum in the word lines, chosen for lower resistance; Samsung first, SK hynix via Tokyo Electron deposition.
TechSpot (2026). “SK hynix moves 375-layer NAND into mass production, replacing tungsten with molybdenum.” Why read? The roadmap: 480- and 604-layer generations will lean on molybdenum more heavily; molybdenum demand is climbing in the supply chain.
Lam Research (2025). ALTUS Halo press release, 19 February 2025. Why read? The deposition platform behind molybdenum in memory, and why its volume remains small.
Jarvis Cutting Tools (2026). “Best Alternatives to Tungsten Carbide Cutting Tools in 2026.” Why read? How mainstream the substitution advice has become in machining practice — and, by contrast, how little the incumbents have moved.
Transformation
Choudhary, P. and Qi, C. (2025). “Evaluation of bound-state β⁻-decay half-lives of fully ionized atoms.” arXiv:2507.08199. Why read? The atoms that change element when stripped bare — and why tungsten is not among the reachable ones.
Gilbert, M. R., Sublet, J.-Ch. and Dudarev, S. L. (2017). “Spatial heterogeneity of tungsten transmutation in a fusion device.” arXiv:1612.03892. Why read? How a fusion wall turns tungsten into rhenium — the one transformation of elements the essay keeps.
Method
Konstapel, J. (2026). The Respondent–Judge Protocol, version 1.0. Constable Research, Leiden. Why read? How predictions, counts and independent reviews were separated across three rounds of work; the corrections of version 3.0 and the verification round of 19 October 2026 (two twelve-query search rounds, preference for primary sources: Federal Register, European Commission, ACS) follow the same discipline.
A settlement on Mars should not be judged only by how much cargo arrives. The more important question is whether it can continue to operate when deliveries from Earth stop. A large shipment can build a settlement, but it cannot by itself make that settlement independent. Independence depends on whether essential systems can renew their own inputs and replace their own losses.
This essay proposes a different way to organise the programme. First, measure how well each essential cycle closes. Second, test life-support and food systems on Earth before sending them to Mars. Third, send a compact industrial seed and the knowledge needed to use local resources. Fourth, count the interfaces between systems, because every interface can introduce a failure point. Fifth, design landed equipment for a useful second life. Finally, judge progress by demonstrated capabilities, not by tonnage alone.
The source paper’s combined model estimates that a settlement of 1,000 people could require roughly 30–114 landings under its stated assumptions, rather than the 10,000 landings implied by a million-tonne transport plan. That is a model result, not a demonstrated mission design. It depends on local manufacturing, early closure of essential life-support loops, and assumptions about industrial growth. The paper also explores a speculative possibility: that the geometry of small gaps might affect measured weight. That proposal is not established physics. It should be separated from the much more practical case for better recycling, local production and disciplined testing.
1. Define success correctly
SpaceX has described the goal as a city that can survive if resupply ships stop arriving from Earth for any reason. This is a demanding standard. It means more than keeping a crew alive until the next launch window. It means having no essential dependency that eventually forces the settlement to request another shipment.
The familiar transport plan measures progress in payload mass and launch frequency. The source paper uses an indicative target of about one million tonnes, delivered over roughly twenty years. With around 1,000 ships per launch window and a window every 26 months, that amounts to about ten major windows and 10,000 landings if each ship delivers 100 tonnes. These figures describe a transport scenario. They do not show whether the resulting city can function independently.
A better measure is closure: the degree to which a system can maintain its function without importing the materials it consumes or loses. Oxygen, water, food, nutrients, energy, machinery and replacement parts each need their own account. A settlement may recycle water well while remaining dependent on imported electronics. It may grow much of its food while losing essential minerals in waste. A single overall percentage can conceal these weaknesses.
The first management tool should therefore be a closure ledger. It should list every essential loop, its measured recovery rate, its losses, its reserve capacity and the evidence behind each figure. Unknown values should be marked as unknown. This ledger should be prepared before the architecture is selected. Otherwise, the team risks choosing a design first and defining success around it afterwards.
2. Close the difficult loops on Earth
A Mars mission should not be the first full test of a settlement’s life-support system. Earth-based trials can reveal failures at much lower cost. They can also be repeated, instrumented and inspected by independent teams.
The source paper points to Lunar Palace 365, a Chinese closed-environment experiment that operated with human crews and crop systems. It reports high closure for oxygen and water, but lower closure for food and solid waste. The paper gives an overall closure figure of 98.2 percent. These figures suggest that recycling air and water is not the only challenge. Food production and the treatment of biological waste remain important sources of dependence.
Small percentages matter. If the unrecovered share of a system falls from 1.8 percent to 1 percent, the required replacement flow can fall substantially, assuming the same accounting boundary and operating conditions. The source paper estimates that improving overall closure from 98.2 to 99 percent reduces imports by a factor of about 1.8. A rise to 99.9 percent would reduce them by about a factor of 18. These are arithmetic illustrations. Real systems may not scale linearly, and different loops may fail in different ways.
The next step should be a sealed greenhouse trial with a crew of realistic size. It should run long enough to reveal cumulative problems. Researchers should measure food output, crop diversity, water recovery, oxygen balance, solid-waste processing, nutrient losses and the need for replacement supplies. The target for each loop should be set in advance. Results should be independently checked.
Past experiments warn against treating a closed habitat as a simple machine. Biosphere 2 experienced a major decline in oxygen, linked in the source literature to biological activity in the soil and carbon dioxide absorbed by concrete. BIOS-3 and other life-support studies also identified persistent losses of nutrients and minerals. These examples do not show that closure is impossible. They show why it must be demonstrated in operation rather than assumed from a design diagram.
3. Send an industrial seed, not the whole city
The central economic opportunity is to use Martian resources for most of the settlement’s bulk material. Mars has regolith, rock, atmospheric carbon dioxide and, in some locations, water ice. These resources are not ready-made construction materials. They must be mined, processed and converted. Yet processing local material may be more practical than lifting every wall, tank and spare part from Earth.
The source paper calls the initial industrial package a seed. It would include robots, power systems, tools, manufacturing equipment and the instructions needed to operate and expand them. Its purpose is to establish a production chain. The seed would make more equipment, which would then expand the production chain. The goal is to reduce the share of each new installation that still has to be imported.
Research cited in the source paper describes robotic processing of regolith, casting in sintered-regolith moulds, sulphur-based concrete, plant growth in Martian soil simulants, cyanobacteria grown under simulated Martian conditions and biological extraction of metals from basalt. Each result addresses a part of the problem. None, by itself, demonstrates a complete Martian industrial economy.
The crucial metric is the imported share of each class of equipment. Call this share f. It should be measured separately for electronics, bearings, seals, sensors, catalysts, motors and other components. A factory may produce most of its structure locally while remaining dependent on imported control electronics. Such a factory is productive, but not yet independent.
The source paper uses a growth model derived from earlier work on self-replicating manufacturing systems. It estimates that a 41-tonne seed could support a much larger installed industrial base over successive generations. The paper applies that model to Mars, while acknowledging that the growth rate has not been demonstrated under Martian conditions. Solar energy is weaker on Mars than on the Moon, and the availability of materials, maintenance and process control may change the result. The growth schedule must therefore be tested with realistic power and material constraints.
The operational rule is simple: ship the knowledge and the minimum equipment needed to begin production. Then measure, generation by generation, which components can be made locally. Do not count a component as locally available merely because its housing can be manufactured on Mars. The full component must work to the required specification and lifetime.
4. Count interfaces and reduce failure points
Complex systems fail not only because individual components are weak. They also fail at the interfaces between components. Materials expand at different rates. Seals leak. Electrical connectors corrode or loosen. Fluids cross boundaries. Thermal protection must remain attached to the structure beneath it.
The source paper calls these interfaces seams. It borrows a formal result from the Vacuum.Net Theory: in the theory’s mathematical model, two differently closed triples cannot both close across the boundary in the same way. The paper then reads this as a reason to count engineering interfaces. The mathematical result concerns the theory’s defined trits. Applying it to actual machines is an interpretation, not a proven law of engineering. Still, interface counting is a useful engineering practice in its own right.
Consider a heat shield. The metal structure and ceramic tiles have different thermal properties. Their attachment points and gaps require careful design. The source paper proposes a test using photographs of recovered upper stages. Investigators would define the seam area before inspecting tile damage. They would then test whether damage is concentrated at seams more often than the seam area alone would predict. The test should specify its failure threshold in advance. If the public photographs are too few, the result should be reported as inconclusive.
The same approach can compare transport architectures. A design may involve launch, staging, docking, propellant transfer and landing. Each transition introduces hardware, procedures and opportunities for failure. The source paper estimates 13–18 seams for one Starship design case and 5–7 for a minimum case. These counts depend on the counting rule and mission assumptions. They should not be treated as universal values.
Before selecting an architecture, engineers should publish a clear interface-counting rule. They should count every candidate design using the same rule. They should repeat the comparison under a second reasonable rule. If the ranking changes, the result is too sensitive to the definition to support a confident choice. When mass and performance are otherwise comparable, a design with fewer difficult interfaces may be easier to operate and maintain.
5. Treat landed hardware as industrial stock
A one-way cargo vehicle does not have to be waste. Its structure, tanks, engines and other components may become useful material on Mars. The source paper estimates a dry mass of about 100 tonnes for a landed Starship, but explicitly marks this figure as an assumption that requires confirmation from primary SpaceX data.
The idea is to design for a second life from the beginning. Tanks might become pressure vessels or storage units. Structural sections might become supports. Selected engine components could become spare parts. Steel could be processed as feedstock. The exact uses would depend on material properties, contamination, inspection and the tools available on Mars.
The same principle applies to waste from other activities. Crop residues may become inputs for biological materials. Sulphur may serve as a binder. Processed regolith may become bricks or construction elements. These are promising directions, but each requires testing for strength, durability, toxicity and performance in Martian conditions.
A useful design review should ask three questions about every major item sent to Mars. What is its primary function? What can it become after that function ends? What equipment is needed to reuse it safely? A second life has value only when the conversion process is practical. Reuse should be engineered, not assumed.
6. Choose transport methods with a common scorecard
The source paper discusses several ways to reduce the burden of transport. These include one-way ships, propellant production on Mars, water ice instead of water-bearing regolith, oxygen production from local resources, lunar water as a possible supply and alternative launch infrastructure. The options differ in maturity and should not be treated as equally available.
Making propellant on Mars could reduce the need to transport return fuel from Earth. NASA studies have examined production of methane and oxygen from local resources. The source paper calculates that a large return load would require substantial plant capacity and power. That makes local propellant a strategic option, but not a free resource. The settlement must first build the extraction, processing, storage and power systems needed to produce it.
Water ice may require less energy to process than extracting small amounts of water from regolith. But the relevant ice deposits must be located, characterised and reached. The choice of landing site therefore affects the whole industrial plan. A site with accessible ice may be preferable to a site with better conditions in another category, but only after the trade-offs are quantified.
Aerocapture could reduce the amount of propellant needed for arrival by using the atmosphere to remove orbital energy. The source paper notes that the approach has not been demonstrated in the relevant operational context. It should be treated as a technology-development option, not as guaranteed savings.
All transport choices should use one scorecard. It should include landed mass, propellant, energy, interface count, operational maturity, repairability, local resource requirements and consequences of failure. The cheapest option on paper may not be the most reliable. The best option is the one that performs well under a transparent set of assumptions and can be tested before it becomes mission-critical.
7. Measure progress by demonstrated capabilities
A programme can report many launches and still fail to establish a self-sustaining settlement. It can also make important progress before delivering large amounts of cargo. A successful test of a closed water loop, a reliable local manufacturing process or a replacement part made from Martian feedstock can reduce future dependence.
The programme should use milestones that can be observed and independently verified. Examples include: a life-support loop that operates for a defined period; a greenhouse that meets a pre-set food target; a manufacturing unit that produces a specified part; a second-generation machine that builds a working third-generation unit; and a maintenance team that repairs equipment without imported components.
Tonnage and cost remain important. They should be treated as outputs of the design, not as substitutes for operational independence. The programme should report both the amount delivered and the amount of essential function that can be sustained locally. Each claim should state whether it is measured, calculated from stated assumptions, inferred from a model or still open.
The source paper notes a large gap between the Starship flight rate it calculates for a million-tonne programme and the flight rate represented by the launch history available to it. That gap is a reminder to separate demonstrated capability from planned scale. A projection is not an operational rate. The schedule should change when the evidence changes.
8. The unresolved question of weight
The paper also explores a more speculative idea. In the Vacuum.Net Theory, matter is described as a stable closure in a larger net. The theory proposes that some effective constants may depend on the local state of that net. The paper asks whether the geometry of small gaps could alter the measured weight of an object.
It distinguishes two routes. The first would require changing or opening the closure associated with mass. The paper concludes that the energy cost of this route is vastly greater than the energy needed to lift a kilogram into low Earth orbit. It therefore treats this route as impractical.
The second route proposes that geometry might change the state of the net around an object. Under a set of explicit assumptions, the paper’s path model predicts a possible difference in weight for objects built with narrow gaps. The paper compares this idea with earlier measurements of stacks of metal discs and a solid cylinder. It acknowledges that the existing measurements do not settle the question.
This proposal is not an established method of reducing gravity or launch energy. The connection between the theory’s mathematical variables and gravitational coupling remains unproven. The predicted magnitude also depends on assumptions about length scales, path duration and which parts of an object respond to a gap. A mathematical calculation inside a model does not establish that the effect exists in nature.
The paper proposes a useful scientific discipline: define the prediction before taking the measurement. Compare the same discs using spacers of about 2 millimetres and shims of about 20 micrometres. Use the same mass comparator, preferably under controlled vacuum conditions. Specify the expected signal, uncertainty and failure criterion in advance. Independent laboratories should repeat the experiment.
A null result would constrain or reject the specific version of the proposal being tested. A positive result would require replication and checks for ordinary causes, including geometry, temperature, air currents, magnetic effects, contact forces and calibration. Until independent results support the claim, no Mars transport plan should assume a reduction in weight or launch energy.
9. The human settlement is part of the engineering
A settlement is not a collection of machines. It is a community that must maintain those machines, make decisions and care for its members. The source paper identifies several unresolved human questions: radiation exposure, life at 0.38 g, long-term reproductive health, isolation, delayed communication and the loss of familiar institutions.
Radiation is a major constraint during transit and on the Martian surface. Shielding with local material may reduce exposure, but it adds requirements for excavation, structure and access. Partial gravity raises another uncertainty. The source paper cites animal studies, but long-term human outcomes at Martian gravity remain unknown. Short experiments cannot establish the safety of living and reproducing at 0.38 g for decades.
Isolation and group dynamics also require planning. Antarctic stations and controlled-environment studies provide useful evidence, but they are not identical to a permanent settlement on Mars. Communication delays make direct supervision from Earth difficult. Medical care, education, conflict resolution and governance must work locally.
The source paper discusses a founding group of around 110 adults as a possible labour minimum, based on secondary reporting of demographic analysis. That figure should not be read as a complete design for a resilient society. Skills, health, redundancy, age structure, social compatibility and emergency capacity all matter. The population model must be tested against the work that actually needs to be done.
Biological diversity also matters. Plants, microbes and other organisms may be necessary parts of food production, waste treatment and nutrient cycling. Research on frozen mouse sperm and embryos in space offers limited evidence about biological preservation. It does not demonstrate safe human reproduction on Mars. That remains an open question requiring careful, staged research.
A credible plan should make human safety a design requirement. It should not treat people as interchangeable cargo or assume that local institutions will emerge automatically. A self-sustaining settlement needs both closed material cycles and a community capable of sustaining them.
10. A staged programme
The source paper’s practical sequence is to carry out the cheapest tests before committing to expensive launches. A revised programme can be organised as follows.
Stage 1 — Establish the baseline. Create a closure ledger for oxygen, water, food, nutrients, waste, energy, metals, polymers, electronics and spare parts. Record measured values separately from assumptions.
Stage 2 — Improve Earth-based life support. Run a sealed greenhouse and life-support system at realistic crew scale. Set targets and failure criteria before the trial. Publish the full material balance.
Stage 3 — Validate industrial processes. Test regolith processing, metals, ceramics, polymers, biological systems and repair methods. Include power demand, wear, maintenance and product quality.
Stage 4 — Compare architectures. Use common assumptions to compare transport options. Count interfaces, identify failure modes and test whether the ranking survives alternative reasonable counting rules.
Stage 5 — Test local manufacturing growth. Track each equipment class across successive generations. Verify that locally produced parts meet specifications and that the imported fraction is falling.
Stage 6 — Resolve human risks. Develop a staged research programme for radiation, partial gravity, medical care, reproduction, social organisation and emergency response.
Stage 7 — Send the industrial seed. Only after the enabling systems have met pre-defined requirements should the programme commit to a seed package. Match crew arrival to verified production capacity.
Stage 8 — Expand by evidence. Increase the settlement’s population and industrial capacity only as the closure ledger and operational record justify it. Treat each expansion as a new engineering decision.
Conclusion
The central issue is not whether a million tonnes can be transported to Mars. It is whether the resulting settlement can survive without further deliveries. Transport capacity is necessary, but it is not the same as independence.
The most practical route begins with measurement. Establish which systems are closed and which remain dependent on imports. Improve food and waste cycles on Earth. Develop local manufacturing around a compact industrial seed. Reduce unnecessary interfaces. Reuse landed hardware. Compare transport options with consistent assumptions. Advance the population only as life support, production and human safety permit.
The source paper’s combined model estimates that a 1,000-person settlement might need only a small fraction of the million-tonne plan if local production works as assumed. The estimate is worth investigating, but it is not yet a validated mission budget. Its strongest contribution is the question it puts first: what must still come from Earth, and why?
The proposed weight effect belongs in a separate research track. It has a defined experimental question, but its physical basis remains open. It should neither be presented as established science nor used to justify a transport architecture. The settlement programme can make substantial progress without it.
A robust Mars programme should therefore be steered by verified capabilities. Count the loops that work, the parts that can be made locally, the failures that have been removed and the generations of equipment that have been demonstrated. Tonnage and cost should follow from those results. The goal is not simply to deliver a city. It is to build one that can continue to function when Earth is no longer supplying it.
Annotated reference list
The annotations below explain the role of the sources in the argument. As in the source paper, some entries are cited through secondary reports. Such sources should be checked against the original publication before being used to support a critical engineering decision.
A. Theoretical framework and methods
Konstapel, J. (2026). The Vacuum.Net Theory: Definition, Theorems and Tests, version 4. Constable Research, Leiden. The source paper’s conceptual framework. It defines closure, seams, scale and the theory’s tests. It is a primary source for the author’s own theory, not independent confirmation of its physical claims.
Konstapel, J. (2026, 25 September). Applied Vacuum Theory: Setting the State of the Net. constable.blog. https://constable.blog/2026/09/25/applied-vacuum-theory/ Develops the proposed relation between coupling and the local state of the net. The paper itself labels parts of the bridge as assumed. Read the status ledger before treating the proposed relation as a physical result.
Konstapel, J. (2026, 28 September). The Vacuum.Net Theory: Complete Derivation. https://constable.blog/2026/09/28/the-vacuum-net-theorycomplete-derivation/ Presents the broader derivation behind the theory, including the distinction between passage through a mesh and opening a closure. Useful for understanding the framework’s internal logic.
Konstapel, J. (2026, 9 October). Transmutation: Elements on Demand. https://constable.blog/2026/10/09/transmutation-elements-on-demand/ Explores the idea of an element as an address plus a state and compares energy scales. The numerical interpretation should be checked against the derivation and stated units.
Watson, G. N. (1939). “Three triples of the lattice.” Quarterly Journal of Mathematics, 10, 266–276. https://doi.org/10.1093/qmath/os-10.1.266 Mathematical background cited for a three-dimensional lattice-return probability used in the paper’s path model. It supports a mathematical input, not the proposed gravitational interpretation.
Pólya, G. (1921). “Über eine Aufgabe der Wahrscheinlichkeitsrechnung betreffend die Irrfahrt im Straßennetz.” Mathematische Annalen, 84, 149–160. https://doi.org/10.1007/BF01458701 Classic work on random walks and returns to previously visited sites. Relevant to the mathematical model used by the Vacuum.Net Theory.
CODATA / NIST (2018). Fundamental Physical Constants. https://physics.nist.gov/cuu/Constants/ Reference values for physical constants used to define the theory’s natural units. These constants are standard inputs; their use does not independently validate the theory built from them.
Space.com (January 2020). “Elon Musk’s Starship flights to Mars colony.” https://www.space.com/elon-musk-starship-spacex-flights-marscolony.html Reports the concept of roughly 1,000 ships departing over about 30 days every 26 months. It describes a planning concept, not a demonstrated launch cadence.
Musk, E. (2017). “Making Humans a Multi-Planetary Species.” New Space, 5(2). https://static.ukrinform.com/files/1497612705-9227.pdf A primary presentation of an earlier Mars transport and settlement architecture. The source paper uses its estimates for cost per tonne and tanker flights. These are historical design estimates, not current guaranteed costs.
Maiwald, V. et al. (2024). Scientific Reports, 14, 11804, with correction. https://pmc.ncbi.nlm.nih.gov/articles/PMC11116405/ An independent mass-budget study cited as finding no feasible return scenario under its analysed assumptions. Useful as a counterweight to optimistic transport projections; its scope and correction should be read directly.
C. Closed life support and food production
Fu, Y. et al. (2021). “Lunar Palace 365.” bioRxiv. https://www.biorxiv.org/content/10.1101/2021.01.12.426282v1.full-text Reports a human-in-the-loop bioregenerative life-support experiment. The source paper uses its closure figures to identify food and solid waste as important remaining dependencies. Check the study’s definitions and accounting boundaries before comparing percentages.
Johnson, C. et al. (2021). Frontiers in Astronomy and Space Sciences, 8, 734343. https://www.frontiersin.org/journals/aerospace-and-space-sciences/articles/10.3389/fspas.2021.734343/full Provides estimates of crop area required per person in space-settlement contexts. The source paper compares its estimate with a linear extrapolation from Lunar Palace. Actual area depends on diet, lighting, crop choice and system design.
D. Local manufacturing and Martian resources
Metzger, P. et al. (2016). arXiv:1612.03238. https://arxiv.org/abs/1612.03238 Models a self-expanding manufacturing system using a compact initial industrial package. It is central to the source paper’s seed-growth calculation. Applying the model to Mars requires new assumptions about power, materials, reliability and production rates.
Chirikjian, G., Zhou, Y. and Suthakorn, J. (2002). Robotics and manufacturing without conventional microprocessors. https://rpk.lcsr.jhu.edu/wp-content/uploads/2014/08/Chirikjian02_b.pdf Explores simple robotic systems and manufacturing methods that reduce dependence on advanced electronics. Relevant to designing equipment that can be repaired or reproduced with limited industrial resources.
Menezes, A. et al. (2015). Journal of the Royal Society Interface, 12, 20140715. A quantitative source on potential mass savings from biologically produced fuels, food and polymers. Useful for evaluating biological manufacturing, though each process requires its own operating assumptions.
Cockell, C. et al. (2020). “BioRock.” Nature Communications, 11, 5523. https://www.nature.com/articles/s41467-020-19276-w Studies biological interaction with rocks and the potential for biomining. The source paper cites it as evidence for a possible route to obtaining metals from basalt under simulated conditions.
Wamelink, G. et al. (2014). PLOS ONE, 9, e103138. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0103138 Reports plant growth experiments using Mars and Moon soil simulants. It supports further research into regolith-based cultivation, but does not establish that untreated Martian soil is safe or sufficient for crops.
Wan, L., Wendner, R. and Cusatis, G. (2016). “Sulphur concrete for Mars.” arXiv:1512.05461. https://arxiv.org/abs/1512.05461 Examines sulphur-based concrete using Martian regolith simulant. It provides a candidate construction material and a basis for further durability and process testing.
E. Propellant, water and transport infrastructure
Hoffman, J. et al. (2022). MOXIE results. Science Advances, 8, eabp8636. https://pmc.ncbi.nlm.nih.gov/articles/PMC9432831/ Reports oxygen production from the Martian atmosphere using MOXIE. It demonstrates a key process at a small scale, but the source paper correctly treats settlement-scale production as requiring substantial scale-up.
Dundas, C. et al. (2018). “Exposed subsurface ice sheets in the Martian mid-latitudes.” U.S. Geological Survey. https://pubs.usgs.gov/publication/70195019 Documents exposed subsurface ice deposits. It supports the possibility of accessible water at selected sites, not the assumption that usable ice is available at every landing location.
Morgan, G. et al. (2021). “The Subsurface Water Ice Mapping (SWIM) project.” Nature Astronomy, 5, 230. https://pubs.usgs.gov/publication/70226786 Maps evidence for shallow subsurface ice in potential Mars landing regions. Useful for linking resource geography to landing-site selection.
Burke, L., Falck, R. and McGuire, M. (2010). NASA/TM-2010-216764. https://ntrs.nasa.gov/api/citations/20100037210/downloads/20100037210.pdf NASA technical analysis of departure and arrival energy. The source paper uses it in its discussion of transfer requirements and aerocapture. The relevant trajectory and assumptions should be checked for each mission.
Drake, B. (2009). Mars Design Reference Architecture 5.0. NASA. https://ntrs.nasa.gov/api/citations/20090010571/downloads/20090010571.pdf Compares chemical, nuclear-thermal and in-situ resource utilisation options. Provides a useful architecture-level reference rather than a prediction of present-day costs.
Ishimatsu, T., de Weck, O. et al. (2016), reported via Space.com. https://www.space.com/30838-manned-mars-mission-moon-refueling.html Secondary reporting of a study on lunar water refuelling and possible reductions in Mars mission launch mass. The primary paper was not retrieved in the source paper, so the headline percentage should be treated cautiously.
StarTram (2010). StarTram technical concept. https://startram.com/wp-content/uploads/2020/12/StarTram2010.pdf Presents an electromagnetic launch concept. It is useful for comparing alternatives, but its assumptions about infrastructure, materials and deployment must be assessed separately.
Birch, P. (1982). “Orbital Rings, Part I.” https://nss.org/wp-content/uploads/Orbital-Rings.pdf A foundational orbital-ring proposal. Relevant to long-term infrastructure comparisons, but its construction and deployment requirements are far beyond a simple launch-cost comparison.
Zhang, J. et al. (2022). Journal of Geophysical Research: Planets, 127, e2021JE007157. Studies effective radiation dose on Mars and shielding depth. Relevant to habitat siting and regolith shielding.
Tsuji et al. (2026). Science Advances, 12(11), eaed2258. https://www.uri.edu/news/2026/03/do-we-have-the-muscle-for-mars/ Reports partial-gravity-related animal research cited by the source paper. It is relevant to the need for further study, but cannot establish long-term human outcomes at 0.38 g.
Richter, C. et al. (2017). Frontiers in Physiology, 8, 583. https://www.frontiersin.org/articles/10.3389/fphys.2017.00583/full Reviews the evidence on partial gravity and human physiology. It supports the conclusion that long-term effects at Martian gravity remain uncertain.
Salotti, J.-M. (2020). Scientific Reports, 10, 9700. https://www.nature.com/articles/s41598-020-66740-0 Analyses minimum population requirements for a settlement. The source paper notes that its approximate 110-person figure was encountered through secondary reporting, so the number should be checked against the original analysis.
Wakayama group (2023). iScience; reported by New Atlas. https://newatlas.com/space/mouse-embryos-space-first-time/ Reports development of frozen mouse embryos to the blastocyst stage after space exposure. The source paper notes that human application remains open.
Hassler, D. et al. (2014). Science, 343; reported by Scientific American. https://www.scientificamerican.com/article/radiation-on-mars-managea/ Reports Mars surface radiation measurements. The source paper used a secondary account because the original article was not directly available to its authors.
Duan, X. et al. (2016). arXiv:1602.06377. https://arxiv.org/abs/1602.06377 Examines free fall in relation to spin state. Cited as a constraint on possible spin-dependent effects.
Heckel, B. et al. (2008). arXiv:0808.2673. https://arxiv.org/abs/0808.2673 Reports a spin-polarised torsion-pendulum test. Relevant to limits on anomalous spin-related interactions.
Rosi, G. et al. (2017). arXiv:1704.02296. https://arxiv.org/abs/1704.02296 Tests equivalence-principle behaviour involving internal energy states. Relevant to whether internal state can alter free fall.
Schlamminger, S. et al. (2008). arXiv:0712.0607. https://arxiv.org/abs/0712.0607 Reports a comparison of beryllium and titanium free fall. It constrains composition-dependent deviations from the equivalence principle.
Touboul, P. et al. (2022). MICROSCOPE mission results. https://arxiv.org/abs/2209.15487 Provides high-precision tests of the equivalence principle for different materials. It is a key experimental constraint on composition-dependent gravity.
Lee, J. et al. (2020). arXiv:2002.11761. https://arxiv.org/abs/2002.11761 Reports constraints on additional interactions at short distances. Useful for establishing which ranges of parameter space have been tested.
Chen, Y.-J. et al. (2016). arXiv:1410.7267. https://arxiv.org/abs/1410.7267 Examines short-range forces at sub-millimetre scales. The source paper uses it to distinguish experimental limits from the gravitational-strength regime relevant to its proposal.
Berry, J. et al. (2013). NPL Report ENG 46. https://eprintspublications.npl.co.uk/5985/1/ENG46.pdf Describes precision weighing of stacks of metal discs and a solid cylinder. It is the closest existing measurement cited in the source paper for the proposed gap experiment. The objects were adjusted to the same mass, so the original measurement does not by itself establish a gap-dependent weight effect.
Calloni, E. et al. Archimedes project. https://arxiv.org/abs/1511.04269 Describes an experiment designed to measure changes in weight associated with a material-state transition. Relevant as a measurement approach, but not evidence for the specific gap mechanism proposed here.
Moddel, G. et al. (2021). Symmetry, 13, 517. https://arxiv.org/abs/2101.03085 Reports measurements involving electromagnetic cavities and small power outputs. The source paper notes the authors’ caveats about plasmonic effects. This work does not establish a gravitational effect.
ALPHA-g Collaboration (2023). Nature. https://www.nature.com/articles/s41586-023-06527-1 Reports a measurement of antihydrogen’s response to gravity. It is relevant to the broader question of how antimatter falls.
Woods, R. et al. (2001). AIAA 2001-3363, via arXiv:physics/0204031. https://arxiv.org/abs/physics/0204031 Reports replication work on claims of gravity shielding by superconductors. The source paper cites the null or very small results as a reason not to rely on such claims.
Faller, J. et al. (1990). Physical Review Letters, 64, 2115. Reports measurements involving spinning gyroscopes and alleged weight changes. Cited as evidence that the reported effect was not established at the claimed magnitude.
Europa’s batterij-industrie worstelt met een afhankelijkheid van Chinese machines en technici, wat leidt tot faillissementen en vertragingen bij fabrieken zoals Northvolt en ACC.
Nederland kiest daarom voor een nichestrategie: het ontwikkelen van eigen productieapparatuur, componenten en onderzoeksconsortia via BCC-NL en het groeifondsprogramma ‘Material Independence & Circular Batteries’.
De blog analyseert deze inspanningen met een theoretisch ‘vacuum net’-model, dat de fundamentele grens van alle batterijchemie beschrijft: elke energiedichtheidswinst gaat gepaard met een grensvlak dat onder spanning staat en kan breken.
Het model voegt toe dat batterijen geen energie ‘bevatten’, maar een tijdelijke toestand zijn die meetbaar is via een ‘hold time’-ladder, en biedt concrete ontwerpaanwijzingen om grensvlakken te stabiliseren.
On 16 October 2026, Het Financieele Dagblad described Europe’s attempt to make batteries without China. The article’s conclusion was blunt: the difficulty is not the chemistry, it is the machines. European battery factories buy their production equipment from China, and the knowledge needed to run that equipment comes with it — in the form of Chinese technicians who stay on the factory floor.
The record supports this. Northvolt bought about 1,200 machines from the Chinese supplier Wuxi Lead, kept hundreds of Chinese technicians far longer than planned, and went bankrupt in March 2025. ACC still struggles with scrap rates; in September 2026, Stellantis halted two French car plants for lack of ACC cells, and ACC flew in a team of experts from China. PowerCo runs its Salzgitter plant almost entirely on Chinese machines, with roughly 200 supplier specialists on site. Verkor lost its anchor customer after eighteen months of delay.
The numbers behind the headlines are stark:
China made more than 80 percent of the world’s battery cells in 2025 (IEA).
In Europe, 98 percent of cell capacity is owned by Asian companies (Deloitte, 2025).
Battery packs in China were about 35 percent cheaper than in Europe in 2025 (IEA).
A quarter of the European cell capacity announced since 2022 has been cancelled or shelved (Transport & Environment).
Since 9 October 2025, the problem has had a legal edge as well. China’s MOFCOM Decision No. 58 places under export licence: cells of 300 Wh/kg and above, cathode material and artificial graphite, and the machines for winding, stacking, filling, hot pressing and formation. The licence requirement was suspended until 10 November 2026. Whether it stays suspended is a decision made in Beijing, not in Brussels.
2. What the Netherlands is already doing
The Dutch response was not improvised yesterday. In 2022 the national Action Agenda set a goal: by 2030, Dutch machine builders supply the European gigafactories. The cabinet deliberately chose components, machines and niches — not a domestic cell gigafactory. A letter to parliament of February 2026 repeats that choice. The FD article describes the execution of a decision taken four years ago.
The network behind it:
BCC-NL in Helmond is the hub, founded in 2021 by DAF, VDL, Damen, ELEO, TNO, TU/e, Brainport Development, RAI and NMT. Its director since March 2025 is Mustafa Amhaouch, formerly twenty years at ASML.
The growth-fund programme “Material Independence & Circular Batteries” runs from 2024 to 2031 with a budget of at most €296 million, across three lines: raw materials and recycling; materials, components and equipment; and systems for transport and grid. Twenty-two projects have been running since March 2026.
National research consortia cover the field: BatteryNL (the electrode-electrolyte interface, €9.3 m, TU Delft), ANION (nanoscale electrochemical interfaces, €23.6 m, Leiden), NANEXBAT (sodium-ion, €8 m, Twente), SMART-CBAT (materials, processes and equipment, €22.3 m, TU Delft), SLDBatt (8–100 hour storage, >€30 m, Twente), CIMBATT (heavy transport, €30 m, VDL), CRM-Lion (critical material recovery, €23 m, TNO), and 10X-Factor(y) (domain walls and memristors, €30 m, Groningen).
The Dutch research effort divides into eight lines: silicon and carbon anodes (LeydenJar, E-magy, CarbonX, LionVolt); solid state and electrolytes (Thionics, Twente, Utrecht, Delft, Eindhoven); sodium-ion (NANEXBAT, Alfen with CATL in Almere); long-duration storage (Ore Energy, Elestor, AquaBattery); production equipment (SALD, SparkNano, Kalpana, Powall, TNO Holst Centre); recycling (SK tes, Back to Battery); the interface inside the cell (BatteryNL, ANION); and storage in a wall (Twente, Groningen).
A few people hold the network together — Marnix Wagemaker (Delft), Mark Huijben (Twente) and Moniek Tromp (Groningen) each sit in several consortia; André Schilt connects TNO Holst with BCC-NL, and Holst has spun off LionVolt, LeydenJar and SALD.
The gaps are just as telling. No Dutch party was found working on: inline measurement on foil and coating at production scale (the skill exists in chip metrology — Nearfield raised €330 million in June 2026, but for chips); current without applied voltage in an asymmetric cavity; or logic and storage on three values (+1, 0, −1). Lithium-sulfur and quantum batteries are also absent.
3. The model: the vacuum net in plain terms
This essay reads the battery through a theoretical model — the vacuum net — and asks what follows for how batteries are made, measured and used.
Imagine a fishnet under tension. Strands run through it and cross; where windings hold each other there is a knot; between strands there are meshes — openings with a size and a direction. The vacuum net theory reads the vacuum itself as such a net, built on five axioms:
Strand. There is one strand; every distinction is the strand touching itself.
Closure. A configuration persists only if it closes on itself in phase.
Memory. The net keeps the orientation of what was laid down first.
Context. Constants are state functions of the local net.
Scale. The same winding appears at every depth.
Three objects follow. A winding is a strand returning on itself in phase: that is energy. A knot is windings holding each other in three directions: that is mass. A mesh is an opening between strands. Power, being energy per unit of time, scales with one over the open size of a mesh.
Two rules do the work in this essay:
The window rule. Each crossing of strands carries a trit: +1, 0 or −1. Three neighbouring crossings form a window. A window closes when the sum of its values is divisible by three — 9 of the 27 possible windows close; the other 18 stand open under tension. What closes, stays.
The mesh energy. One mesh of size L holds an energy of π²ℏc/(720 L). At the natural mesh size of 45.588 nm, that comes to 9.51 × 10⁻²¹ J.
What a battery is, in this model
In everyday speech a battery “contains” energy. In the net it does not. Energy is a closed motion, not a substance in a jar. A battery is a state that is held for a time — and a hold time can be measured.
That measurement produces a ladder. Each rung has its own carrier:
Hold time
Carrier
Dutch work
Seconds
Flywheel, supercapacitor, polar wall
Twente (dielectric films); Wang, Delft (MXenes)
Hours
Lithium-ion, sodium-ion
LeydenJar; NANEXBAT; Alfen
Days
Iron-air, salt, flow
Ore Energy; Elestor; AquaBattery; SLDBatt
Months
Molecules
Hydrogen pilots
No rung replaces another. The public grid fuses all phases into one — what is generated is used at the same instant, hold time zero. A cell asked to bridge days does the work of another rung, and that is why it grows heavy, expensive and dependent on scarce material.
The one limit every line shares
The lines of battery research differ in chemistry, but they meet the same limit. Every gain in energy density comes from an electrode that changes shape, and the boundary layer must follow that change without breaking:
Silicon anode: volume grows by more than 300 percent; the interphase cracks and re-forms every cycle.
Lithium metal: lithium plates unevenly and grows along cracks.
Solid state: solid-solid contact is lost; filaments grow along grain boundaries.
Lithium-sulfur: intermediate products dissolve away from the cathode.
Magnesium, calcium: the boundary layer seals and stops passing the ion.
Each boundary joins two structures that close differently — crystal and liquid, metal and ceramic, lattice and film. By the window rule, such a boundary is an open window: it stands under tension and does not stay unless something keeps closing it.
The rule does not say why a seam breaks. It says which configurations persist. That is enough to turn the seam into three design questions: Where is the seam? Can it be laid so that fewer windows stay open? And can the motion take place in a mesh — an open space — rather than in the seam itself?
4. The model against the measurements
The rules were set against numbers that Dutch and international groups have published.
Closure orders crystal orientations. Crystal directions are written with three Miller indices; for the low-index directions these are trit triples, so the window rule applies directly. Of the ⟨100⟩ family (6 members), none close; of ⟨111⟩ (8 members), 2 close; of ⟨110⟩ (12 members), 6 close. An open direction carries passage; a closed direction holds. The rule therefore predicts transport through a film in the order ⟨100⟩ > ⟨111⟩ > ⟨110⟩.
The measurements agree:
Lithium titanate (Cunha et al., 2019), at 3C: (100) about 313 mAh/g, (111) about 283, (110) about 277 — exactly the order of the open fraction.
Lithium manganese oxide (Huijben et al., Twente, 2022): after 1,000 cycles at 5C, (100) keeps 55 percent and (110) keeps 33 percent, with more surface-held charge in the closed direction. An earlier Twente series (Hendriks et al., 2018) also found (100) best.
The coupling at a wall is a quantity of that wall. The theory’s coupling law reads Δg/g = γδ, where δ is the relative change in density and γ is a state function of the wall, not one universal number. Palasantzas’s group in Groningen measures Casimir forces on phase-change films:
GST: when the film crystallises, the force rises by about 40 percent below 100 nm, while density goes from 5.87 to 6.27 g/cm³ (δ = 0.068). Measured γ: 5.9.
AIST: force contrast of 17–25 percent against densification of 5.2 percent. Measured γ: 3.3 to 4.8.
Passage scales with one over the gap. Garret Moddel’s group in Colorado measured current without applied voltage across metal-insulator-metal junctions carrying optical cavities from 33 to 1,100 nm. The current rose as the cavity grew thinner. Maclay’s fit (2024) shows conductance proportional to 1/L — the exponent of the passage rule.
A polar wall holds about one mesh energy per lattice cell. The Twente multilayer of Nguyen, Koster and Rijnders stores 128.4 J/cm³ at 81 percent efficiency. With a lattice parameter of 4.15 Å, that is 9.18 × 10⁻²¹ J per cell — 0.97 of one mesh energy. The best films reported up to October 2026 hold between one and 1.65 mesh energies per cell.
Room instead of grip. LeydenJar’s silicon columns, E-magy’s pores and LionVolt’s 3D pillars all leave room for the swelling — in the fishnet image, the motion goes into the mesh, not into the knot. Single-crystal cathodes, which have no internal seams, outlast polycrystalline ones by a wide margin (Dalhousie–Tesla, 2022).
Where the readings stand:
Reading
Status
9 of 27 windows close; 0/6, 2/8, 6/12 per family
counted, exact
Orientation order (100) > (111) > (110)
agrees in two spinels
γ at a wall is a quantity of that wall
GST 5.9; AIST 3.3–4.8
Conductance ∝ 1/L
agrees with Moddel’s series
Polar walls hold about one mesh energy per cell
Twente 0.97; best films up to 1.65
A battery boundary is an open window
reading
Life follows the count of open windows
hypothesis, not yet measured
5. What the net adds: six contributions
The net places the limit of the battery at the seam — and the Netherlands is strongest exactly there. BatteryNL is devoted entirely to the electrode-electrolyte interface; Twente grows crystal films one atomic layer at a time and measures them while they grow; Delft watches lithium move through the interface with neutrons; Eindhoven has the densest cluster of thin-layer equipment builders in Europe; and the chip supply chain has made the country a world leader in measuring and controlling thin layers. That is a different strength from the “paint robot” in the FD headline — and it is the strength the net speaks to.
1. Ranking the seam before it is grown. The window rule ranks crystal orientations by their open fraction before a single film is grown, and it got the order right in two spinels. One experiment would turn this into a design tool: grow all three orientations of one electrode at equal surface area and measure life and rate. Huijben’s group has the films and the pilot line; Wagemaker and Bannenberg in Delft have the instruments to watch the seam while the cell runs. The same count extends to any interface whose two sides can be written as trits — every new coating, interlayer or epitaxial stack gets a ranking at no cost.
2. The coupling of a wall as a design quantity. With γ known per wall, the change in coupling follows from a density measurement alone. That makes γ measurable for any wall that changes density: a phase-change film, an interphase that densifies during formation, a coating that compacts. Palasantzas has the force set-up; Wagemaker’s neutron depth profiles give the density of a plated layer during operation. Together they would give γ for the interphases that decide a cell’s life.
3. Making without the heat step, outside the licence list. A conventional electrode is a slurry coated on foil and dried in long ovens. The drying step alone costs 11.02 kWh per kWh of cell capacity and 27–48 percent of all factory energy; counting yield, each good kWh carries up to 58 kWh of factory energy — a cell must be fully discharged 23 to 58 times before it has returned the energy of its own production.
The net describes making as a three-step cycle: soften (bring the local state near its tipping point), write (impose a mesh, a layering, an asymmetry), harden (return the state so that only closing windings remain). Hardening selects; it does not place. Processes that work layer by layer, at low temperature, follow this cycle directly.
The machines China has licensed belong to one process order: mix, coat, dry, press, wind or stack, fill, form. Four process families do not appear on the list — plasma deposition (LeydenJar), spatial atomic layer deposition (SALD, SparkNano, Kalpana, TNO Holst Centre), thin-film cells (Huijben, Energy01), and electrodeposition. Each is already Dutch. The FD plan rebuilds the existing order from Dutch coaters and laminators, but that order still contains the licensed steps. A second route chooses the processes that do not need them. It is less mature, it touches the Chinese list less, and it plays directly to Dutch strength.
4. The loop instead of the hand. The FD article names the real difficulty: a machine does not run by itself; the knowledge sits in the people who tune it. Fraunhofer FFB puts first-years gigafactory scrap at 15–30 percent, and each percentage point costs about €30,000 per day — €10.95 million per point per year. At 20 percent scrap that is €219 million a year.
In the net, a state that holds itself is a closed loop. A loop that moves one setting until the output is at its maximum, then holds it there, needs no operator. Where the state of each layer is measured inline and the setting follows, the knowledge sits in the loop — not in the hands of technicians flown in with the machine. This is the thin spot of the Dutch network: the skill exists in chip metrology, the battery application is missing. Joining the two would address the cause of Northvolt’s and ACC’s troubles directly.
5. The wall that holds in three states. A ferroelectric wall has three polarisation states — P+, P0, P−: a trit. The Twente films already store about one mesh energy per lattice cell, and no ion moves house, so there is no seam that changes with every cycle. Twente and Groningen (10X-Factor(y), €30 million) hold the materials. What the net adds is the register: a wall with three states can hold and pass on +1, 0 and −1 directly. Storage and logic on three values are not yet worked on in the Netherlands.
6. Passage instead of stock. One mesh at the natural size holds 9.51 × 10⁻²¹ J; one passage per natural time unit gives 6.25 × 10⁻⁵ W per mesh. Moddel’s patent reports 1.4 pW from an active junction of 0.02 µm² — about 70 W/m², or a coupling η of 2.3 × 10⁻⁹ of the full mesh rate. That η, the share of what the mesh closes that leaves as current, is the one parameter that decides how far this route goes. No Dutch group works on it; Palasantzas, AMOLF and Schall have the cavities, the walls and the tipping point as a knob.
6. The future if the readings hold
Two conditions carry the boldest scenario: Moddel’s 70 W/m² must be reproduced, and stacked layers must add. The rest rests on what is already measured.
Less is asked of the cell. When storage follows the ladder of hold times, each rung does its own work. The cell no longer bridges days; it becomes smaller, lighter and less dependent on cobalt, nickel and graphite. A good part of the demand that drives gigafactories falls away, and storage moves from a remote park to the street, the house and the district.
The seam is designed, not suffered. A cell is built from the seam outward. Orientations, interlayers and coatings are ranked before they are grown; the coupling of each wall is known from its density. Life becomes a design target instead of a test outcome after years. Single crystals, epitaxial films and columns with room for motion become the rule rather than special cases.
Factories without fire, near the user. Making by soften-write-harden needs no ovens tens of metres long and no dry rooms. A line that deposits layer by layer at low temperature is small; it fits in a city rather than on the edge of a country. The gigafactory, built around the oven and the formation hall, is no longer the natural form of making a cell. The scale of production follows the scale of use.
The knowledge sits in the loop. Where every layer is measured as it forms and the setting follows, the line holds itself. The knowledge no longer travels with foreign technicians. Scrap falls — and with it the €10.95 million per percentage point per year. A licence decision in Beijing then reaches the spare parts of a machine, not the know-how of a line.
The wall holds and computes. A wall with three states holds energy without moving ions, and holds a trit. Storage and register become one structure, and the fastest rung of the ladder is served by walls that do not wear.
Passage instead of stock. If the passage holds, the change is largest. A house needing 1,450 W continuously needs 20.7 m² of one layer at 70 W/m² — or a plate of about 45 by 45 cm with 100 layers. Such a plate delivers 34.8 kWh per day, where a battery holding one such day at 250 Wh/kg weighs 139 kg. A drone needing 1 kW needs 0.14 m². The battery then stops bridging energy; it only catches peaks — a small buffer in the meter cupboard, the drone, the tool.
What it means for people. In Eindhoven, laboratory staff at LeydenJar assemble test cells by hand all day, at a 95 percent pass rate; in a line that holds itself, their work moves from tuning by feel to reading a state. In Skellefteå, a town of 75,000 built its future on Northvolt and lost 1,000 jobs in one round of cuts, beside technicians from Wuxi who lived in their own camp; a factory without fire, near the user, is less likely to tie a whole town to one plant. In Debrecen, residents raised concerns about CATL’s plant, and in early 2026 the authorities closed parts of it after nine employees were exposed to nickel; a cell asked to do less carries less nickel. In Ukraine, the drones soldiers carry run mostly on Chinese cells, and a licence decision reaches their front line; a plate that delivers continuously would end that dependence. In the household, the base load would no longer depend on a cable, a substation or a battery park — what disappears is the long chain of mine, refinery, gigafactory, ship and grid; what remains is the house and its plate.
Fred Gao, Inside China (9 October 2025). “China Implements New Export Controls on … Lithium Batteries, and Artificial Graphite Anode Materials.”https://www.fredgao.com/p/china-implements-new-export-controls Why read? The list of controlled machines, step by step. Basis for section 5, contribution 3.
Konstapel, J. (2026). “Applied Vacuum Theory, Parts Two to Four.” constable.blog, 25–27 September.https://constable.blog/2026/09/25/applied-vacuum-theory/ Why read? Soften-write-harden, the self-tuning loop and power = 1/S. Reading advice: Part Four, on energy without a reservoir.
Konstapel, J. (2026). “What the Path Gives and What Its Description Gives.” Version 1, 16 October. Why read? Why γ is a state function and not one universal number. Basis for the reading of the Groningen films in section 4.
The measurements
Hassanzadeh, Atul, Wang, Kooi, Tajik, Palasantzas (2025). Physical Review B 112, 155403.https://doi.org/10.1103/zlvn-r3jy Why read? The GST force contrast of about 40 percent.
Torricelli, van Zwol, Palasantzas et al. (2010). Physical Review A 82, 010101(R).https://arxiv.org/abs/1006.4065 Why read? The first switchable Casimir force, with AIST.
Njoroge, Wöltgens, Wuttig (2002). Journal of Vacuum Science and Technology A 20, 230. Why read? The densities of amorphous and crystalline GST that give δ.
Hendriks, Huijben et al. (2018). ACS Applied Energy Materials.https://doi.org/10.1021/acsaem.8b01477 Why read? LiMn₂O₄ in all three orientations; (100) best.
Cunha et al. (2019). Orientation-dependent Li₄Ti₅O₁₂ thin films. Why read? The three-orientation capacity series that follows the closure order. Reading advice: compare at the same C-rate.
Nguyen, Birkhölzer, Houwman, Koster, Rijnders (2022). Advanced Energy Materials 12, 2200517.https://doi.org/10.1002/aenm.202200517 Why read? 128.4 J/cm³ in a Twente multilayer: one mesh energy per lattice cell.
Lv, Wagemaker et al. (2018). Nature Communications. Why read? Plated lithium seen with neutrons during operation. The instrument for the density of a seam in contribution 2.
Zheng, Zhao, Archer et al. (2019). “Reversible epitaxial electrodeposition of metals in battery anodes.” Science 366, 645.https://news.cornell.edu/node/315386 Why read? An underlayer that sets the direction of growth; cycles rise tenfold.
Schütte, Degen, Walter (2024). Batteries 10(2), 64.https://www.mdpi.com/2313-0105/10/2/64 Why read? Energy per production step per kWh of cell. Source of the 11.02 kWh in contribution 3.
Moddel, Weerakkody, Doroski, Bartusiak (2021). “Optical-Cavity-Induced Current.” Symmetry 13, 517.https://arxiv.org/abs/2101.03085 Why read? The cavity series from 33 to 1,100 nm, with eight controls for artefacts.
Moddel, G. US Patent 12,166,434, “Enhanced quantum vacuum energy devices.”https://patents.google.com/patent/US12166434B2/en Why read? The 1.4 pW and the active area of 0.02 µm² behind the 70 W/m².
Maclay, G. J. (2024). “Modeling Electronic Devices with a Casimir Cavity.” Physics 6(3), 1124–1131.https://doaj.org/article/31210466294848b0b30d3380b0056c2c Why read? The fit of conductance against one over thickness. Maclay questions the proposed mechanism and leaves the measurement standing.
Hertlein, Helden, Gambassi, Dietrich, Bechinger (2008). Nature 451, 172. Why read? Critical Casimir forces measured directly. The tipping point as a knob, which Schall uses in Amsterdam.
Europa wil eigen batterijen. Op 16 oktober 2026 beschreef Het Financieele Dagblad hoe het continent probeert los te komen van China: het Nederlandse LeydenJar dat in Eindhoven een celfabriek bouwt, de Europese gigafabrieken — ACC, Verkor, PowerCo, Northvolt — die hun machines in China kochten en ze niet aan de praat kregen, en een plan van het Battery Competence Cluster Nederland om uit bestaande Nederlandse machines een productielijn samen te stellen.
Achter dat artikel staat één hard feit. Op 9 oktober 2025 plaatste China een reeks artikelen onder exportvergunning (Besluit nr. 58): cellen vanaf 300 Wh/kg, dicht LFP-kathodemateriaal, kunstmatig grafiet, en de machines voor het wikkelen, stapelen, vullen, heetpersen en vormen. De vergunning werd geschorst tot 10 november 2026; of een latere Amerikaans-Chinese regeling, verlengd tot 10 januari 2027, ook dit besluit dekt, is niet bekendgemaakt.
De krant vraagt: hoe kopieert Europa de batterij? Dit essay stelt een andere vraag. Wat is een batterij als je er door de bril van het vacuum net naar kijkt — en wat volgt daaruit voor hoe we haar maken en gebruiken? Het antwoord valt uiteen in zes richtingen, die hier telkens naast het onderzoek worden gelegd dat al onderweg is. De rekenkunde wordt volledig gegeven, en een statusregister houdt aan het einde bij wat gemeten, afgeleid, gelezen of hypothetisch is.
1. Het net in een handvol regels
Beeld je een visnet in: draden die elkaar kruisen. Waar windingen elkaar vasthouden, ontstaat een knoop. Tussen de draden liggen mazen: openingen met een grootte en een richting. De vacuum net-theorie leest het vacuum als zo’n net onder spanning, gedragen door vijf axioma’s: er is één draad, en elk onderscheid is die draad die zichzelf raakt (slinger); een configuratie blijft alleen bestaan als ze in fase op zichzelf sluit (sluiting); het net onthoudt de richting van wat eerst werd neergelegd (geheugen); constanten zijn toestandsgrootheden van het lokale vacuum (context); en hetzelfde patroon keert op elke diepte terug (schaal).
Uit die regels volgen drie objecten. Een winding is een draad die in fase op zichzelf terugkeert: energie, met dimensie T/S. Een knoop is windings die elkaar in drie richtingen vasthouden: massa. Een maas is een opening tussen draden — geen materie, maar wel een grootte S.
Eén gevolg is direct belangrijk: vermogen is energie per tijd, dus (T/S)/T = 1/S. Vermogen door een maas schaalt met één gedeeld door haar open grootte. Nergens in die regel komt een reservoir voor.
Verder draagt elke kruising één van drie waarden: +1, 0 of −1. Drie naburige kruisingen vormen een venster. De vensterregel zegt: een venster sluit als de som van zijn drie waarden deelbaar is door drie. Van de 27 mogelijke vensters sluiten er 9 — drie homogene ((−1,−1,−1), (0,0,0), (+1,+1,+1)), statisch en kristalachtig, en zes heterogene (de volgordes van −1, 0 en +1), die roteren en een levendige beweging dragen. De overige 18 vensters zijn open: ze staan onder spanning. En nu de zin waar alles in dit essay op draait: de grens tussen een homogene en een heterogene sluiting is altijd een open venster. Wat sluit, blijft.
2. Wat een batterij is in het net
In het dagelijks taalgebruik “bevat” een batterij energie. In het net bestaat dat niet: energie is een gesloten beweging, geen substantie in een pot. Een batterij is een toestand die een tijdlang wordt vastgehouden. Zo gelezen heeft ze vier fasen — laden, vasthouden, ontladen, rusten — waarvan het gewone elektriciteitsnet er maar twee kent, tot één gesmolten: wat opgewekt wordt, wordt op hetzelfde ogenblik gebruikt. De houdtijd τ van het net is nul.
Houdtijden vormen een ladder, en elke sport heeft haar eigen drager:
Houdtijd
Drager
Stand der techniek, 2026
seconden
vliegwiel, supercondensator
30 MW-vliegwiel op het net; supercondensatoren met een miljoen cycli
uren
lithium-ion, natrium-ion
commercieel
dagen
ijzer-lucht, zand, warmte
eerste netwerksystemen (Form Energy, Ore Energy, Polar Night)
maanden
moleculen (waterstof)
zoutgrot-piloten
Geen sport vervangt een andere. Een cel die dagen moet overbruggen, wordt gevraagd het werk van een andere sport te doen — en wordt daarom zwaar, duur en afhankelijk van schaars materiaal. Een batterij van 200 MW in Winschoten die vier uur vasthoudt, is 800 MWh: één houd op één sport. Ze maakt geen ladder.
3. De naad
Eén muur voor elke lijn
De huidige onderzoekslijnen verschillen in chemie maar delen één limiet. Elke winst in energiedichtheid komt van een elektrode die van vorm verandert — en de grenslaag moet die verandering volgen zonder te breken:
Lijn
Wat er aan de grens gebeurt
silicium-anode
volume groeit met meer dan 300%; de grenslaag (SEI) scheurt en hervormt elke cyclus
lithium-metaal, anode-vrij
lithium plaatst ongelijk en groeit langs scheuren en korrelgrenzen
solid-state
vast-vast contact raakt verloren; lithium-filamenten groeien langs korrelgrenzen
lithium-zwavel
tussenproducten lossen op en verlaten de kathode
magnesium, calcium
de grenslaag sluit af en laat geen ion meer door
kwantumbatterij
de koppeling die laadt, ontlaadt ook
De naad in het net
Elke grens uit deze tabel verbindt twee structuren die verschillend sluiten: een kristal en een vloeistof, een metaal en een keramiek, een rooster en een film. Volgens de vensterregel is zo’n grens een open venster — hij staat onder spanning en blijft niet staan, tenzij iets hem voortdurend sluit. Dit is een correspondentie, geen mechanisme: de regel zegt niet waarom een naad scheurt, alleen welke configuraties blijven bestaan. Uit de lezing volgen drie ontwerpvragen: kan de cel zo worden gelegd dat er minder vensters open blijven staan, en kan de beweging in een maas — een open ruimte — plaatsvinden in plaats van in de naad?
Antwoorden die al bestaan
Ruimte in plaats van greep. LeydenJar laat silicium als zuiltjes op koperfolie groeien; de ruimte ertussen vangt de zwelling op. E-magy doet hetzelfde met poriën, LionVolt met een driedimensionale structuur. In het beeld van het visnet: de beweging gaat de maas in, niet de knoop.
Eén richting. Zheng, Archer en collega’s lieten zink epitaxiaal groeien op een dunne grafietlaag (Science, 2019). De afzetting werd omkeerbaar, en onder netwerk-schommelingen steeg het aantal stabiele cycli van ongeveer 200 tot meer dan 2.000. De onderlaag zet de richting, en de afzetting volgt. Het net heeft dezelfde regel voor ringen: een wand of zaad met de juiste richting geeft alleen de passende ring; een fles zonder richting geeft een mengsel.
Homogene sluiting. Kathodedeeltjes als enkelkristallen NMC532 overleven polykristallijne ruim. Een Dalhousie–Tesla-groep projecteerde levensduren richting een eeuw bij 25 °C (Journal of the Electrochemical Society, 2022). Een enkelkristal heeft geen korrelgrenzen, dus geen interne naden: de homogene, kristalachtige sluiting van de vensterregel.
Geen naad die moet houden. In een nikkel-waterstofcel is de ene kant een gas dat op een katalysator stuit. Er is niets dat kan scheuren. EnerVenue specificeert zulke cellen op 30.000 cycli.
Wat de regel nog niet geeft
De vensterregel telt open vensters, maar zet die telling nog niet om in een aantal cycli; het register van spanning dat grootheden zou moeten dragen, is nog open. De regel doet wél één ordeningsvoorspelling: neem een reeks grenslagen, tel hun open vensters en meet hun leven — de volgorde van het leven zou de volgorde van de telling moeten volgen. Die meting is niet gedaan.
4. Maken zonder vuur
Hoeveel vuur een cel draagt
Een conventionele elektrode is een pasta van poeder, bindmiddel en oplosmiddel, gesmeerd op folie en gedroogd in ovens van tientallen meters. Daarna: persen, snijden, stapelen, vullen met elektrolyt en de eerste lading — de vorming.
De energieprijs valt af te lezen aan één gepubliceerd cijfer. Schütte, Degen en Walter (Batteries, 2024) geven 11,02 kWh droogenergie per kWh celcapaciteit; drogen is in hun studie 27% van het fabrieksverbruik, dus de fabriek verbruikt in totaal 40,8 kWh per kWh cel. Andere studies stellen drogen op 47 tot 48%, wat neerkomt op 23,4 kWh. Het bereik is dus ongeveer 23 tot 41 kWh per kWh cel. Schroot maakt het erger: passeert 70% van de cellen, dan draagt elke goede kWh 40,8 / 0,70 = 58,3 kWh fabrieksenergie. Een cel van 1 kWh moet dus 23 tot 58 keer volledig worden ontladen voordat ze de energie van haar eigen fabricage heeft terugverdiend — grondstoffen nog niet meegerekend.
De vorming voegt tijd toe: drie tot zeven dagen, gevolgd door tot twee weken rijping; ze kan een kwart van de vloer en tot 30% van de maakkosten beslaan. En het gebruikelijke anodemateriaal, kunstmatig grafiet, wordt wekenlang gebakken op tot 3.000 °C.
Processen zonder de oven
Proces
Wat het weghaalt
Stand, oktober 2026
droge elektrode (Tesla, LEAD, Dürr)
oplosmiddel en droogoven
in productie in China en de VS; pilot in Europa
plasmadepositie (LeydenJar)
pasta, bindmiddel, oven
fabriek gepland voor 2027
ruimtelijke ALD (TNO Holst Centre)
laag-op-laag bij lage temperatuur, ca. 1 nm per seconde
pilot-apparatuur
elektrodepositie van kathodemateriaal (Xerion)
calcinatie
pilot
koolstofnetwerk uit raffinageresidu (CarbonX)
grafitering op 3.000 °C
opschaling
Verweken, schrijven, verharden
Het net beschrijft maken als een drieluik. Verweek de lokale toestand tot dicht bij haar kantelpunt. Schrijf er een maas in: een vorm, een maat, een gelaagdheid, een asymmetrie. Verhard de toestand, zodat alleen de sluitende windingen overblijven. Verharden selecteert onder configuraties die kunnen sluiten; het plaatst niets — het zet de toestand zo dat alleen het doel blijft staan.
De vorming past in de derde stap. Tijdens de eerste lading ontstaat de grenslaag: wat sluit, blijft; de rest wordt verbruikt. Snelle vormingsprotocollen bij Oak Ridge bekorten de stap minstens zesmaal — in de taal van de cyclus: een korter verharden.
5. De lus in plaats van de hand
De machine draagt de kennis niet
Het FD-artikel noemt de eigenlijke moeilijkheid: een machine loopt niet vanzelf. De kennis zit bij de mensen die haar afstellen.
Northvolt kocht circa 1.200 machines bij Wuxi Lead en had honderden Chinese technici nodig om ze te bedienen — die weken tot 70 uur maakten, waarvoor hun werkgever in Zweden werd beboet. Na vier jaar bereikte Northvolt ongeveer 70% opbrengst; een Chinees directeur zei dat zijn fabrieken 96% halen binnen vier maanden. ACC heeft schroot als voornaamste probleem; in september 2026 moest Stellantis twee Franse autofabrieken stilleggen om gebrek aan ACC-cellen, waarna ACC een team experts uit China liet overkomen.
De prijs: Fraunhofer FFB rekent 15 tot 30% schroot in de eerste jaren en ongeveer 10% na vijf jaar. Elk procentpunt kost ongeveer €30.000 per dag: 30.000 × 365 = €10,95 miljoen per jaar. Bij 20% schroot is het verlies 20 × 10,95 = €219 miljoen per jaar.
Een lus die haar vaste punt vindt
De toegepaste theorie beschrijft een zelfstemmende maas: een lus die één instelling verschuift tot de uitgang maximaal is, en haar daar vasthoudt. Ze corrigeert drift en leest geen plek, geen tijd, geen operator en geen materiaal. De lus vervangt de hand van de maker. Ze is ontworpen, nog niet gebouwd.
Op een productielijn overgezet verschuift het zwaartepunt van de machine naar meting en regeling: waar de toestand van elke laag inline wordt gemeten en de instelling volgt, zit de kennis in de lus — niet meer bij de operator. De vooraanstaande Chinese fabrieken werken al zo: CATL rapporteert defecten op parts-per-billion-niveau, met inspectie van 1,7 seconden per cel.
En Nederland is sterk precies in dit segment: meten, positioneren en gesloten regelkringen, van de lithografieketen tot metrologiebedrijven. Dat is een andere sterkte dan de verfrobot uit de krantenkop.
6. Vasthouden in drie toestanden
Een ferro-elektrische wand als bariumtitaat heeft drie polarisatietoestanden: P+, P0 en P−. In de theorie vormen die een trit. Het teken van de perovskiet-tolerantiefactor minus één geeft dezelfde drie waarden: +1, het centrum geeft sluiting op (ferro-elektrisch); 0, rust (kubisch); −1, rotatie neemt de misfit op. Twee onderzoeklijnen slaan energie op in zo’n wand, zonder ion die moet verhuizen.
Diëlektrische opslag. Relaxor-ferro-elektrische films slaan in het laboratorium 182 J/cm³ op bij 78% rendement (Nature Materials, 2022). Er is geen naad die met elke cyclus beweegt.
Kwantumbatterijen. Moleculen die samen aan een lichtveld zijn gekoppeld, laden sneller naarmate er meer zijn. CSIRO en partners toonden in maart 2026 de eerste elektrische ontlading uit zo’n apparaat, bij kamertemperatuur; de opslag hield ongeveer een miljoen keer langer stand dan het laden. De limiet: de koppeling die laadt, ontlaadt ook. Het werkende antwoord is een donkere toestand — één kant die niet terugstraalt. Het vacuum net gebruikt dezelfde asymmetrie voor zijn generatorelement: een gesloten kant en een open kant met bewust verschillende randvoorwaarden.
Geen van beide levert energie per kilogram uit het net; de correspondentie zit in de structuur, nog niet in de getallen.
7. Doorgang in plaats van voorraad
De vraag
Een batterij is nodig omdat opwekking en gebruik niet samenvallen in tijd of plaats. Maar als een maas continu zou leveren op de plaats van gebruik, verandert de rol van de batterij ingrijpend: ze brugt dan geen energie meer over, maar vangt alleen pieken.
De energie van één maas
De energie tussen twee parallelle wanden op afstand L wordt gegeven door de Casimirtelling: per oppervlak π²ℏc/(720 L³). Voor één maas met oppervlak L² resteert π²ℏc/(720 L). Stap voor stap: ℏc = 3,1615 × 10⁻²⁶ J·m; maal π² = 9,8696 geeft 3,1203 × 10⁻²⁵; gedeeld door 720 geeft 4,3337 × 10⁻²⁸ J·m. Bij de natuurlijke maasgrootte van het net, S₀ = 45,588 nm, is de energie van één maas 4,3337 × 10⁻²⁸ / 4,5588 × 10⁻⁸ = 9,51 × 10⁻²¹ J. De natuurlijke tijdseenheid is T₀ = 1,5207 × 10⁻¹⁶ s; één doorgang per T₀ geeft 9,51 × 10⁻²¹ / 1,5207 × 10⁻¹⁶ = 6,25 × 10⁻⁵ W per maas. Dat is de volle snelheid; wat er als stroom uit de maas treedt, is daar een fractie van.
Het gemeten apparaat
De groep van Garret Moddel aan de University of Colorado plaatste een optische holte aan één kant van een metaal-isolator-metaal-junctie en mat een stroom bij nul aangelegde spanning (Symmetry, 2021). Acht tests op artefacten kwamen negatief terug; de stroom steeg naarmate de holte dunner werd. Het patent (US 12,166,434) geeft een maximumvermogen van 1,4 pW; het per elektronenmicroscoop gemeten actieve oppervlak is 0,02 ± 0,006 μm². Per oppervlak: 1,4 × 10⁻¹² W / 0,02 × 10⁻¹² m² = 70 W/m² (met de oppervlaktemarge 54 tot 100).
Vergelijk dat met het net. Eén maas van 45,588 nm beslaat 2,078 × 10⁻¹⁵ m²; de junctie bevat dus 0,02 × 10⁻¹² / 2,078 × 10⁻¹⁵ = 9,6 mazen, die elk 1,45 × 10⁻¹³ W leveren. Tegen de volle snelheid is dat 1,45 × 10⁻¹³ / 6,25 × 10⁻⁵ = 2,3 × 10⁻⁹. Die fractie η is de koppeling van de open kant: het aandeel van wat de maas sluit dat als stroom vertrekt.
Een tweede apparaat: Casimir Inc. in Houston specificeert een chip van 5 × 5 mm die 1,5 V levert bij 25 μA — 37,5 μW op 25 mm², oftewel 1,5 W/m². Het is een specificatie, geen gepubliceerde meting, en per oppervlak 70 / 1,5 = 47 maal onder de junctie van Moddel.
Wat het zou betekenen
Het volgende geldt alleen als twee dingen kloppen: 70 W/m² wordt gereproduceerd, en lagen stapelen zich op.
Een huis. Een woning die voortdurend 1.450 W nodig heeft, vraagt 1.450 / 70 = 20,7 m² aan één laag. Met 100 lagen wordt dat 1.450 / 7.000 = 0,21 m² — een plaat van ruwweg 45 × 45 cm.
Tegen een batterij. Zo’n plaat levert 1,45 kW × 24 h = 34,8 kWh per dag. Een batterij van 250 Wh/kg die één zo’n dag vasthoudt, weegt 34.800 / 250 = 139 kg.
Een drone. Een drone die 1 kW nodig heeft, vraagt 1.000 / 70 = 14,3 m² aan één laag, of 0,14 m² met 100 lagen.
Gewicht van één laag. Eén element is 232 nm dik; bij dichtheden van 2,7 tot 12 g/cm³ weegt dat 0,6 tot 2,8 g/m² — 25 tot 112 kW per kilogram film. Op een drager van 25 μm polyimide (35,5 g/m²) is dat ongeveer 1,9 kW/kg.
Wat er eerst moet kloppen
Vijf punten bepalen of deze rekenkunde telt:
Reproductie. Geen tweede groep heeft Moddels meting herhaald.
De helling. Het net leidt af dat vermogen schaalt als 1/L. Moddels figuur 4b toont stroom met een helling van −0,9; vermogen gaat met stroom in het kwadraat, dus zijn helling is ongeveer −1,8. De stroom matcht de afgeleide exponent; het vermogen nog niet.
Stapeling. Dat N lagen N keer het vermogen van één laag geven, is nooit gemeten — en er hangt een factor 100 mee aan de cijfers hierboven.
Het mechanisme. Maclay (Physics, 2024) betoogt dat het door Moddel voorgestelde mechanisme een stroom in de tegengestelde richting voorspelt; de meting zelf betwist hij niet.
De open kant. De koppeling η is de experimenteel onbekende parameter van het hele ontwerp.
De volgorde van de beslissende metingen volgt uit deze punten: eerst de uitgang tegen de holte-grootte, op vaste stapels van 10 tot 1.000 nm; daarna een stapel van twee lagen tegen één.
8. De lijst en de route
De machines die China onder vergunning plaatste, horen bij één procesorde: mengen, coaten, drogen, persen, wikkelen of stapelen, vullen, vormen. Vier procesfamilies ontbreken op die lijst: plasmadepositie, ruimtelijke ALD, dunne-film-cellen en elektrodepositie. De vorming staat er wél.
Het Nederlandse plan uit de krant volgt de bestaande orde met bestaande machines — een lijn uit coaters, foillamineerders en verpakkingslijnen — en bevat daarmee de stappen die onder vergunning staan.
Maar er is een tweede route mogelijk: processen kiezen die die stappen niet nodig hebben (zie paragraaf 4), en de Nederlandse sterkte in meting en regeling in de lus leggen (paragraaf 5). Die route is minder rijp, raakt de Chinese lijst nauwelijks, en de stukken liggen er al: LeydenJars plasmalijn, E-magy’s poreuze silicium, de ruimtelijke ALD van TNO Holst Centre, LionVolts driedimensionale cel en de proeffabriek OBIC in Helmond.
9. Wat er voor mensen verandert
Een batterij wordt gemaakt en gebruikt door mensen.
In Eindhoven assembleren laboratoriummedewerkers in witte jassen de testcellen van LeydenJar de hele dag met de hand; ongeveer 95% slaagt. Hun vakmanschap is wat de nieuwe machines zullen moeten dragen. In Skellefteå bouwde een stad van 75.000 inwoners haar toekomst op Northvolt; in één bezuinigingsronde alleen al verloren 1.600 mensen in Zweden hun baan, 1.000 van hen daar, terwijl de technici van Wuxi ernaast werkten, in hun eigen kamp, zorgend voor de machines. In Debrecen maakten bewoners zich zorgen over de CATL-fabriek; begin 2026 sloten de autoriteiten delen ervan nadat negen werknemers aan nikkel waren blootgesteld. In Oekraïne lopen de drones die soldaten dragen nog grotendeels op Chinese cellen; een vergunningsbesluit in Peking reikt tot aan hun frontlijn.
Elk van de zes richtingen verplaatst het werk. Minder vragen van de cel verhuist de opslag van de verre gigafactory naar de straat, het huis en de wijk. Maken zonder vuur maakt fabrieken schoner en kleiner, zodat ze in steden passen in plaats van aan de rand van het land. De lus in plaats van de hand verandert wat een fabrieksarbeider doet: van afstellen op gevoel naar het lezen van een toestand die zichzelf houdt. En doorgang in plaats van voorraad verandert, als ze klopt, het meest: een huishouden zou niet langer afhangen van een kabel, een onderstation of een batterijpark voor zijn basislast — de batterij wordt een klein buffertje in de meterkast.
10. Statusregister
Stelling
Status
Energie is een winding; vermogen = 1/S
afgeleid binnen de theorie
Een batterij is een houdtoestand; vier fasen; τ-ladder
afgeleid construct van de theorie
Vensterregel: 9 van 27 vensters sluiten
geteld (exact)
Een grenslaag in een cel is een open venster
lezing
Ruimte in plaats van greep; één richting; homogene sluiting; geen naad
correspondenties met gemeten resultaten (bronnen in de referenties)
Volgorde van het leven volgt de telling van open vensters
hypothese, ongemeten
Fabrieksenergie 23–41 kWh per kWh cel; vorming 3–7 dagen
Konstapel, J. (2026). “The Vacuum.Net Theory: Complete Derivation.” constable.blog, 28 september. https://constable.blog/wp-content/uploads/vdefThe-Vacuum.Net-Theory.pdf Waarom lezen? Axioma’s, definities, natuurlijke eenheden en de maasenergie. Lees eerst de statustabel aan het eind.
Konstapel, J. (2026). “Applied Vacuum Theory, Parts Two to Four.” constable.blog, 25–27 september. https://constable.blog/2026/09/25/applied-vacuum-theory/ Waarom lezen? Deel Twee geeft verweken-schrijven-verharden. Deel Drie de zelfstemmende lus. Deel Vier vermogen = 1/S en de volgorde van metingen — met name de sectie over energie zonder reservoir.
Zheng, J., Zhao, Q., Archer, L. A. e.a. (2019). “Reversible epitaxial electrodeposition of metals in battery anodes.” Science 366, 645. https://news.cornell.edu/node/315386 Waarom lezen? Het duidelijkste geval van een onderlaag die de groeirichting zet: tienmaal stabieler bij lage stroom.
Schütte, A., Degen, F. & Walter, L. (2024). Batteries 10(2), 64. https://www.mdpi.com/2313-0105/10/2/64 Waarom lezen? Energie per productiestap per kWh cel; de bron van de 11,02 kWh in paragraaf 4.
Moddel, G., Weerakkody, A., Doroski, D. & Bartusiak, D. (2021). “Optical-Cavity-Induced Current.” Symmetry 13, 517. https://arxiv.org/abs/2101.03085 Waarom lezen? De meting achter paragraaf 7, met haar acht controletests. Let op: figuur 4b toont stroom, geen vermogen.
Maclay, G. J. (2024). “Modeling Electronic Devices with a Casimir Cavity.” Physics 6(3), 1124–1131. https://doaj.org/article/31210466294848b0b30d3380b0056c2c Waarom lezen? Een kritiek op het voorgestelde mechanisme die de meting overeind laat.
Hymas, K. e.a. (2026). “Superextensive electrical power from a quantum battery.” Light: Science & Applications. https://arxiv.org/abs/2501.16541 Waarom lezen? De eerste volledige cyclus van een kwantumbatterij, en waarom de laadkoppeling ook ontlaadt.
From Particle to Cosmos: The Measured World from Below the Atom to the Edge of the Universe
An essay based on J. Konstapel, “From Particle to Cosmos on the Net: A Reference Work on the Particles, the Earth, the Sky and the Vacuum.Net” (Constable Research, Leiden, 2026).
1. A book of measured cases
From Particle to Cosmos on the Net sets out to do one thing: tell the physical world through its measured cases — the masses, sizes, lifetimes, dates and counts that experiment has actually produced. The book is organised in four parts. Part I, Below the Nucleus, covers light, the electron, quarks and nucleons, neutrinos, the force carriers, the Higgs, antimatter, the constants, the measured vacuum and the lifetimes of the unstable. Part II, The Earth, follows the planet from its assembly 4,567 million years ago to its present state. Part III, The Sky, runs from the first moments of the universe to the galaxies of today. Part IV places all of it on a single scale, “the ladder”.
Two structural choices give the book its character. First, every number carries its unit and its calculation; the reader needs nothing else. Second, the book reads its history alongside a private interpretive framework, the “Vacuum.Net”: a strand that touches itself, whose closures are said to persist while their carriers are replaced. The author is explicit that this is a reading, placed beside the history rather than offered as its cause. That separation is maintained honestly, and it is what allows the measured record to stand on its own.
The scale itself is unusual and effective. The position of a length is the logarithm to base 3 of the length divided by S₀ = 4.558816 × 10⁻⁸ m — half the inverse of hydrogen’s Rydberg constant. The position of a time uses T₀ = 1.520655 × 10⁻¹⁶ s, so that S₀ = c·T₀ and a length and its period share one coordinate. One step on the ladder is a factor of three. The day sits at position 43.50, the year at 48.87, the age of the universe at 70.12. The catalogue in chapter 36 orders 2,407 such positions, and every case in the book is a row in a companion catalogue of 7,035 cases. The result is that a photon of 2.5 PeV (position −29.27) and the observable universe (71.86) can be discussed in one breath, as points on one ruler.
2. Below the nucleus: precision as the working method
Part I is a history of measurement, told in the order things were measured. Light is the model case. Its speed went from Rømer’s 22 minutes across Earth’s orbit in 1676 — 24 % below today’s value — through Fizeau’s toothed wheel, Foucault’s and Michelson’s rotating mirrors, Essen’s cavity resonator, to the 1972 laser measurement, and in 1983 the number was fixed outright by the definition of the metre: 299,792,458 m/s exactly. Each order-of-magnitude gain in precision arrived with a new method. The book generalises this pattern everywhere: precision is not accumulated, it is replaced.
The same discipline applies to the photon’s other properties. Every measured property apart from speed and spin is a bound: mass below 1 × 10⁻¹⁸ eV (from solar-wind magnetohydrodynamics extended to Pluto’s orbit), charge below 1 × 10⁻³⁵ e (from the isotropy of the cosmic microwave background). The book is careful to show that published bounds differ by many orders of magnitude and that the tightest ones rest on models — a Proca photon and a Higgs-mass photon admit different limits at all.
The electron chapter carries the strongest claim of the part: the electron’s magnetic moment, g/2 = 1.00115965218059(13), measured on a single trapped electron in 2023, is the most precisely tested number in particle physics, known to 0.13 parts per trillion. Yet the book also records where precision disagrees with itself: the two atomic-recoil measurements of the fine-structure constant (caesium 2018, rubidium 2020) differ by 5.5σ, even though each is good to a few parts in 10¹⁰. Open questions are kept open. The proton-radius puzzle — muonic hydrogen’s 0.84184 fm against the pre-2010 electronic 0.8768 fm — is followed through its resolution (CODATA 2022 now recommends 0.84075(64) fm) and its residue (a 2.8σ tension remains inside the CODATA analysis).
The quark chapter makes the book’s most striking physical point with a single number: the proton weighs 938.27 MeV, but its three valence quarks weigh 9.02 MeV — 0.96 %. A 2018 lattice calculation splits the remainder into quark energy (33 %), gluon field energy (37 %), the trace anomaly (23 %) and quark masses (9 %). Mass, in other words, is not stuff; it is the binding and motion of what holds the proton closed. The neutron gets the mirror-image treatment: free, it lives 878.4 ± 0.5 s, and the beam and bottle methods still differ by 4.4σ; bound in a deuteron, it does not decay at all.
Neutrinos supply the part’s cleanest narrative arc: proposed in 1930, caught in 1956 at three events per hour from a reactor flux of 5 × 10¹³ neutrinos per square centimetre per second; missing from the Sun for decades (Homestake found a third of the prediction); restored by SNO in 2001–2002, when the total flux held at the solar value while only 34.6 % arrived as electron type — the total closes, the kind flows. The Z boson’s invisible width counts the kinds at 2.9963 ± 0.0074: three, and only three.
3. The Earth: a planet read through its clocks
Part II is the most original section as a synthesis, because it treats geology the way Part I treats particle physics: as a set of measured quantities with positions on one scale. The Earth is dated from its first solids at 4,567.30 ± 0.16 million years; its core closed within about 30 million years of them (the tungsten isotope excess of the silicate shell is the receipt); the Moon followed from a late collision between 30 and 207 million years after the first solids, and the oxygen and titanium isotopes show Earth and Moon are one stock to within a few parts per million.
The record of the deepest time is not rock but crystal: single zircon grains of 4,404 ± 8 Ma, eroded from a vanished crust, carry both the age and — through their oxygen isotopes — evidence of liquid water and granite-like crust within about 160 million years of the first solids. Water, air and life are then followed as one system. Oxygen rose in three steps (the Great Oxidation Event between 2.46 and 2.06 Ga; a billion-year plateau at ≤ 0.1 % of present levels; modern levels by the Devonian, peaking near 35 % around 280 Ma). The ocean’s salinity has held near 35 g per litre for billions of years; the water in it passes through in about 3,100 years.
The closing chapters bring the record to the present, without drama and without evasion: 426.07 ppm of carbon dioxide, 1.42 times the upper bound of the last eight glacial cycles (172–300 ppmv); a sea rising at 4.4–4.75 mm per year, about twice the rate of the 1990s; an ocean that took 91 % of the heat added to the climate system; a human population of 8.09 billion carrying a made mass of 1.1 teratonnes. The book’s own reading is that these numbers count “the carbon that stands outside closed forms” — a poetic gloss, but the measurements underneath are standard and correctly cited.
4. The sky: from one second to 13.787 billion years
Part III is a history in which almost every quantity has been measured within living memory. The universe is 13.787 ± 0.020 billion years old; its matter fills a sphere 46.5 billion light-years in radius. The first three minutes fixed hydrogen, helium (a mass fraction of 0.245 ± 0.003, agreeing with the frozen neutron-to-proton ratio of 1/7) and a trace of lithium. At 371,800 years the light was set free, and it has been measured since as a blackbody to better than 50 parts per million — varying across the sky by 3.7 × 10⁻⁵, the closure that is close but never exact, and the seed of everything later.
The same quantity is then read at two depths: the light elements give Ω_b h² = 0.02205 ± 0.00043, the background light 0.02237 ± 0.00015 — one amount of matter, seen at positions 37.88 and 60.55 on the ladder. The book records the field’s live tension without taking sides: the early universe gives H₀ = 67.4–68.5 km s⁻¹ Mpc⁻¹, the local ladder 73.0–73.5, and the ratio between them is 1.084.
Stellar physics is told as a chain of residues: hydrogen burns to helium (26.73 MeV per chain), the triple-alpha process closes two helium nuclei into carbon-12 (the beryllium-8 intermediate lasts 8.19 × 10⁻¹⁷ s — what does not close does not stay), and the elements beyond iron are built by neutron captures — slow ones at ten to a hundred years apart in dying stars, rapid ones at up to a hundred per second in mergers, where GW170817 threw out an estimated 0.078 solar masses of r-process matter. The Sun, seen from close by, dates to 4,567.30 ± 0.16 Ma from grains preserved in meteorites; its neutrinos leave the core in 2.3 s while its light takes 10,000 to 170,000 years to reach the surface — a factor of 10¹¹ in the travel times of the two carriers of the same energy.
The outer parts scale the story up without changing its grammar: 6,445 confirmed exoplanets in 4,834 systems as of October 2026, most of them between Earth and Neptune in size and orbiting in under 100 days; a Galaxy of ~2 trillion counterparts; a cosmic web of filaments, clusters and voids; and one acceleration, g† = 1.20 × 10⁻¹⁰ m s⁻², that fits 2,693 points across 153 galaxies of very different shapes — a single scale on which the Sun’s own orbit sits at 1.7 times the threshold.
5. The ladder, and what the reading adds
Chapter 36 assembles the whole book on one axis: from the Planck length at −57.53 to the observable universe at 71.86, about 130 positions — a factor of three, 130 times over. On it, the proton’s charge radius (−16.21), the Bohr radius (−6.15), the Earth’s mean radius (29.65), the Sun’s radius (33.92), the Galaxy’s disc (59.99) and the edge of the observable universe (71.86) are points on one ruler, and the times run from the Z boson’s 2.6 × 10⁻²⁵ s (−18.36) to the age of the universe (70.12).
What does the Vacuum.Net reading add? As physics, nothing it claims to add — the author says so himself, repeatedly and plainly: “a reading is called a reading.” Its forms (what closes stays; the form lasts while its carriers are replaced; one form at three depths; loading, holding and discharge) function as an index: a way of noticing that the proton does not decay while the W does, that the neutrino flux holds while its kind changes, that the ocean’s form holds while its water passes through every 3,100 years, that the continents keep their area while their material is recycled. Whether or not one adopts the framework, the pattern-spotting it drives is consistently illuminating, and the book’s candour about the distinction between measurement and interpretation is its most professional quality.
Three criticisms are fair. First, the reliance on Wikipedia as a secondary source (the book marks it as such) is unusual in a reference work, though the primary papers are cited beside it wherever a number matters. Second, the net-text apparatus (trits, windings, knots, closure depths 41 and 53) is introduced briefly and will be opaque to readers who do not consult the author’s companion papers; the essayist’s advice is to read around it on first pass. Third, the book’s habit of stating open tensions without resolving them (α’s 5.5σ, the neutron’s 4.4σ, H₀’s 1.084) is correct practice, but a reader must bring patience.
Against this stands the achievement: a single volume in which the same intellectual procedure — measure, bound, date, count, place on the scale — is applied from the Compton wavelength to the cosmic web, in which every number shows its derivation, and in which the 20th century’s central lessons (constants become definitions; mass becomes binding energy; the vacuum becomes a measurable medium; the universe becomes a dated object) appear not as doctrine but as the accumulated result of specific experiments, each with its name, its year and its error bar.
6. Conclusion
From Particle to Cosmos on the Net is best understood as two works bound together. One is a compact, self-contained reference work of measured physical fact, from 1676 to 2026, with 1,027 sources and an unusual but coherent logarithmic index. That work succeeds on its own terms and would be worth having without the second. The second is an interpretive essay in the author’s Vacuum.Net framework, offered explicitly as a reading and kept separable from the first. The binding is the ladder: one scale, in factors of three, from hydrogen’s light to the edge of the observable universe. Its lesson is the one the book keeps repeating in different words: what closes, stays — and what stays, we can measure.
Annotated References
The book cites 1,027 sources, numbered [1]–[1027]. The selection below annotates the sources most used in this essay, grouped by kind. Bracketed numbers are the book’s own reference numbers.
Standards and primary data compilations
[1] NIST (2024), CODATA 2022 adjustment, “Fundamental Physical Constants — Complete Listing.” The recommended values of all fundamental constants, the source of nearly every constant quoted in Part I (electron mass, α, Rydberg constant, proton radius). The book’s backbone for numbers.
[2] Mohr, Newell, Taylor & Tiesinga (2025), “CODATA recommended values…: 2022,” Rev. Mod. Phys. 97:025002. The peer-reviewed account of the same adjustment, used for the internal CODATA analysis (e.g. the proton-radius tension between electronic and muonic data).
[7][8][13][59][69][113] Particle Data Group (2024/2025), Review of Particle Physics. The standard reference for particle masses, widths, lifetimes and bounds. The book notes explicitly that PDG 2024 tables still carry CODATA 2018 values, and tracks the differences — a detail many summaries miss.
[25] PDG (2024), Astrophysical constants and parameters. Age of the universe (13.797 ± 0.023 Gyr), CMB temperature and photon density, baryon-to-photon ratio; the anchor of Part III’s numbers.
Discovery and precision-measurement papers (physics)
[38] Fan, Myers, Sukra & Gabrielse (2023), measurement of electron g/2 to 0.13 parts per trillion. The most precise property of any elementary particle; one electron in a Penning trap.
[51] Pohl et al. (2010) and [52] Antognini et al. (2013), muonic hydrogen spectroscopy. The measurements that opened the proton-radius puzzle (0.84184(67) fm, 5σ from the then-accepted value).
[110] Cowan, Reines et al. (1956), “Detection of the Free Neutrino: a Confirmation,” Science 124:103. The first detection, three events per hour against a 5 × 10¹³ cm⁻² s⁻¹ flux.
[116] Super-Kamiokande (1998), “Evidence for oscillation of atmospheric neutrinos,” PRL 81:1562; [117][118] SNO (2001–2002). The discovery that neutrinos change kind, and the measurement that closed the solar-neutrino problem while the total flux held.
[119] Daya Bay (2012), sin²2θ₁₃ = 0.092, 5.2σ. The third mixing angle, from 55 days of reactor data.
[121] Esteban et al. (2024), NuFIT 6.0 global fit, JHEP 12:216; [122] KATRIN (2025), direct neutrino-mass limit m < 0.45 eV. The current state of the oscillation parameters and of the direct mass bound.
[156] Christenson, Cronin, Fitch & Turlay (1964), “Evidence for the 2π Decay of the K₂⁰ Meson,” PRL 13:138. The discovery of CP violation.
[169][170] ATLAS and CMS (2012), observation of the Higgs boson. The two simultaneous discovery papers; the book dates the “kind opened” here and follows its filling (8–9 million events per experiment by 2018).
[162] CDF (2022), m_W = 80,433 ± 9 MeV. The W-mass measurement that stood 7σ from the Standard Model prediction as quoted; the book sets it honestly beside the 2024 ATLAS and CMS values that sit near the prediction.
[180] Andreassen, Frost & Schwartz (2018), “Scale Invariant Instantons and the Complete Lifetime of the Standard Model.” The computed vacuum lifetime (~10¹³⁹ years central, >10⁵⁸ years at 95 %), the longest time in the book.
[196] BASE (2022), antiproton-to-proton charge-mass ratio to 16 parts per trillion. The sharpest test of the matter–antimatter mirror.
[216] Lange et al. (2021), atomic-clock limits on drift of constants (|α̇/α| < 1.0(1.1) × 10⁻¹⁸ per year); the stiffest measured case of the book’s “context” form.
[224] Lamb & Retherford (1947) and Bethe (1947), the Lamb shift (cited via a secondary source): the 1,057 MHz split that opened modern QED and returns in the book as a measurement of the vacuum.
[233] Wilson et al. (2011), “Observation of the Dynamical Casimir Effect,” Nature 479:376. The moving mirror that turned vacuum fluctuation into real photons.
[253] Gonzalez et al. (2021), UCNτ neutron lifetime; with the beam measurements, the source of the still-open 4.4σ beam–bottle gap.
Earth and planetary sciences
[265] Connelly et al. (2012), “The absolute chronology and thermal processing of solids in the solar protoplanetary disk,” Science 338:651. The 4,567.30 ± 0.16 Ma anchor date for the first solids, used for Earth, Moon, Sun and small bodies alike.
[272] Hazen & Morrison (2022), “On the paragenetic modes of minerals,” American Mineralogist 107:1262. The mineral-count framework (23 species before the Sun, 86 in the nebula, 5,659 today) behind the book’s “opened once, filled many times” reading of mineral evolution.
[916] Brown et al. (2002), “The flux of small near-Earth objects colliding with the Earth,” Nature 420:294. The impact-rate scale (5-kiloton strikes yearly, Chicxulub at 66.043 Ma).
[656] Our World in Data, Population (HYDE, Gapminder, UN WPP 2024). Source of the demographic figures in the chapter on the Earth now.
Cosmology and astronomy
[686] Riess et al. (2022), SH0ES, H₀ = 73.0 ± 1.0 km s⁻¹ Mpc⁻¹; set against Planck’s 67.4 — the Hubble tension, reported at its measured value of 1.084 in ratio.
[721] Schouws et al. (2025), [O III] detection in JADES-GS-z14-0 at z = 14.1793. Among the sources for the record early galaxies of chapter 26.
[758] Rozwadowska, Vissani & Cappellaro (2021), Milky Way core-collapse supernova rate (one per ~61 years — below the book’s closure depth 53 at 93.4 years).
[783] GRAVITY Collaboration (2022), Galactic-centre stellar orbits; with the EHT imaging results, the observational basis for the black-hole sections.
[1026] Lopez, Clowes & Williger (2022), “A Giant Arc on the sky,” MNRAS. One of the large-scale-structure anomalies the book lists without endorsement.
The author’s own framework texts (secondary within the book)
[26]–[33] Konstapel (2026), the Vacuum.Net papers (Complete Derivation; Foundational Paper 7th ed.; Clerk Maxwell and the Vacuum.Net; The Climate on the Ladder of Three; The Heat Beneath; The Balanced Three; How to Make Matter). These supply the interpretive forms used throughout — trits and windows, windings and knots, the ladder of three — and the reader should treat them exactly as the book treats them: readings beside the history, cited as such.
[27] within these, the electron-as-confined-photon reading follows Williamson & van der Mark (1997); a speculative model, flagged by the book as the net texts’ adoption rather than as established physics.
Secondary encyclopaedic sources
[3]–[6], [9]–[12], [17], [60], [67], [72], [84]–[85], [109], [134], [761], [793], [939] and ~200 similar entries: Wikipedia pages, read 15 October 2026. Used deliberately as secondary sources and marked as such in every citation; the historical narrative (Rømer to Bradley to Fizeau to Michelson; the quark timeline; the geoneutrino and supernova compilations) leans on them, with the primary papers cited alongside wherever a specific number is at stake. This is the book’s most unconventional practice, and the annotation policy — every Wikipedia entry labelled “secondary” — is at least transparent about it.
Het boek beschouwt cultuur als één doorlopend archief, van de eerste stenen werktuigen 3,3 miljoen jaar geleden tot het huidige netwerk van zes miljard mensen.
Most histories of human culture are told as a sequence of inventions: the stone tool, the farm, the city, the script, the machine. This book takes a different position. It treats culture as one continuous record — from the first broken stones in Kenya 3.3 million years ago to a present in which six billion people share a single network — and reads that record against a small set of recurring forms drawn from one model, the “Vacuum.Net.”
The book’s method is an audit rather than a narrative. Every claim carries a date, a size, a count, and a source. Every length and duration is placed on one logarithmic “ladder” in which each step is a factor of three (the day sits at position 43.50, the year at 48.87, a generation of 26.9 years at 51.87). The result makes genuinely comparable things that are usually told in separate genres: a business cycle, a dynasty, and an orbital cycle end up on one scale.
The central proposition is compact enough to state in one line: what closes, stays. A form that achieves closure — a stone edge, a cell membrane, a city wall, a script — persists while its carriers are replaced, generation after generation. Beside it stand three companion forms that recur through the whole book: the residue of one depth becomes the alphabet of the next; a new kind of union opens once and is then filled many times; and reading keeps the relations and drops the rest.
The book is explicit that the model is a lens, not a cause: the forms are “read in the history… described beside it, not as its cause.” This essay follows the book’s own three-part structure — the long beginning, the kinds of culture, and what lies ahead — and closes with an assessment.
Part I — The Long Beginning (Chapters 1–8)
Before culture (ch. 1). The record begins before Homo sapiens. At Lomekwi 3 in West Turkana, worked stones lie in place at 3.3 million years ago — 700,000 years older than the Oldowan industry long taken as the start of technology. The makers were not yet our genus: the oldest Homo jaw, from Ledi-Geraru in the Afar, is only 2.8–2.75 million years old. Tool-makers left Africa early, reaching the Chinese Loess Plateau by about 2.1 Ma and the Caucasus (Dmanisi) by 1.85–1.78 Ma.
The handaxe follows. The earliest measured Acheulean assemblage lies at Garba IV, Melka Kunture, Ethiopia, around 1.95 Ma — in levels only 50,000 years above the Oldowan, showing that the new form opened before the old one closed. At Olorgesailie in Kenya the tradition was practiced for roughly 800,000 years through at least sixteen major environmental shifts; regionally the Acheulean ended between about 174 and 50 ka, at different times on different continents. The makers of the late handaxes were not the makers of the first: the form lasted; its carriers were replaced.
The same stretch delivers fire burned in place at Wonderwerk Cave about 1.0 Ma, fire made with pyrite at Barnham about 400 ka, the first engraving (a shell from Trinil, Java, 540–430 ka, by Homo erectus), the first hafted compound tools (Kathu Pan 1, ~500 ka), the first joined wood (the notched logs of Kalambo Falls, at least 476 ka — a structure, not a tool), and the hunting spears of Schöningen, re-dated by amino acid geochronology to about 200 ka. Fire also changed the energy budget of a person from about 3.5 GJ per year (the body’s own 111 W) to 7–15 GJ — the first doubling of human energy capture.
The Neanderthals (ch. 2). A parallel human line lived from the Atlantic to the Altai, few in number (effective population 3,000–12,000), with brains as large as ours, a hyoid bone matching ours, tar-hafted tools at the North Sea coast, and a tool tradition — the Mousterian — that lasted more than a quarter of a million years and ended with them, between 41,030 and 39,260 cal BP.
The other humans (ch. 3). Between about 300 and 50 ka, at least eight lineages of Homo lived at once: Denisovans, late H. erectus, Homo naledi, the small humans of Flores and Luzon. The chapter frames them with a genetic bottleneck of about 1,280 breeding individuals, 930–813 ka, and ends with the meeting of all these lines inside living genomes.
Sapiens in Africa (ch. 4). The oldest fossils lie at the ends of the continent: Jebel Irhoud in Morocco at 315 ka, Florisbad at 259 ka, Omo and Herto in Ethiopia. Modern brain size came first; the modern globular shape only in fossils younger than 35 ka. The Middle Stone Age delivers, each first in one place and then in others: pigment, beads, engraved signs, a drawing, the bow and arrow, poison, bedding, and the care of the dead.
Out of Africa (ch. 5). Early presences outside Africa stand at Apidima (>210 ka) and Misliya (≥177 ka); the lasting exit populated every continent except Antarctica within about 60,000 years — into Sahul by perhaps 1,300–1,550 founders, and across the Americas as far as the White Sands footprints in New Mexico, dated to ~21,000 years ago — and the last large lands, Madagascar, Iceland, Aotearoa, only in the last 1,400 years.
Hunters and gatherers (ch. 6). Foraging filled 95% of the species’ history. The comparative measures: a mean experienced band of 28.2 adults across 32 societies, most band members not genetically related, kinship patterns like those of Sunghir 34,000 years ago, and a repertoire of settled foragers — Jōmon, Natufian, Göbekli Tepe, Amazonia, the Arctic and the Pacific Northwest.
Settling and farming (ch. 7). Villages and monuments came before farming: the Natufians lived in half-sunken huts for over 2,000 years and ate bread 4,000 years before agriculture. Domestication then began independently in at least eleven regions with different species — wheat and barley in southwest Asia, millet and rice in China, taro in New Guinea, maize in Mexico, manioc in Amazonia, potato and llama in the Andes.
The first cities and states (ch. 8). Here the book’s second signature form appears: the village, itself closed and durable (median occupation in the Fertile Crescent about 1,000 years), became one letter of a larger closure — the city, then the state. Uruk grew from 20 to 600 hectares within a millennium. Six regions formed primary states independently, each when rule first reached beyond a day’s round trip. Egypt’s state form held for ~2,500 years through dynasties with a median length of 127 years; China’s through 21 houses. Over 4,919 years the largest polity grew 355-fold — not smoothly, but in long plateaus broken by sudden rises, with clustered collapses around 2200 BC (Akkad, the Old Kingdom, Harappan urban life). The record keeps relations and drops the rest: 85.6% of Uruk’s tablets were administrative; Teotihuacan’s plan kept the sun for 850 years.
Part II — The Kinds of Culture (Chapters 9–24)
Part II tells each kind of culture as its own history, first trace to present.
The sacred (ch. 9) begins with the care of the dead — ochre, beads, dug pits (with the Homo naledi burials presented alongside their published rebuttal) — and runs through the Natufian graves, the megaliths of Atlantic Europe, the moralizing gods of the Axial Age, the great temples and pilgrimages, to the present counts of adherents by tradition.
The spiritual path (ch. 10) is measured comparably: in a survey of 488 societies, 437 — about 90% — had institutionalized trance. Texts on method begin in India in the first millennium BCE; in the last half-century the path entered the laboratory, with thermometers on Tibetan monks practicing tummo heat yoga and electrodes on long-term meditators.
Music and dance (ch. 11) runs from the debated Divje Babe flute (presented with both readings) and the certain Swabian Jura flutes (>35 ka) to 253 million recorded tracks, with about 106,000 new tracks arriving each day. Whale and birdsong show that learned, transmitted song is not exclusively human: humpback song was transmitted population to population across the Pacific for 11 years.
Image (ch. 12) opens with the Trinil engraving and Ice Age caves (Sulawesi, at least 51,200 years) and ends with about 1.9 trillion photographs taken in 2024. An image is the book’s plainest case of reading: “a homomorphic image: it keeps the relations and drops the rest.”
Building (ch. 13) runs from the Kalambo Falls joint through Djoser and Khufu to a tower of 828 m, tracking size on the ladder. The Pantheon’s dome, self-healing in its lime, still stands after nineteen centuries — form and carrier in concrete.
Cities (ch. 14) states its thesis in one line — “a city is a place that holds its form while its people pass through” — and traces the largest city through 5,000 years to a ceiling that held at roughly one million inhabitants for some 1,800 years (from ancient Rome to Beijing), before industrial energy broke it.
The shaped landscape (ch. 15) follows fields, canals, terraces and raised beds to the share of the Earth that people hold now. Canal irrigation began independently in at least four regions; the qanat of Gonabad has delivered water for some 2,700 years and still serves about 40,000 people, and the Budj Bim eel channels in Australia are at least 6,600 years old. Amazonia holds ~4,700 forest islands shaped by pre-Columbian hands, and the American demographic collapse after 1492 (about 90% of 60.5 million people) let forest regrow over 55.8 million hectares, drawing CO₂ down 7–10 ppm. Today humanity appropriates 23.8% of land’s primary production (HANPP) — a landscape that is, measurably, one human artefact.
Language and writing (ch. 16) moves from the Kebara hyoid and FOXP2 (with both the original selection claim and its later reanalysis) through 7,159–7,674 living languages in 246 families to writing’s four independent inventions — Mesopotamia, Egypt, China, Mesoamerica — where, as in Mesopotamia, the number signs come before the other signs. Compression is the signature of the line: roughly 800 proto-cuneiform signs collapsed to 22 Phoenician letters (a factor of 36), and from Brahmi grew 198 scripts. The Cherokee syllabary made a majority literate within 3–5 years; worldwide, literacy rose from 12.05% (1820) to 87.74% (2024). Across 17 languages, spoken information flows at a strikingly uniform ~39 bits per second.
Number and mathematics (ch. 17) starts with counting older than writing (the Lebombo bone, ~42 ka), passes Babylonian base-60 and the tablet YBC 7289 — √2 computed to six decimal places in clay — and notes that zero was invented three times (Babylon, the Maya, and India/Cambodia, in stone by 683 CE). Madhava of Kerala carried π to eleven places centuries ahead of Europe. The line ends with a world literature of nearly five million items (zbMATH, ~125,000 added yearly) and Setun, the Soviet balanced-ternary computer that counted in +1, 0 and −1 — the same three values the book’s own model takes as elementary.
Knowing nature (ch. 18) follows two lines: the sky, from Nabta Playa’s solstice circle (~4800 BC) through Newgrange, the Nebra disc, Chankillo’s thirteen towers, Teotihuacan’s 260+105-day count, the Antikythera mechanism and the Maya Venus tables (584 days per period, a residue of 0.08 day) to 3.3 million articles a year; and the body, from trepanation — practised on four continents, with survival rates in Peru up to 91% — through Ibn Sīnā’s Canon, taught for 690 years, to smallpox eradication (1980) and the length of life rising from 28.5 years in 1800 to 73.2 in 2023. Of 12,495 medicinal plant uses, more than 75% are known to a single language only: the botany and the linguistics are one record.
Technique (ch. 19) spans 3.3 Ma of stone to a language model trained with 4.6 × 10²⁶ operations, tracking two measures: how often a technique began on its own (pottery at least three times — China ~20 ka, the Sahel, Amazonia; metal at least twice — Balkans copper and Andes gold; the wheel known in Mesoamerica only as a toy), and its spread time, which keeps shortening: paper took ~1,216 years to travel from China to Europe, the landline 43 years, the smartphone 2 — and a chatbot reached 100 million users in two months.
Energy (ch. 20) carries the per-person series from muscle (3.5 GJ, 111 W) through fire, farming and coal to 640 EJ of primary energy in 2025, 2,426 W per person, with human-made mass doubling every twenty years.
Exchange (ch. 21) begins as carried stone: at Olorgesailie the raw material of the handaxe makers came from at most 5 km away; by about 320 ka obsidian moved from sources up to 95 km distant — a nineteenfold widening, and the first measured trace of social networks. The line runs through gift systems (kula, potlatch, hxaro), coins struck almost simultaneously in Lydia and China (~700–550 BC), an Inca empire that ran without money on mit’a labour and knotted khipu, Potosí silver (~60% of world silver, 13,000 conscripts a year), the company, the bank and the bourse in Amsterdam, to world trade at 56.72% of global output (2024), moving in Kondratieff waves of ~48–53 years.
Living together (ch. 22) measures the nested groups of humanity — camps of ~30 adults, then layers at roughly 1.5, 5, 15, 50 and 150 (each a factor of ~3, as the ladder predicts) — through kinship, marriage, rank, slavery, law oral and written, and violence (from 2% lethal at origin to 12% in medieval societies). Six regions formed primary states; in the Seshat databank 77% of polities form one connected component. Empires die exponentially with a mean lifetime of ~220 years; Chinese major dynasties run a median 268; secular cycles take 200–300. The Atlantic slave trade, the audit’s darkest entry, embarked 12.5 million people and landed 10.7 million.
Story and memory (ch. 23) opens with the oldest dated oral memory — 21 Australian coastal stories that match sea-level curves 7,250–13,070 years old, some 290 generations of tellers — passes the Rigveda (2,039 years between composition and oldest surviving manuscript), the ~500,000-line Manas, and printing’s four steps (Diamond Sutra 868, Bi Sheng’s movable type 1041, the Jikji 1377, Gutenberg ~1455; 641 million printed books in 1750–1800 alone), and closes with a single web crawl of September 2026: 2.17 billion pages. Through the whole line, the story is the closed form; its carriers are replaced.
What was lost (ch. 24) is the audit’s balancing entry. The Americas lost about 90% of 60.5 million people within a century of contact; Tasmania’s population fell from thousands to 154 in thirty years; the Great Andamanese from 5,000 to 19; 984 languages are extinct — one lost every two weeks — and 196 uncontacted peoples remain (October 2025), half at risk within a decade. Indus script, Linear A and Rongorongo are still unread; of the Maya books burned at Maní in 1562, four codices survive. Some closed forms were lost with their carriers and found again only from outside: Egyptian blue, Maya blue, Roman concrete. Cuneiform held for 3,279 years and then stopped with its last scribes. Yet where carriers survived, the form refilled — Aboriginal Australians grew 12.5-fold from their 1921 nadir; Linear B was read again after 3,151 years; the American demographic collapse let forest regrow over 55.8 million ha and cooled the whole Earth’s air by 0.15 °C.
Part III — Ahead (Chapters 25–28)
The rhythms of history (ch. 25) gathers every measured rhythm on one ladder of time, from the Kitchin business cycle (3.3 yr) and dynasties to the beats of sun and sea and the mahāyuga of 4.32 million years. The striking result is structural: several rhythms stand on the same ladder layers — the Saros, the Sothic cycle, the Bond cycle, the Calendar Round and Kondratieff waves all cluster near layer 52, the male generation at 51.99 — and the depths at which cycles close (1, 2, 5, 12, 41, 53, 306) include, at depth 12, a residue of 1.01364: the Pythagorean comma, music’s own leftover from closing the octave in twelve steps. The Maya Long Count (5,125 years) matches precession divided by five. Climatic beats — the 8.2-ka, 5.9-ka and 4.2-ka events — line up with the collapses of Akkad, the Old Kingdom and Harappan urban life around 2200 BCE; the Green Sahara pulsed ~230 times in 8 million years; and the great migrations out of Africa paced to astronomy.
The human net now (ch. 26) stops the clock on a measured present: 8.09 billion people in 2023, 57.8% urban across 33 megacities (Jakarta, ~42 million, the largest); 9.2 billion phone subscriptions; 6 billion people — 74% — online and 2.2 billion offline; 8.8 zettabytes of traffic; 1.9 trillion photographs a year; 106,000 new tracks a day. Growth peaked at 2.283% in 1963 and fell to 0.871% in 2023, with a projected peak of about 10.3 billion in the mid-2080s.
The next depth (ch. 27) reads the line forward. The step the history has taken many times has one movement: a closed form that once lived on its own becomes a letter of a larger closure. The bacterium became the mitochondrion; the cell, the body; the sign, the script; the village, the city. The same handoff now runs through the human functions themselves: memory went to clay around 3200 BCE, calculation went to machines, and ordering is going to the network. The human net stands at such an opening: humans and their livestock are 94% of mammal mass; agriculture and settlement cover almost 39% of ice-free land; made mass (~1.1 Tt) equals all dry living biomass, crossed around 2020 and doubling every twenty years; human energy use is 15–16% of all photosynthesis. Every earlier opening, once made, was filled many times — farming at least eleven times, the state six times, writing four times. The next depth the book names is “a reading that counts itself, bound to no single body”; it names no dates to come, noting only that “a ladder driven by a rest cannot end.”
The ladder of culture (ch. 28) sets 3,290 positions in order across four tables — lengths, areas, durations and ages of the whole catalogue (lengths alone account for 427 positions) — from a 0.5 mm cord at Abri du Maras to the orbital cycles, closing with the statement that the next depth stands above the last position written: “A ladder driven by a rest cannot end.”
Assessment
Read as a whole, the book makes three contributions worth stating plainly.
First, it holds a single scale over the whole record. By placing the day, the dynasty, the energy budget and the orbital cycle on one ladder, it makes comparable what is usually told in separate genres, and it lets the reader see that the same forms recur at every size — the cell membrane, the handaxe edge, the city wall.
Second, it is disciplined about evidence. Every number stands with its unit, its calculation and its source; contested cases (the Homo naledi burials, the Divje Babe flute, the age of the Schöningen spears, FOXP2 selection) are presented with both readings rather than resolved by fiat. Where sources disagree — Terra Amata’s dates, Çatalhöyük’s population — the disagreement is shown, with the factor between them.
Third, its thesis is falsifiable in a way most grand histories are not. “What closes, stays,” “the residue of one depth becomes the alphabet of the next,” and “a new kind of union opens once and is then filled many times” are claims about recurrence, and the 6,995-case catalogue is built so they can be checked against it.
The limitations are the mirror of the strengths. The Vacuum.Net model is described beside the history rather than derived from it; a reader who does not accept the trit-and-closure formalism still receives the full measured history, but may find the recurring form-naming interrupt the flow. And the breadth is bought with compression: 633 pages for three million years means each case arrives as a dated, sourced datum rather than a told story — a reference work to consult, not a narrative to live in.
The book’s own summary may serve as the essay’s close: the record keeps the relations and drops the rest. What persists across three million years is not any particular people, language or technique — it is the closed forms they made and handed on, each becoming a letter of the next closure, up to the one that is opening now.
Annotated Reference List
The sources below are selected from the book’s own numbered list and annotated for the role each plays in the argument. Together they illustrate the evidentiary range of the work: primary archaeological reports in Nature and Science, comparative databases, statistical syntheses, and international data series. Bracketed numbers are the book’s source list.
Harmand, S., et al. (2015). “3.3-million-year-old stone tools from Lomekwi 3, West Turkana, Kenya.” Nature 521:310–315. [2] The anchor for the book’s start: worked stones 700,000 years older than the previously earliest known tools, resetting the beginning of the culture record and showing that technology precedes the genus Homo.
Villmoare, B., et al. (2015). “Early Homo at 2.8 Ma from Ledi-Geraru, Afar, Ethiopia.” Science 347:1352–1355. [6] The oldest dated Homo jaw (LD 350-1). Used to show that the first tool-makers predate the oldest known members of our genus by about half a million years.
Mussi, M., et al. (2023). “Early Homo erectus lived at high altitudes and produced both Oldowan and Acheulean tools.” Science 382:713–718. [12] Establishes the earliest known Acheulean assemblage (~1.95 Ma) with Oldowan and Acheulean in successive levels 50,000 years apart — the book’s evidence that a new form opens before the old one closes.
Key, A.J.M., Jarić, I. & Roberts, D.L. (2021). “Modelling the end of the Acheulean at global and continental levels.” Humanities and Social Sciences Communications 8:55. [20] Statistical modelling of the handaxe tradition’s end across 44 late sites. Supplies the key illustration of regional asynchrony: the same form ending at different times on different continents, after a total run of ~1.7 million years.
Berna, F., et al. (2012). “Microstratigraphic evidence of in situ fire in the Acheulean strata of Wonderwerk Cave.” PNAS 109:E1215–E1220. [22] The earliest secure evidence of fire burned in place (~1.0 Ma), the basis of the energy argument that fire was the first energy source outside the body.
Lenton, T.M., Pichler, P.-P. & Weisz, H. (2016). “Revolutions in energy input and material cycling in Earth history and human history.” Earth System Dynamics 7:353–370. [26] Supplies the per-person energy series (3.5 GJ for the body alone; 7–15 GJ with fire; ~2,426 W today) on which the whole ladder of energy capture in chapters 20 and 26 is built.
Wilkins, J., et al. (2012). “Evidence for Early Hafted Hunting Technology.” Science 338:942–946. [27] Dates hafted compound tools to ~500 ka. The book’s example of “residue becomes alphabet”: the finished flake, itself closed, becomes one letter of the larger closure, the spear.
Barham, L., et al. (2023) on the Kalambo Falls joined wood; and Hutson, C.M., et al. (2025) on the amino-acid re-dating of the Schöningen spears. [28][30] Together they carry the book’s treatment of perishable materials and of moving dates: the earliest joined wood (~476 ka, a structure rather than a tool) and spears whose age halved from 400 to 200 ka as methods improved — both readings kept.
Plummer, T.W., et al. (2023). “Expanded geographic distribution and dietary strategies of the earliest Oldowan hominins.” Science 379:561–566. [3] Extends the earliest Oldowan map by over 1,300 km and links tool-makers to hippopotamus butchery sites — used for the spread of forms across landscapes before they spread across continents.
Brooks, A.S., Potts, R., et al. (2018) on the Olorgesailie transition; and Brooks, A.S. (2018), Springer Nature Research Communities. [17][34][120] Document the shift from local raw material (≤5 km) to obsidian from up to 95 km, with ochre and pigments — the first measured widening of a supply chain and, on the book’s reading, of a social network: the origin of exchange in chapter 21.
Bourguignon, E. & Evascu, T. (1977) on trance in 488 societies, with the HRAF summary. [295][428] The cross-cultural survey behind the claim that ~90% of societies institutionalized trance. Anchors the chapter on the spiritual path in countable, comparative data rather than anecdote.
Hill, K.R., et al. on co-residence in 32 foraging societies (5,067 individuals), and the Binford hunter-gatherer database of 339 societies. [222][220][221] The comparative basis for the measured forager present: a mean experienced band of 28.2 adults, mostly unrelated, and home ranges scaling at exponent ~0.70 with population.
Nunn, P.D. & Reid, N.J. on Aboriginal coastal-drowning stories matched to sea-level curves. [251][1044] The source of the book’s most striking memory case: 21 oral stories datable to 7,250–13,070 years ago, ~290 generations of carriers — the form of “form and carrier” over the longest measured span.
The Seshat Global History Databank and related comparative polity datasets. [410][421 et seq.] The comparative database behind the counts of hierarchy levels across 1,155 societies and the finding of long stasis in polity size broken by sudden rises — the empirical support for the book’s “loading and holding” model and its six primary states.
United Nations Population Division and Our World in Data series (2022–2025 editions). [727][731][973][1110] The measured present of chapter 26: 8.09 billion people, urban–rural parity in 2007, a projected peak near 10.3 billion, growth rates from 2.283% (1963) to 0.871% (2023). Every figure carries its date and source.
Elhacham, E., et al. (2020). “Global human-made mass exceeds all living biomass.” Nature 588:442–444, with the 2023–2025 energy and connectivity data. [721][946][965][1145] The quantitative threshold claims of chapter 27: made mass equal to all dry living biomass (~1.1 Tt), doubling every ~20 years; humans and livestock 94% of mammal mass; 6 billion people online.
J. Konstapel (2026). The History of Culture on the Net: A Reference Work on Human Culture and the Vacuum.Net. Constable Research, Leiden. The source work itself: 28 chapters, 6,995 catalogued cases and 3,290 ladder positions, with every number sourced and every calculation shown. This essay summarizes all three parts and follows the book’s practice of naming both readings where the evidence is contested.
Het essay verbindt het FD-gesprek tussen Dolf van den Brink en Paul Scheffer over het Europese onbehagen: eens over de diagnose dat gemeenschapszin en politieke levering ontbreken, oneens over de remedie (Singapore en liberale maakbaarheid versus richting en een migratieakkoord).
Dolf van den Brink (53) was tot juni 2026 ceo van de Nederlandse bierbrouwer Heineken. Paul Scheffer (72) was hoogleraar Europese studies. Foto: Roger Cremers voor het FD
Op 15 oktober 2026 bracht het FD twee mannen bij elkaar. Paul Scheffer, oud-hoogleraar Europese studies, schreef De dag na de bevrijding. Dolf van den Brink, tot juni ceo van Heineken, las het en maakte veel aantekeningen. Ze zijn het eens over de diagnose. Het onbehagen in Europa reikt verder dan de economie. De gemeenschapszin is weg. De politiek levert niet.
Over de remedie verschillen ze. Van den Brink wijst naar Singapore en wil een nieuwe liberale maakbaarheid. Scheffer wil richting zonder recepten en een migratieakkoord naar het model van het klimaatakkoord.
Dit essay legt het gesprek naast één model: het visnet van de Vacuum.Net-theorie. Het model heeft weinig begrippen. Een draad. Een winding die op zichzelf sluit. Een knoop. Een maas. Een kring waarin een daad pas telt als een ander haar afmaakt.
Met die begrippen valt het gesprek in zes punten uiteen. Het “superieure model” blijkt een eindpunt dat het net niet kent. Singapore sluit de korte kring, niet de lange. Liberale en autoritaire maakbaarheid liggen op één as en missen dezelfde nul. Een akkoord werkt alleen met een doorlaat. Een belofte reikt niet verder dan de kring die haar draagt. En de herinneringsgemeenschap van Scheffer is de trage laag waarop alles rust.
Het essay sluit met een statusgrootboek: wat gemeten is, wat gerekend, wat een lezing is, en wat open ligt.
1. Het gesprek
Van den Brink opent relativerend. Is dit niet hoe geschiedenis altijd gaat? Tussen perioden van voorspoed zit steeds een instabiele tijd. De jaren zeventig waren ook somber.
Scheffer ziet het anders. De instabiliteit duurt al sinds 2001, zeker sinds 2008. En de context is nieuw: oorlog in Oekraïne, en een autoritair China dat economisch slaagt. Volgens zijn boek staat Europa aan het eind van de naoorlogse periode.
Dan komt de zin die de kop van het artikel werd. Van den Brink woonde in Congo, de VS, Mexico en Singapore. Het Europese zelfvertrouwen dat ons model superieur is, noemt hij arrogant.
Hij geeft een voorbeeld. In Singapore wilde de minister van handel best met de directie van de brouwerij praten. Op één voorwaarde: hij wilde ook een vragenuur met het voltallige personeel in de kantine. Daarom vertrouwen Singaporezen hun overheid meer dan wij de onze. Die overheid is aanspreekbaar.
Scheffer werpt tegen dat Singapore een stadsstaat is met zes miljoen inwoners. Een technocratisch bestuur lijkt hem voor Europa geen oplossing. Maar hij erkent dat Europese regeringen te vaak niet leveren.
Daarna volgen de thema’s. Gemeenschapszin als wederkerigheid. De Stichting van de Arbeid en de SER als voorbeeld van samenwerking. Het pensioenakkoord. Scheffers ergernis over het zinnetje “wen er maar aan”. Merkels “es gibt keine Alternative”, gevolgd door de AfD. Een migratieakkoord. Ongelijkheid en vermogensbelasting. De diplomademocratie. Het Draghi- en het Wennink-rapport. Oekraïne in de EU. Het populisme als protectionisme.
Het gesprek eindigt bij hoop. In zijn Sijthoff-lezing van 2024 zei Van den Brink dat hoop geen strategie is. Nu voegt hij toe: zonder hoop is er ook geen oplossing.
2. Het model in tien regels
Dit essay leest het gesprek met één model. Het staat hier volledig, zodat de lezer niets hoeft op te zoeken.
Regel 1. Er is één draad. Alles wat verschillend lijkt, is dezelfde draad die zichzelf op een andere manier ontmoet.
Regel 2. Wat blijft, sluit op zichzelf. Een draad die terugkomt in zijn eigen fase vormt een winding. Wat niet sluit, blijft niet als zelfstandige toestand bestaan.
Regel 3. Knoop en maas. Waar windingen elkaar vasthouden, ontstaat een knoop. Tussen de draden zitten openingen: mazen. Het geheel lijkt op een visnet in water.
Regel 4. Adres en spanning. Elke knoop heeft twee lagen. Het adres is traag: geografie, grenzen, verdragen, bevolking, herinnering. De spanning is snel: prijzen, stemming, incidenten. De trage laag zet het kader voor de snelle, nooit andersom.
Regel 5. Vier toestanden. Spanning laadt, houdt vast, ontlaadt of rust.
Regel 6. De open maas. Spanning loopt door de zwakste open maas. Dat is een punt waar de last samenkomt en geen tweede pad bestaat.
Regel 7. Drie standen. Het register van het net heeft drie standen: +1, 0 en −1. De 0 is geen leegte. Het is de wettige rust waarin een tegenstelling sluit.
Regel 8. De kring. Een daad sluit pas als een ander haar afmaakt. Een gerepareerd dak telt als de eigenaar zegt: klopt. Zonder die ander blijft de daad open.
Regel 9. De horizon. Een verwachting reikt niet verder dan de sluitingstijd van de kring die haar draagt.
Regel 10. Zelfde patroon, andere schaal. Een winding op het ene niveau komt terug op een ander niveau. Een buurt, een stad en een land kunnen dezelfde vorm hebben.
Met deze regels wordt niets “verklaard” in de zin van oorzaak en gevolg. Er wordt beschreven wat sluit en wat niet, op welke schaal, en waar de spanning heen gaat.
3. Het superieure model is een eindpunt
Van den Brink noemt het geloof in de superioriteit van het Europese model arrogant. Zijn argument is ervaring: hij zag elders samenlevingen die anders werken en toch werken.
Het model geeft dat argument een vloer. Achter het “superieure model” zit een oudere aanname. De mens zou gedreven worden door een behoefte aan erkenning. De geschiedenis zou naar een eindpunt lopen waar die behoefte vervuld is. Fukuyama zette dat eindpunt in 1989 op de liberale democratie.
Die aanname heeft een herkomst. Ze is in de Westerse traditie gelast rond het jaar 400. Augustinus verbond het Griekse streven met het Paulinische tekort, en gaf de geschiedenis een richting (Konstapel, 2026a). Via Hobbes, Rousseau, Hegel en Kojève kwam ze bij Fukuyama terecht.
Oudere menselijke tradities kennen die aanname niet. In de overlevering van de San is een mens geen wezen dat erkenning zoekt. Een mens is een plek in een weefsel van mensen, dieren en land.
In het net is er geen eindpunt. Er is alleen sluiting (regel 2). Een model kan niet “superieur” zijn aan een ander model. Het kan alleen sluiten of niet sluiten, op een bepaalde plek en schaal.
Dat maakt de vraag anders. Niet: welk model wint? Maar: waar sluit dit model, en waar niet? Een knoop wil geen aanzien. Een knoop wil een kruising vasthouden, een plek waar lijnen samenkomen die zonder hem niet samenkomen.
Een status is geen kruising. Een samenleving kan iedereen formeel gelijk behandelen en toch knopen overhouden die niets dragen. Daar zit het onbehagen waar Van den Brink en Scheffer over spreken. Het gaat verder dan de economie, omdat het geen tekort aan middelen is. Het is een tekort aan kruising.
4. Singapore sluit de korte kring
Het voorbeeld uit de kantine is precies. Een minister die met het personeel spreekt, maakt een kring met een sluitingstijd van één middag. Een vraag wordt gesteld. Een antwoord wordt gegeven. Iemand zegt: klopt, of: klopt niet. De daad is afgemaakt door een ander (regel 8).
De cijfers bevestigen dat Singapore die korte kring goed sluit. In de Edelman Trust Barometer 2026 vertrouwt 76 procent van de Singaporezen de regering. Dat is het hoogst van alle instituties in het land. De vertrouwensindex staat op 64. Voor Nederland is die 57.
Maar dezelfde meting laat iets anders zien. Slechts 31 procent van de Singaporezen gelooft dat de volgende generatie het beter krijgt. Dat is 11 punten minder dan een jaar eerder. De regering van Singapore stelt dat haar eigen enquêtes geen vergelijkbare daling tonen.
Die twee getallen passen bij regel 9. Een belofte reikt niet verder dan de kring die haar draagt. De kring van de kantine is kort. De kring van een volgende generatie is lang. Singapore sluit de eerste beter dan de tweede.
Hier komt Scheffers tegenwerping terug. Singapore is een stadsstaat. In netwoorden: Singapore is één winding. Wat op één winding werkt, werkt niet vanzelf op een continent van 450 miljoen mensen met vele windingen.
Toch hoeven beide sprekers niet te kiezen. Regel 10 biedt een derde weg. Dezelfde vorm kan op elke schaal terugkomen. Het Nederlandse voorstel voor fractale democratie werkt dat uit (Konstapel, 2025a). Wijkcirkels besluiten via consent over hun eigen leefomgeving. Ze verbinden zich met een hogere cirkel alleen als het lokaal niet lukt. Elke cirkel heeft een dashboard: wat besloten, wat gedaan, wat het kostte, wat het opleverde.
Zo wordt de kantine van Singapore herhaald op elke laag. Niet als regime dat wordt overgenomen. Als vorm die op elke schaal sluit.
5. Twee handen van één as
Het scherpste moment in het gesprek gaat over maakbaarheid. Scheffer ergert zich aan “wen er maar aan”. Als het politieke midden de maakbaarheid loslaat, zegt hij, krijg je autoritaire maakbaarheid terug. Van den Brink antwoordt met een nieuwe liberale maakbaarheid.
Er staan dus drie posities op tafel:
Geen alternatief. Merkel, en het midden dat zegt dat alles complex is.
Autoritaire maakbaarheid. De AfD, het Kremlin, en volgens beide sprekers ook trekken van het huidige Amerika.
Liberale maakbaarheid. Van den Brinks antwoord.
In het net liggen de laatste twee op dezelfde as. De ene is +1, de andere −1. Beide zeggen: wij maken het, van bovenaf, naar ons beeld. Ze verschillen in teken, niet in vorm. En de eerste positie is geen 0. Het is de as zonder richting, die de ruimte openlaat voor een van beide uitersten.
Dit patroon is in Nederland met getallen gemeten. Burgers geven hun eigen leven een 7,7. Het vertrouwen in de politiek krijgt een 4,9 (Konstapel, 2025a). Met mij gaat het goed, met ons gaat het slecht. Dat is het onbehagen van Van den Brink, in twee cijfers.
Hetzelfde stuk laat zien dat de grote blokken in de Nederlandse politiek, hoe verschillend ook, hetzelfde proces hanteren: sturen van bovenaf. Het ene blok combineert markt met hiërarchie. Het andere combineert gemeenschap met bureaucratie. Ze blokkeren elkaar, omdat ze op dezelfde as liggen.
Wat ontbreekt is de wettige 0 (regel 7). In de Nederlandse traditie heeft die een naam: consent. Kees Boeke ontwierp na de oorlog een bestuursvorm waarin een besluit valt als niemand een fundamenteel bezwaar heeft. Een besluit staat dan niet meer ter discussie, behalve als de kring dat wil (Konstapel, 2026b).
Consent is geen compromis tussen +1 en −1. Het is de stand waarin de tegenstelling sluit. Een strategie die de impasse werkelijk opheft, loopt daarom niet via een betere maakbaarheid. Ze loopt via de 0.
Dezelfde vorm is zichtbaar in de geopolitiek. Het HCSS-rapport Europe’s Strategic Trap beschrijft een Russische dwangcampagne onder de drempel van oorlog. Het stelt een Europese tegenstrategie voor. Gelezen in het net zijn beide strategieën spiegelbeelden (Konstapel, 2026c). Beide stellen eisen, beide verhogen de prijs, beide stoppen pas als de ander toegeeft. Geen van beide bevat rust.
6. Van polder naar polarisatie
In zijn Sijthoff-lezing van november 2024 sprak Van den Brink over zijn terugkeer in 2020. De polder, de cultuur van overleg en consensus, had plaatsgemaakt voor polarisatie. Feiten werden openlijk ontkend. Groepen trokken zich terug in hun eigen waarheid.
De polder was een kring. In de SER en de Stichting van de Arbeid zaten werkgevers, werknemers en overheid aan één tafel. Een akkoord telde als alle drie het afmaakten. Dat is regel 8 op landelijke schaal.
Polarisatie is de as zonder 0. Twee polen, en geen plek meer waar ze sluiten.
Het punt over feiten verdient extra aandacht. In het net heeft een feit een eenvoudige definitie: een sluiting die door een ander is afgemaakt. Een regel in de geheugenbank van een kring luidt bijvoorbeeld: Karim hielp Fatima met het dak, dinsdag. Het telt pas als Fatima bevestigt (Konstapel, 2026d).
Een “eigen waarheid” is dan een winding die alleen op zichzelf sluit. Ze heeft geen ander nodig. Daarom staat ze buiten de kring. Het probleem is niet dat mensen verschillende meningen hebben. Het probleem is dat er geen plek meer is waar een bewering door een ander wordt afgemaakt.
Volgens het FD-verslag van de lezing ziet Van den Brink bedrijven als verbinders en als oplossingenmachine. Hij werkte met andere ceo’s aan een deltaplan voor grootschalige her- en bijscholing. Zo’n plan is een concrete kring. Wie leert, krijgt werk dat telt. Wie het werk ziet, bevestigt het.
Elders is gemeten wat zulke kringen doen (Konstapel, 2026e). In Liberia kregen 999 mannen met een hoog risico acht weken groepstherapie en 200 dollar. Tien jaar later was hun crimineel gedrag ongeveer gehalveerd. In Chicago kregen jongeren een zomerbaan met een mentor. Arrestaties voor geweld daalden met 43 procent. Geen van beide programma’s gaf vooral inkomen. Beide gaven een daad die telde, iemand die haar zag, en tijd om te herhalen.
7. Het akkoord en de doorlaat
Van den Brink noemt het pensioenakkoord als bewijs dat samenwerking werkt. Scheffer wil een migratieakkoord, met bedrijven en ngo’s aan tafel, naar het model van het klimaatakkoord.
Beide voorstellen vallen onder regel 8. Een akkoord is een kring. Maar een kring kan open blijven aan de uitgang. De vraag is dan hoeveel van wat de kring afspreekt werkelijk wordt uitgevoerd. Die maat heet hier de doorlaat.
De doorlaat is gemeten bij burgerberaden (Konstapel, 2026f). Het Franse klimaatberaad van 2020 deed 149 voorstellen. President Macron beloofde ze “zonder filter” over te nemen. Een telling van Reporterre vond er 15 terug in de klimaatwet. De doorlaat was ongeveer 0,10. In Ierland liep een burgerberaad uit op een referendum, en de doorlaat was vrijwel 1.
Een kring zonder doorlaat is geen neutraal experiment. Het verhoogt de verwachting, en levert niet. De spanning neemt dan toe. Een migratieakkoord zonder doorlaat zou hetzelfde doen.
Een tweede les komt uit de geschiedenis van de VVD. In 2005 schreef de commissie-Dales een Liberaal Manifest. Het stelde zelf de diagnose: het vertrouwen in de politiek was gezakt van 65 procent in 2000 naar 30 procent in 2003. Het noemde dat het grote politieke probleem van onze tijd.
Daarna regeerde de VVD veertien jaar onafgebroken. Van de voorstellen uit het manifest werd vrijwel niets uitgevoerd (Konstapel, 2026g). Het ene voorstel dat wel werd uitgevoerd, het Ministerie van Veiligheid, concentreerde macht. De voorstellen die macht naar de burger brachten, bleven liggen.
Een relationeel probleem kreeg procedurele antwoorden. Er bestond geen geheugen dat de afstand tussen belofte en levering jaar na jaar zichtbaar maakte. Dat geheugen is wat een akkoord nodig heeft om een doorlaat te hebben.
8. Vier problemen onder één woord
Scheffer houdt vast aan een stelling uit zijn eerdere werk. Om het sociaal contract vorm te geven, is grip op de instroom nodig.
In 2007 besprak de historicus Leo Lucassen zijn boek Het land van aankomst. Hij maakte drie punten die hier van belang zijn.
Het eerste punt. Scheffer stelde dat de integratieproblemen voortkwamen uit naïef multiculturalisme. Lucassen zag daar geen bewijs voor. In Duitsland en Frankrijk was de situatie vergelijkbaar, terwijl die landen veel eerder op assimilatie aandrongen.
In netwoorden: verschillend beleid geeft dezelfde uitkomst. Dan beschrijft het beleid de uitkomst niet. Wat terugkomt, is hetzelfde patroon op een ander adres (regel 10).
Het tweede punt. Scheffer leunde op Oscar Handlins The Uprooted uit 1951. Volgens Lucassen was die these sindsdien weerlegd. John Bodnar schreef in 1985 The Transplanted.
In het net zijn dat twee toestanden van één draad. Ontworteld: de winding is losgemaakt en sluit nergens meer. Verplant: de winding sluit opnieuw, op een nieuw adres. De vraag is dus niet of migratie ontwricht. De vraag is of de winding op de nieuwe plek een kring vindt.
Het derde punt. Te vaak, schreef Lucassen, kregen moslims de boodschap dat zij hooguit voorwaardelijke Nederlanders waren. Ze moesten steeds bewijzen dat ze deugden. Dat blokkeert identificatie eerder dan dat het haar helpt.
Een voorwaardelijke Nederlander heeft een status zonder kruising. Hij staat in het net, maar niets loopt via hem. Dezelfde vorm komt terug in onderzoek naar jongeren elders. In Afghanistan, Colombia en Somalië gaf niet werkloosheid de doorslag naar geweld, maar onrecht (Mercy Corps, 2015). Een jongere wordt dan wel gezien, maar verkeerd.
Het woord “immigratie” bevat in Nederland vier verschillende problemen (Konstapel, 2025a):
Probleem
Wat wordt gevoeld
Waar het speelt
Wonen
Onrechtvaardigheid
Twintig jaar te weinig gebouwd
Veiligheid
Verlies van vertrouwdheid
Snelle verandering zonder begeleiding
Waarden
Bedreigde levenswijze
Onduidelijke grondwet, falende integratie
Arbeidsmarkt
Economische dreiging
Minimumloon niet gehandhaafd
Een migratieakkoord dat deze vier onder één label houdt, kan geen van de vier sluiten. Elk vraagt een eigen kring met eigen deelnemers. Scheffer wijst zelf op het vierde. Bedrijven profiteren van goedkope arbeidsmigranten, maar bekommeren zich minder om huisvesting en zorg. Hier ontbreekt de ander die de daad afmaakt.
De Staatscommissie Demografische Ontwikkelingen 2050, waarvan Scheffer lid was, adviseert gematigde groei naar 19 à 20 miljoen inwoners in 2050. Dat is een keuze in de trage laag (regel 4). De bevolking is adres. Wat daar wordt besloten, zet het kader voor alles wat sneller beweegt.
9. Wie per maand rondkomt
Het gesprek raakt aan ongelijkheid. Scheffer citeert Martin Wolf: elites moeten meer respect tonen voor wat burgers bij elkaar houdt, en meer belasting betalen. Van den Brink erkent dat de ongelijkheid groter is geworden. Als een samenleving het systeem onrechtvaardig vindt, zegt hij, werkt het niet.
Over de vermogensbelasting is hij voorzichtig. Hij noemt een effectief tarief van 2,1 procent, een van de hoogste in Europa. En hij waarschuwt dat kapitaal mobiel is.
Dat getal is niet bevestigd. Onder het oude forfaitaire stelsel lag de effectieve heffing in box 3 tussen 0,87 en 1,65 procent, afhankelijk van de schijf (Jacobs, 2015). Vanaf 2028 wordt in box 3 het werkelijke rendement belast tegen 36 procent.
Belangrijker is waar de mobiliteit zit. De 1 procent meest vermogenden heeft tachtig procent van het vermogen in box 2: aandelen in eigen bv’s (ESB, 2025). Het debat in het interview gaat over box 3. De open maas (regel 6) is box 2.
Aan de andere kant van de vergelijking staat een andere groep. Campbell en Mankiw maten in 1989 dat ongeveer de helft van de consumptie meebeweegt met het inkomen van de maand zelf. Volgens het standaardmodel van Hall zou dat nul moeten zijn.
Regel 9 geeft dat getal een vorm. Een huishouden dat H maanden vooruit rekent, laat zijn uitgaven voor een deel 1/H meebewegen met het inkomen. Een gemeten waarde van 0,5 past bij een verdeling waarin de helft van de huishoudens niet verder kijkt dan de volgende maand (Konstapel, 2026h).
Dat zijn mensen. Het zijn huishoudens die aan het eind van de maand de huur betalen en dan pas weten wat er over is. Voor hen bestaat een belofte over vijf jaar nauwelijks.
Het standaardmodel van Woodford rekent met een weegfactor van 0,99 per kwartaal. Een belofte voor over vijf jaar, twintig kwartalen, weegt dan nog 0,99²⁰ = 0,82. Met een horizon die volgt uit hoe lang prijzen werkelijk stilstaan, houdt zo’n belofte 8 procent van de kracht van een besluit vandaag (Konstapel, 2026h).
Kapitaal en huishouden handelen dus allebei binnen de kring die hen draagt. Het mobiele kapitaal heeft een lange kring door een smalle maas. Het huishouden heeft een kring van een maand. Een nieuw vooruitgangsverhaal moet in beide kringen landen. Anders blijft het een belofte zonder ontvanger.
10. De herinnering als trage laag
Scheffers boek heeft drie delen: de herinneringsgemeenschap, de vrijheidsgemeenschap en de veiligheidsgemeenschap. In het FD-gesprek ligt de nadruk op de laatste twee. Zonder veiligheidsgemeenschap, zegt hij, geen vrijheidsgemeenschap.
Volgens een bespreking door Alicja Gescinska stelt Scheffer dat duurzame vrede een gedeelde herinneringscultuur vraagt. Duitsland kwam in het reine met zijn wandaden. Willy Brandt knielde in 1970 in Warschau. Rusland deed dat niet. Er staan in Rusland meer dan honderd beelden van Stalin, en er komen er nog bij.
In het net past de herinnering op de trage laag (regel 4). Het is adres. Vrijheid is doorgang door open mazen. Veiligheid is het sluiten van mazen die anders de last doorlaten. De volgorde in Scheffers zin klopt dan zonder oorzaakpijl. Een maas laat pas door als de knopen eromheen houden. Dat is een gelijktijdige voorwaarde, geen opeenvolging.
Een herinnering die niet gedeeld wordt, is een winding die niet sluit tussen twee partijen. Er is dan geen gemeenschappelijk adres om vrede op te bouwen.
Scheffer zei het in 2016 al anders. Een samenleving moet altijd een evenwicht zoeken tussen wat plaatsgebonden is en wat in beweging is. Dat is het onderscheid tussen adres en spanning, in zijn eigen woorden.
Het boek opent met de verwijzing naar wat Scheffer de naoorlogse tijd noemt. Die begon op de conferentie van Jalta in februari 1945. Volgens Hans Wansink eindigt het boek met een pleidooi voor een zo snel mogelijke opname van Oekraïne in de EU. Oekraïne hoort volgens Scheffer feitelijk al bij de veiligheidsgemeenschap die Europa wil zijn.
11. Oekraïne en februari 2028
Scheffer zegt dat de keuzes van nu nog decennia doorwerken. Het model geeft daar een rekenbare grens aan.
De ladder van drie meet tijd in lagen. De maat is de natuurlijke tijdseenheid T₀ = 1,521 × 10⁻¹⁶ seconde. Dat is de helft van de periode van de Rydbergfrequentie. Laag n duurt 3ⁿ × T₀.
Eén jaar is 3,156 × 10⁷ seconde. Dat is 2,075 × 10²³ T₀. De positie op de ladder is log₃ daarvan: 48,87. Een duur hoort bij laag n als de positie tussen n − 0,5 en n + 0,5 ligt. Een jaar ligt dus op laag 49.
De oorlog in Oekraïne begon op 24 februari 2022. Na 4,6 jaar ligt hij op positie 50,26, dus op laag 50. Die laag eindigt bij positie 50,5. Dat is 3^(50,5 − 48,87) = 3^1,63 ≈ 6,0 jaar na de start: rond 21 februari 2028.
De oorlog met Iran begon op 28 februari 2026. Hij ligt op laag 48 en 49. Laag 49 eindigt bij positie 49,5. Dat is 3^0,63 ≈ 2,0 jaar na de start: rond 27 februari 2028.
Beide oorlogen bereiken dus in dezelfde maand een laaggrens. Hun grenzen van zes en twee jaar verschillen precies een factor drie: één laag. Wat op die grens niet gesloten is, hoort daarna bij de tragere laag. Voor Oekraïne betekent dat een winding van ongeveer tien jaar (Konstapel, 2026i).
Dat geeft Scheffers zin een datum. “De tijd staat open”, citeert hij Kouwenaar. In het net staat de tijd open tot februari 2028.
Van den Brinks zorg past in hetzelfde beeld. Lidmaatschap door de strot van de eigen bevolking duwen, zegt hij, gaat backfiren. Marine Le Pen en de AfD voeren campagne tegen de steun aan Oekraïne. Een besluit zonder tweede pad is een open maas onder last (regel 6).
Het HCSS-rapport laat zien hoe zo’n maas eruitziet. Van de 210 miljard euro aan bevroren Russische tegoeden lag 193 miljard bij één instelling, Euroclear in Brussel. Dat is 92 procent in één gebouw, onder één staat. De spanning kwam daar terecht. België blokkeerde de grote lening in december 2025 (Konstapel, 2026c).
Een veilig pad voor Oekraïne vraagt dus tweede paden. Een tweede depot. Een tweede spoorlijn. Een betaalsysteem dat offline werkt. Leden die elkaar kunnen vervangen. In Fiske’s termen is dat het vermogen tot delen en beurtelings geven. Juist die twee vormen van relatie hebben in de huidige strategieën geen bureau.
12. Vier risico’s, één net
In zijn Sijthoff-lezing noemde Van den Brink vier risico’s: geopolitiek, structurele schaarste aan arbeid, kapitaal en energie, klimaatverandering, en ontwrichting door AI. Hij behandelt ze als vier dossiers.
In het net zijn het geen aparte dossiers. Het is één net waarin de spanning door de zwakste maas loopt (Konstapel, 2026i).
Risico
Waar het in het net zit
Geopolitiek
Twee oorlogen die in februari 2028 dezelfde laaggrens bereiken; open mazen in Europees gas en in Euroclear
Schaarste aan arbeid
Het spiegelbeeld van werk dat elders niet telt: 800 miljoen jongeren zonder kring waarin hun werk telt
Schaarste aan energie
Energie als lange keten door smalle corridors; Banga’s eigen les dat regionale afhankelijkheid beter te beheersen is
Klimaat
Water als trage laag onder de warmte; bodem en water als voorwaarde voor elk klimaatbeleid
AI
Een poort die alleen onderbouwd antwoordt of zegt: weet ik niet
De arbeidsschaarste in Nederland en het werk dat elders niet telt, horen bij elkaar. Volgens de Wereldbank bereiken de komende tien jaar 1,2 miljard jongeren in ontwikkelingslanden de werkende leeftijd. Er komen ongeveer 420 miljoen banen bij. Bijna 800 miljoen blijven over (Konstapel, 2026e).
Werk is er genoeg: daken, oogsten, water, zorg. Het telt alleen niet, omdat er geen kapitaal rendement op verwacht. In Nederland is er werk zonder handen. Elders zijn er handen zonder werk dat telt. Beide gaan over een ontbrekende kring tussen werk en de erkenning ervan.
13. Waar de lijnen samenkomen en waar ze uiteenlopen
Van den Brink en Scheffer zoeken allebei een kring. Van den Brink zoekt hem in samenwerking tussen overheid en bedrijfsleven, en in de dialoog met alle lagen van de samenleving. Scheffer zoekt hem in herinnering, veiligheid en een sociaal contract.
Op drie punten sluiten hun lijnen aan op het net.
Aanspreekbaarheid. De kantine in Singapore en de dialoog buiten de directiekamer zijn dezelfde vorm. Het is een korte kring waarin een ander de daad afmaakt.
Akkoorden. Pensioenakkoord, SER, migratieakkoord: het zijn kringen. Ze werken als ze een doorlaat hebben, en een geheugen dat bijhoudt wat ervan terechtkwam.
De trage laag. Herinnering, bevolking en de lange lijnen in Scheffers boek liggen op het adres. Daar wordt het kader gezet.
Op één punt lopen de lijnen uiteen. Van den Brink geeft het bedrijf de rol van verbinder. In de relatietheorie van Alan Fiske werkt een bedrijf met rangorde en prijs. Verbinden hoort bij twee andere vormen: delen en beurtelings geven.
Het model van panarchie legt de verhouding andersom. Daar bepaalt de samenleving wat er gebeurt met wat bedrijven maken (Konstapel, 2026b). Het verschil is geen tegenstelling die moet worden beslecht. Het is een open vraag: in welke kring krijgen delen en beurtelings geven een vaste plek?
Ook Scheffers weigering om recepten te geven past in het model. “Richting schept ruimte”, zegt hij. Een recept is een maakbaarheidsplan. Een richting zet het adres, en laat de sluiting over aan de kringen die volgen. Wat ontbreekt, is de vorm van die kringen.
14. De mensen in het net
Een analyse in termen van windingen en mazen kan te ver van de mensen komen te staan. Daarom op een rij wie hier leeft.
Het huishouden dat aan het eind van de maand de huur betaalt. Voor hen is stabiliteit geen belofte over 2031. Het is een maand die sluit.
De arbeidsmigrant in een distributiecentrum. Hij werkt, maar zijn huisvesting en zorg vallen buiten de kring van het bedrijf dat van hem profiteert.
De jongen in de tweede generatie die steeds moet bewijzen dat hij deugt. Hij heeft een paspoort. Hij heeft geen kruising.
Het personeel in de kantine van de brouwerij in Singapore. Zij stelden een vraag aan een minister en kregen antwoord.
De Oekraïense leraar die maandenlang niet wist of de begroting zou bestaan.
De Singaporese jongere die zijn regering vertrouwt, maar niet gelooft dat zijn kinderen het beter krijgen.
De Nederlander die zijn eigen leven een 7,7 geeft en de politiek een 4,9.
Al deze mensen vragen hetzelfde. Niet om erkenning als status. Om een plek waar wat zij doen door een ander wordt afgemaakt.
15. Conclusie
Het onbehagen waar Scheffer en Van den Brink over spreken, is geen tekort aan welvaart. Het is een tekort aan sluiting. Er zijn te weinig kringen waarin een daad, een belofte of een bewering door een ander wordt afgemaakt.
Het “superieure model” zocht een eindpunt. Het net kent geen eindpunt, alleen sluiting op een plek en een schaal.
Singapore laat zien dat een korte kring vertrouwen geeft. Het laat ook zien dat een korte kring de lange niet vervangt.
Liberale en autoritaire maakbaarheid liggen op één as. Beide missen de 0. In de Nederlandse traditie heeft die 0 een naam: consent, en een vorm: de kring van Boeke.
Een akkoord werkt alleen met een doorlaat en een geheugen. Zonder die twee wordt het een versterker van de verwachting die het niet kan inlossen.
Een belofte reikt niet verder dan de kring die haar draagt. Voor de helft van de huishoudens is die kring een maand.
De herinnering is de trage laag. Wat daar niet sluit, maakt elke vrede wankel.
En de tijd staat open tot februari 2028. Wat dan niet gesloten is, wordt een trage winding van ongeveer tien jaar.
De opgave die uit dit gesprek volgt, is daarom concreet. Niet een nieuw verhaal over de toekomst. Wel kringen op elke schaal, van wijk tot Europa, waarin een daad telt omdat een ander haar afmaakt.
16. Statusgrootboek
Uitspraak
Status
Inhoud van het FD-gesprek
Gemeten: volledige tekst, 15-10-2026
Inhoud van de Sijthoff-lezing
Tweedehands: samenvatting en FD-verslag van 6-11-2024; vier van de vijf suggesties niet beschikbaar
Opbouw en kernstellingen van De dag na de bevrijding
Tweedehands: interview, Wynia’s Week, Gescinska, flaptekst; de boektekst zelf is niet geraadpleegd
Lucassens bezwaren tegen Het land van aankomst
Gemeten: volledige recensie, 2007
Vertrouwen in Singapore en Nederland (Edelman 2026)
Gemeten
Effectief tarief 2,1 procent vermogensbelasting
Niet bevestigd; oud forfait 0,87–1,65 procent
Concentratie in box 2 bij de top 1 procent
Gemeten (IBO 2022, via ESB)
Laaggrenzen in februari 2028
Gerekend, in dit essay nagerekend
0,99²⁰ = 0,82
Gerekend
Euroclear 193 van 210 miljard = 92 procent
Gerekend
Doorlaat Frans klimaatberaad ≈ 0,10
Gemeten (telling Reporterre)
Superieur model als eindpunt, Singapore als korte kring, as zonder 0, herinnering als adres
Lezing in het model, geen uitkomst
Of Scheffer in zijn nieuwe boek op Lucassen ingaat
Open
De vorm van de kring waarin delen en beurtelings geven een vaste plek krijgen
Open
Geannoteerde referenties
Het gesprek en de sprekers
Couwenbergh, P. en Geels, B. (2026). “Paul Scheffer en Dolf van den Brink: ‘Ons Europese model superieur? Dat is behoorlijk arrogant gedacht’.” Het Financieele Dagblad, 15 oktober 2026.Waarom lezen? De aanleiding. Een zeldzaam gesprek tussen een publicist en een topbestuurder die het over de diagnose eens zijn en over de remedie niet. Leesadvies: lees de passage over Singapore naast die over “wen er maar aan”. Daar zit de kern.
FD (2024). “Topman Heineken hekelt polarisatie en ziet verbindende rol voor bedrijfsleven.” Het Financieele Dagblad, 6 november 2024.Waarom lezen? Het verslag van de Sijthoff-lezing, met het deltaplan voor her- en bijscholing als enige uitgeschreven voorstel. Leesadvies: kort; lees het als de voorganger van het gesprek van 2026.
Scheffer, P. (2026). De dag na de bevrijding. Een drieluik over Europa. Amsterdam: Prometheus, 608 p.Waarom lezen? Een ideeëngeschiedenis van Europa sinds 1945, verweven met vijftig jaar eigen ervaring als correspondent en hoogleraar. Leesadvies: begin bij het deel over de herinneringsgemeenschap; daar ligt de trage laag onder de andere twee.
Wansink, H. (2026). “Een rijk boek over een halve eeuw bemoeienis met Europa.” Wynia’s Week, 1 oktober 2026.Waarom lezen? De uitgebreidste bespreking van het nieuwe boek. Laat zien hoe Scheffer van Havemann in Oost-Berlijn naar Matviichuk in Kyiv loopt. Leesadvies: de passage over Jalta geeft het begin van “de naoorlogse tijd”.
Gescinska, A. (2026). “Schandelijke eer.” De Morgen, 2 september 2026; herplaatst door het Humanistisch Verbond.Waarom lezen? De helderste weergave van Scheffers herinneringsgemeenschap, toegepast op Servië en Rusland. Leesadvies: lees de alinea over de beelden van Stalin naast die over Brandts knieval.
Lucassen, L. (2007). Recensie van P. Scheffer, Het land van aankomst. Tijdschrift voor Sociale en Economische Geschiedenis 4(4), 190–191.Waarom lezen? Twee pagina’s die drie zwakke plekken in een invloedrijk boek precies aanwijzen: de ontbrekende vergelijking, het verouderde beeld van ontworteling, en het effect van voorwaardelijk burgerschap. Leesadvies: de laatste alinea over “voorwaardelijke Nederlanders” is de kern.
Het model
Konstapel, J. (2026). “Vacuum.Net Theory: hoe massa, energie en ruimte één net vormen.” constable.blog, 28 september 2026.Waarom lezen? Het net zelf: draad, winding, knoop en maas, en de regel dat wat blijft op zichzelf sluit. Leesadvies: de eerste paragrafen volstaan voor dit essay.
Konstapel, J. (2026i). “Global Tensions 2026: February 2028 as a Turning Point.” constable.blog, 11 oktober 2026.Waarom lezen? De ladder van drie, de laaggrenzen van beide oorlogen, en de lezing van vier regio’s als één net. Leesadvies: paragraaf 3, “The clocks”; reken de twee data zelf na.
Konstapel, J. (2026h). “Waarom een belofte niet verder reikt dan de kring die haar draagt.” constable.blog, 12 oktober 2026.Waarom lezen? De sluitingshorizon toegepast op het werk van de Nobelprijswinnaars Hall en Woodford, met de meting dat de helft van de huishoudens binnen een maand leeft. Leesadvies: het statusgrootboek aan het eind scheidt bewezen van gemeten.
Erkenning, kring en werk
Konstapel, J. (2026a). “De Waarheid Achter Fukuyama’s ‘Einde van de Geschiedenis’.” constable.blog, 12 september 2026.Waarom lezen? Dateert de aanname van erkenning en laat zien waar het “superieure model” zijn eindpunt vandaan heeft. Leesadvies: de paragraaf “De las” over Augustinus.
Konstapel, J. (2026e). “Waarom 800 miljoen jongeren geen banen tekortkomen, maar een kring waarin hun werk telt.” constable.blog, 12 oktober 2026.Waarom lezen? Werk dat niet telt, de vier routes waarlangs spanning ontlaadt, en drie getoetste programma’s. Leesadvies: de paragraaf “Wat aantoonbaar werkt”.
Konstapel, J. (2026d). “The MAZE-Bank: A Radical Reimagining of Economic Systems.” constable.blog, 9 september 2026.Waarom lezen? De kring als geheugen: een daad telt als een ander haar bevestigt. Leesadvies: lees het als uitwerking van regel 8.
Mercy Corps (2015). Youth & Consequences: Unemployment, Injustice and Violence.Waarom lezen? Het bewijs uit drie landen dat onrecht, niet werkloosheid, jongeren naar geweld drijft. Leesadvies: de samenvatting volstaat.
Campbell, J. en Mankiw, G. (1989). “Consumption, Income, and Interest Rates.” NBER Macroeconomics Annual 4.Waarom lezen? De meting dat de helft van de consumptie het inkomen van nu volgt. Leesadvies: de schattingstabellen.
Democratie, consent en doorlaat
Konstapel, J. (2025a). “Fractale Democratie: Van Vertrouwenscrisis naar Wijkcirkels.” constable.blog, 2 oktober 2025.Waarom lezen? De paradox van 7,7 tegen 4,9, de twee blokken die allebei van bovenaf sturen, en het ontwerp van wijkcirkels. Leesadvies: paragraaf IV, met de voorbeelden wonen, energie en zorg.
Konstapel, J. (2026b). “Hoe het Volk de Macht weer krijgt.” constable.blog, 26 maart 2026.Waarom lezen? Consent, de dialoog als gezamenlijke inspanning, en panarchie als model voor de verhouding tussen volk, bedrijf en overheid. Leesadvies: de paragraaf over Kees Boeke.
Konstapel, J. (2026f). “Het Burgerberaad is een Versterker, geen Ventiel.” constable.blog, 18 augustus 2026.Waarom lezen? De doorlaat als maat voor elke kring die adviseert. Leesadvies: lees de cijfers over Frankrijk en Ierland naast elkaar.
Konstapel, J. (2026g). “Het Vergeten Manifest: Waarom de VVD Haar Eigen Diagnose Twintig Jaar Heeft Genegeerd.” constable.blog, 11 juni 2026.Waarom lezen? Een gecontroleerd experiment in politiek niet-leren, en waarom een akkoord een geheugen nodig heeft. Leesadvies: de paragraaf over het ene voorstel dat wel werd uitgevoerd.
Konstapel, J. (2026). “De Neutrale Staat en haar Grenzen.” constable.blog, 14 mei 2026.Waarom lezen? De fractale staat als antwoord op Frissen: condities scheppen zonder uitkomsten voor te schrijven. Leesadvies: paragraaf V.
Konstapel, J. (2026). “Democratie in Vrije Val: Zwemmen in plaats van Verzuipen.” constable.blog, 3 oktober 2026.Waarom lezen? Het overzicht van de hele democratielijn, met een lijst van de onderliggende stukken. Leesadvies: de lijst “Gebruikte blogs” als kaart.
Europa en geopolitiek
Konstapel, J. (2026c). “The Mesh That Stays Open: How Russia Exploits Europe’s Weak Links.” constable.blog, 8 oktober 2026.Waarom lezen? De open-maastoets op Euroclear en Polen, en de twee strategieën als twee handen van één as. Leesadvies: paragraaf 5 met de tabel van Fiske.
Iven, M., Mansfield, E., Eichstaedt, N. en Sweijs, T. (2026). Europe’s Strategic Trap. Den Haag: HCSS.Waarom lezen? De bron van alle data over België en Polen, eerlijk over de grenzen van de eigen toeschrijving. Leesadvies: paragrafen 3.4 en 4.4.
Fiske, A. P. (1992). “The Four Elementary Forms of Sociality.” Psychological Review 99(4), 689–723.Waarom lezen? Delen, rangorde, beurtelings geven en prijs: de vier vormen van relatie die in elke cultuur terugkomen. Leesadvies: dit artikel volstaat; het boek uit 1991 is de lange versie.
McWhinney, W. (1997). Paths of Change. Thousand Oaks: Sage.Waarom lezen? De vier wereldbeelden en de regel dat verandering een gesloten pad door alle vier is. Leesadvies: hoofdstuk 1 tot 3.
Cijfers
Edelman Trust Institute (2026). 2026 Edelman Trust Barometer, met landenrapporten voor Nederland en Singapore.Waarom lezen? Vertrouwen in 28 landen, en het verschil tussen vertrouwen in de regering en vertrouwen in de volgende generatie. Leesadvies: leg de cijfers voor Singapore en Nederland naast elkaar.
Jacobs, B. (2015). “Partiële begrotingsanalyse belastingen.” ESB.Waarom lezen? De berekening van de effectieve vermogensbelasting onder het oude box 3-stelsel. Leesadvies: de passage over de drie schijven.
ESB (2025). “Volgend kabinet moet de belasting op vermogensinkomsten, vermogen en erfenissen pragmatisch hervormen.”Waarom lezen? Waar het vermogen in Nederland werkelijk zit, en hoe ongelijk de belasting op verschillende vormen uitvalt. Leesadvies: de passage over box 2.
Staatscommissie Demografische Ontwikkelingen 2050 (2024). Gematigde groei. Den Haag.Waarom lezen? Het advies over 19 à 20 miljoen inwoners, met de onderbouwing per scenario. Leesadvies: de samenvatting en hoofdstuk 3 over migratievarianten.
Achtergrond
Handlin, O. (1951). The Uprooted. Boston: Little, Brown.Waarom lezen? Het klassieke beeld van migratie als ontworteling, waarop Scheffer in 2007 leunde. Leesadvies: lees het naast Bodnar.
Bodnar, J. (1985). The Transplanted. Bloomington: Indiana University Press.Waarom lezen? Het tegenbeeld: migranten die opnieuw wortelen. Leesadvies: de inleiding geeft het hele argument.
Gunderson, L. H. en Holling, C. S. (red.) (2002). Panarchy. Washington: Island Press.Waarom lezen? De cyclus van groei, vasthouden, loslaten en vernieuwen in natuur en samenleving. Leesadvies: de inleiding en hoofdstuk 1.
Boeke, K. en Endenburg, G. Sociocratische ontwerpprincipes; zie Endenburg, G. (1998). Sociocracy: The Organization of Decision-Making. Delft: Eburon.Waarom lezen? Consent als besluitvorm, en de kring als bouwsteen van bestuur. Leesadvies: de hoofdstukken over consent en dubbele koppeling.
The history of life is one line of closures. Again and again, a form that once lived on its own became a letter of a larger whole. What closes, stays.
How the book reads
The book tells the history of life with numbers. Every case gets a date, a size and a duration. And every length and time gets a place on one ladder: a logarithmic scale with base 3. Lengths are measured against 4.56 × 10⁻⁸ m, times against 1.52 × 10⁻¹⁶ s.
One step up the ladder is a factor of three. The day sits at position 43.50. The year at 48.87. The age of the Earth at 69.11. Because atoms and geological ages share one axis, relations become visible that ordinary units hide.
An example. All covalent bonds of life run from 74 to 204 picometres. That is one step of three. The chemical frame of life fits inside a single trit of the ladder.
The book is self-contained and honest about disagreement. Where the record has two readings — the Late Heavy Bombardment as spike or as argon artefact, sponge steranes or algal ones, 700 or zero new hippocampal neurons per day — it states both and keeps going.
The form language: the Vacuum.Net
On top of the facts sits one reading frame. It is called the Vacuum.Net. Its parts are simple:
The trit. Every crossing of a strand holds one of three values: +1, 0 or −1.
The window. Three trits in a row form a window. Of 27 possible windows, 9 are closed: they add up to zero.
The closure. What closes, stays. What is open carries tension.
Windows that repeat one value three times (+1 +1 +1, 0 0 0) are crystals: order that repeats. Windows that hold all three values once are metabolism: the content changes, the form stays closed.
Several principles run through all 24 chapters:
Form and carrier. A form outlasts its carriers. A water molecule lasts about 11 hours. Its hydrogen bonds exchange every picosecond — 4 × 10¹⁶ times faster. The molecule is the form; the bond is the carrier.
One direction. A living strand has one direction or it does not exist. The rule is symmetric: it does not say which direction wins. It keeps what was laid down first. Life uses L amino acids and D sugars; meteorites carried an L excess of up to 18.5 % before life; autocatalysis raised 0.00005 % to 99.5 % in three rounds.
Rest and discharge. A closure cannot load twice in a row. Every excitable tissue shows this as a refractory period: 3.5 ms in a medusa giant axon, 40–50 ms in a sea-pen net, 29 s in a glass sponge. Extinctions work the same way: long holding, short discharge.
Context. The same rule takes other values in other surroundings. A cell doubles in 37 minutes at 100 °C and in decades in rock. Fire is off below 16 % oxygen and fast above 19 %.
Opened once, filled many times. A new kind of union opens once. Its depth is then filled again and again.
Depths and residues. Each new depth uses the residue of the previous one as its alphabet. A bacterium became a letter of the cell. The cell became a letter of the body. The nerve cell became a letter of the brain. “The residue of depth n is the alphabet of depth n+1.”
The book states plainly that this language is a way of reading, not a cause: the forms are described beside the history, not as its driver.
The story, chapter by chapter
1. Before the cell. The oldest solids of the solar system formed 4,567.30 ± 0.16 million years ago. Within 267 million years — the first 6 % of Earth’s history — liquid water stood at the surface, recorded by a 4,404 Ma Jack Hills zircon. Chemistry without a cell is surprisingly rich: Miller’s archived vials, re-read in 2008, held 22 amino acids; the 1958 sulfide run 23. Cyanide and sulfide under light give nucleotide, amino acid and lipid precursors from one network. RNA catalyses at every size measured: 19 nucleotides (hammerhead), 93 (ligase core), 2,904 (the ribosome’s catalytic core, where no protein atom lies within 18 Å of the bond being made). In 2009 two RNA enzymes were built that make each other indefinitely: the pair, doubling each hour. Fatty-acid vesicles of 65–200 nm grow, compete and divide. The chapter names the net’s basics here: trit, window, ring, depth, edge.
2. The first cell. The record has two tiers. Older, 3.7–4.28 Ga: single signals in heated rock — light carbon in a zircon at 4.10 Ga, Isua graphite, disputed cones. Younger, 3.40–3.48 Ga, in the Pilbara: hot-spring deposits, sulfate reduction, methanogenesis below δ¹³C −56 ‰, microfossils, a 10 km stromatolite reef. LUCA is dated 4.09–4.33 Ga by one clock, with a gene set of 355 families (one reconstruction) to 2,657 proteins (another) — already carrying viruses and a CRISPR defence. The cell is “the first ring that stays closed while all its content changes”: a 4 nm membrane holding −105 to −150 mV, a field of 3.75 × 10⁷ V/m set by about 9,400 ions per µm².
3. Bacteria and Archaea. Prokaryotes alone held the Earth for about 2.36 billion years, and still hold most of its cells: 1.2–6 × 10³⁰ of them, 58 % below the surface. They grow from −15 °C to 122 °C, double in 9.8 minutes to 2,000 years. Genes move sideways: the strict tree of descent holds in about 1 % of prokaryotic genes; 75–88 % of gene family expansions came from transfer. A biofilm of a million cells loads and releases potassium in waves, and a second species answers within 22–26 minutes: community as shared prefix. Prokaryotes carry the element cycles: nitrogen fixation by 3.2 Ga, oxygen in the air from about 2.33 Ga, with Prochlorococcus (2.9 × 10²⁷ cells) making up to 20 % of today’s oxygen.
4. Viruses. Virus hallmark genes put virus-like entities before modern cells, and LUCA’s virome already held the main groups of bacterial and archaeal viruses beside 19 CRISPR effector families. The particle is the crystal: one value repeated in a closed shell, tobacco mosaic virus crystallised like a salt in 1935. Genomes run from 1.7 kb (circoviruses) to 2.5–2.8 Mb (pandoraviruses) — near LUCA’s own reconstructed genome. A giant virus can host a virophage (Sputnik), and partners with their own genome sit at most two deep. In the ocean, 10²³ infections per second kill about 20 % of microbial biomass daily; a particle succeeds in a cell about once per 116 days. Viruses also write themselves into genomes: ~8 % of the human genome is retroviral in origin, and syncytins — captured viral envelope genes — now build the placenta.
5. The cell that took in a cell. The keystone event. The host lineage nests within the Asgard archaea (sister to Hodarchaeales); the mitochondrial symbiosis is dated 2.053–1.21 Ga; the account by Lynn Sagan (1967) is now widely accepted. The bacterium became a letter: against 834 protein genes in Rickettsia, human mitochondria keep 13 — 1.6 % — while more than 95 % of mitochondrial proteins are now coded in the nucleus. The plastid repeated the step: of 3,168 cyanobacterial genes, the Arabidopsis chloroplast keeps 87. Plastids then passed from hand to hand at second and third hand — at least 8 to 10 times — and the serial red chain (red alga → cryptophyte → ochrophyte → haptophyte) shows memory with reach two: each hand inherits only from the one before. Younger cases stand at earlier stages: the Paulinella chromatophore (about 94 Ma, 29 % of genes kept), the Azolla cyanobiont (100 Myr, 68 % intact), the nitroplast UCYN-A (2024: a nitrogen-fixing organelle). In the mealybug a bacterium holds a bacterium — the deepest measured nesting, two deep, like the virophage. Stolen plastids (kleptoplasty) keep working 15 days to 30 months without the algal nucleus.
6. Becoming many. Multicellularity arose more than 25 times in eukaryotes — a serious underestimate, one review says — and first in bacteria: multicellular cyanobacteria before 2.4 Ga, with the Great Oxidation. Bangiomorpha (1.047 Ga) is the oldest recognised crown-group eukaryote, with holdfast and gametes. Hyphae evolved once early in fungi; fruiting bodies many times. Two experiments frame it: a bacterial sulfonolipid at 10⁻¹⁵ M switches S. rosetta to rosettes; 600 days of selection grew snowflake yeast 20,000-fold in size under low oxygen and only 6-fold in the other line — the same rule, other values. The maximum size of life rose 16 orders of magnitude in two steps, at about 1.9 Ga (after the eukaryotic cell) and 0.6–0.45 Ga (after multicellularity), both tracking oxygen. Bodies now span 2 µm to a 3,810 m Armillaria genet: 19 steps of the ladder.
7. The first animals. Four kinds of evidence: molecules, clocks, fossils, living animals. Crown animals by clock: 833.5–649.8 Ma; most bilaterian phyla diversified in the Ediacaran, over 100 Myr before the Cambrian. Sponge choanocyte chambers (16–56 µm) pack 1,876–35,175 per mm³; sponges filter up to 900 body volumes per hour, and individuals stand for thousands of years. The comb jellies keep a nervous system apart from all others (no serotonin, acetylcholine, dopamine; L-glutamate as candidate transmitter). Trichoplax was carried through 2,000–4,000 generations in 11 years of clonal culture and stayed the same sheet. Within some 35 Myr after the first complex fossils (574 Ma), bodies with muscle, tentacles, skeletons and bored holes had all appeared. A recurring note: the parts of the synapse (scaffold, secretion, channels) were whole in single cells before any nerve — the residue became the alphabet.
8. Plants take the light. From cyanobacteria through red, green and brown algae, onto land, into trees and flowers. Prochlorococcus divides once a day, replication in the afternoon, division at night — metabolism held in a day. Land plants: cryptospores at 473–471 Ma; the arbuscular union with fungi at about 480 Ma by clock, “morphologically identical” arbuscules at Rhynie by 411.5 Ma. The oldest forests (Gilboa, Cairo, about 385 Ma) already show three root systems and 8 m trees. Cones and sunflower heads follow Fibonacci parastichies: 73.6 % of 768 counts, and cone-scale quotients cluster near 0.618. Extreme ages: a bristlecone pine of 4,900 years, a Lomatia clone of at least 43,600, a Posidonia genet of up to 200,000. The tallest tree stands near the mechanical limit of 122–130 m. Today’s forests hold about 3.04 trillion trees, 46 % fewer than before civilisation.
9. Fungi. The sister kingdom of animals. Fungus-like filaments in 2.4 Ga basalt; crown Fungi about 1,020 Ma; the arbuscular symbiosis at 480 Ma — before the land plants it now serves, with plant signalling genes already closed in green algae. Plants pass 3.58 Gt C per year into the mycorrhizal mesh; ectomycorrhizal mycelium reaches 2,000 m per cm³ of soil. Lignin-degrading peroxidases closed once, around the end of the Carboniferous, then were lost at least four times; the depth filled with about a thousand white-rot species (the coal question gets both readings, fungal and climatic). Fungus-farming ants went from lower farming (61–57 Ma) to leaf-cutting (19–18 Ma); termites farm once, in Africa. Chytridiomycosis has contributed to the decline of 501 amphibian species, 90 of them extinct in the wild. Named fungi: 120,000–155,000 of an estimated 2.2–3.8 million.
10. Ashore together. The land was not colonised once. Microbial mats by 3.22 Ga; animals burrowed in Ordovician soils before any rooted plant; arthropods walked out with spiracles by about 414 Ma (Pneumodesmus, redated); tetrapod trackways at Zachełmie are 18 Myr older than the first body fossils. Herbivory opened once, at 417 Ma, then filled tissue after tissue. Trees with deep roots cut CO₂ from about 2,000 to 278 ppm over 400–300 Ma; oxygen rose above 26 %, and giant insects flew in it. The chapter’s planetary lesson: the air’s oxygen turns over in 0.8–2.1 Myr, yet charcoal shows the composition held for hundreds of millions of years — the content flows through while the form holds.
11. Earth and life together. The elements in time order, each step carried by one kind of organism: sulfur (sulfate reduction, 3.47 Ga), nitrogen (nitrogenase by 3.2 Ga), oxygen (2.33 Ga, within roughly 10 Myr), carbon, phosphorus, water. The Azolla bloom in the Arctic Ocean (about 48.5 Ma, 800 kyr) drew CO₂ down by 55–470 ppm. Today: NPP 104.9 Pg C per year, split evenly land–sea, while 450 of 550 Gt of living carbon stands on land. Soil turns its carbon over in 15 years and keeps a radiocarbon age of 4,830. Seabirds carry 591 Gg N and 99 Gg P from sea to land each year. People now fix about half of all reactive nitrogen: Haber-Bosch alone is 2.07 times the natural land fixation.
12. Partners. Every union with its date: the lichen before the plant (spores 460 Ma, arbuscule 411.5 Ma); insect symbionts with the smallest cellular genomes (Nasuia 112 kb; Sulcia + Nasuia together make 10 essential amino acids and have both lost their own ATP synthesis); coral-dinoflagellate symbiosis by 212 Ma; Wolbachia in 40–66 % of arthropod species. Each kind of union opens once and fills many times: ectomycorrhiza at least 66 times in fungi, lichenisation 20–30 times, nodulation about 8 times after one predisposition. The human body is a community too: 3.8 × 10¹³ bacteria in 4,644 species against 3.0 × 10¹³ human cells. Coral bleaching, with its median return time now 6 years, is the same form breaking under context.
13. Eating and being eaten. The oldest parasitism: Lower Cambrian kleptoparasites on a brachiopod, 47.8 % infested. Drilling rises through the Palaeozoic, falls almost silent for 120 Myr, returns with a higher ceiling — rest and discharge. Insect herbivory relates to diversification across 31 orders and explains about 25 % of richness; in all five sister pairs the more herbivorous clade is the richer. Plants answered: spines at least 55 times, latex in about 10 % of species, and a wasp reads the plant’s answer to the caterpillar as a sign of the caterpillar — reading. Parasitism itself opened at least 223 times. Today herbivores eat 5.3 % of the world’s leaf tissue per year on land and 51 % of aquatic production; trophic transfer efficiency averages 10 %; and the measured scaling of metabolic rate changes exponent by grade of organisation — no universal allometry.
14. Flowers and their animals. Pollination begins on gymnosperms: long-proboscid scorpionflies (about 165 Ma) on five hosts, a 62-Myr interval. Bees at about 123 Ma; the oldest direct insect pollination of a eudicot at 99 Ma. Fleshy seeds precede their carriers; obligate pairs follow, each dated on the insect side first. Ant dispersal arose at least 101 times, and ant-dispersed lineages hold about 2.24 times the species of their sister groups. Megafaunal fruits survive their carriers: 103 Brazilian species still carry a syndrome for animals that are gone. Pollination networks lose plants only after 70–80 % of random pollinators are removed — but roughly linearly if the most-linked go first. Excluding vertebrate pollinators cuts fruit or seed production by 63 %. Defaunation has cut by more than half the seeds dispersed far enough to track climate change: the fruit stays; its carriers go.
15. Animals spread out. Nine groups, two measures: repeated units and handedness. Segment counts: millipedes to 330 segments and 750 legs (Illacme); a homeotic module of 16 leg pairs; centipede counts odd and stepping by 2. Handedness everywhere: snail faunas 75–99 % dextral, one Partula population 97.9 % sinistral; a snake’s right-jawedness spares 87.5 % of sinistral snails and no dextral ones; fiddler crabs hold 50.3 : 49.7 over 8,088 males; flounder eye sides fixed in 708 of about 715 species; human left-handedness 10.60 % over 2.4 million people. Each body keeps one direction for life; populations differ in which one. Myelin arose at least four times (vertebrates, two annelid lines, copepods); the fastest axon known, a penaeid shrimp fibre of 120 µm, conducts over 200 m/s. The Greenland shark matures at 156 ± 22 years.
16. The great dyings and the refilling. Five events, one pattern. Two ways to fall: raised extinction (end-Ordovician, end-Permian at 60 ± 48 kyr, end-Cretaceous with the impact within 33 kyr) or stalled origination (“mass depletion” in the late Frasnian and end-Triassic). Refilling lags in three bands: years to 300 kyr at the Chicxulub crater (“Life reappeared in the basin just years after the impact”), 0.65–3 Myr for reefs and mammal richness, 8–26 Myr for stable ecosystems and body size. Background loss is about 0.1–2 E/MSY; today’s rate is near 1,000 times that, and on current trends the threshold of the big five could be reached “in as little as three centuries.”
17. The excitable cell becomes a nerve. Every cell measured carries a membrane voltage; the nerve adds nothing new — it is a cell whose edge tension travels along its length. Channels were already in prokaryotes (KcsA: one ion per 10 ns, 10,000 K⁺ per Na⁺; bacterial Na⁺ and Ca²⁺ channels exist). The synapse’s parts predate the neuron; the sponge holds a nearly complete post-synaptic scaffold with dlg’s 13 ligand-binding residues 100 % identical to human. Speeds run 3 cm/s in a sea-pen to 200 m/s in myelinated shrimp. Energy: one spike and its aftermath cost 4.5 × 10⁸ ATP; one bit about 24,000. The FlyWire connectome holds 138 million neurons’ worth of wiring in 149 m of total path; the human brain holds 86 billion neurons, 69 billion of them in the cerebellum. Adult cortical neurons do not divide — the body keeps them for life.
18. Senses, rhythms, memory. A sense is reading: “a homomorphic image that keeps the relations and drops the rest.” The ear reads 20–20,000 Hz; the eye 380–720 nm. Opsins: more than a thousand, nine paralogs already in the bilaterian ancestor; a focused eye can evolve in a few hundred thousand years; the first dated eyes (515 Ma) carry 16,000 ommatidia. The intrinsic human clock runs 24.18 h — 10.8 minutes slow per day, reset by light; food sets the liver’s clock. Memory has two phases (early, protein-free; late, needing transcription) everywhere measured — Aplysia, bee, rat — and its durations run from 20 ms to 50 years in a human “permastore,” 10.9 decades of the ladder.
19. Many minds. Seven groups carry complex brains — insects, octopuses, teleosts, corvids and parrots, whales, elephants, primates — in three great bilaterian lineages, each with its own multimodal centre. The smallest insect keeps a brain 37.4 µm on edge and dissolves the nuclei of 95 % of its neurons; the pilot whale’s neocortex holds about 37 billion neurons, almost twice a human’s; the elephant brain holds 3 times our neurons but one third our cortical density. And the record of consciousness declarations widens by date: mammals, birds, octopuses (2012); insects (2016); all cephalopods and decapods (2021–2022); “at least a realistic possibility” in all vertebrates and many invertebrates (2024).
20. The human. From Toumaï (7.2–6.8 Ma) through Australopithecus (a brain flat at ~450 cc for 1.5 Myr — then a step at 2.10 Ma to eleven times the rate: rest and discharge), Lomekwi tools at 3.3 Ma, fire secured in Wonderwerk at 1.0 Ma. A narrow neck: effective size about 10⁴; Neandertal DNA 1.5–2.1 % outside Africa; Denisovan up to 5 % in Melanesia. Omo at ≥233 ka; Misliya at ≥177 ka. The brain costs 20 % of resting energy at 2 % of mass, 129 g of glucose a day. Each child is a new carrier with about 73 new mutations. Weaning at 2.8 years; menopause shared with wild chimpanzees at Ngogo (0.195 post-reproductive representation, about half ours); the longest verified life 122 years.
21. Language and culture. Codes are older than the human: three monkey calls for three predators that narrow with age; vocal learning in 3 bird and 5 mammal groups; a “Chimpanzee Stone Age” at 4,300 years — and human disturbance erases the traditions before it erases the animals: the form goes before the carriers do. The dated record steps down: eagle talons at Krapina 130 ka; shell beads at Skhul 100–135 ka; the Blombos ochre workshop ~100 ka; rock art at 67.8 ka minimum; flutes at 35 ka. Writing begins as bookkeeping: 85.6 % of early Uruk texts are administrative. From ~800 proto-cuneiform signs to 22 Phoenician letters — each letter the residue of a sign, one alphabet opened once and filled six scripts. Print doubled manuscript output by a factor of 51 in three centuries; world literacy rose from 12 % (1820) to 87.7 % (2024).
22. The human net. One species gathered other species, land, energy and mass into one connected whole. The dog first (two branches, 14,000–6,400 ya); cereals and four livestock in the Near East within 3,000 years; maize by 8,700; rice by 6,900. Twelve thousand years ago nearly three quarters of Earth’s land was already shaped by people. Population doubled eight times since 1800. Human energy use, 600 EJ, equals 15 % of all photosynthesis. Humans and livestock hold 94 % of mammal mass. Megacities rose from 8 (1975) to 33 (2025). Computing information grew at rates doubling in 1.5 to 12 years; one frontier training run equals 8.1 years of El Capitan. Six billion people are online; 11,000 active satellites stand over them. And each body of the net is itself a community of partners — up to 400,000 protein families in a single mouth, most of unknown function.
23. The next depth. The whole history has one movement: a closed form that once lived on its own became a letter of a larger closure. The human net now stands where those larger closures once stood: at an opening. After every opening came a filling — multicellularity 25+ times, eyes at least 40, eusociality 8, nectaries at least 457 — and the filling at the human depth has begun: three regions of domestication, four of writing, 12,000 cities, four exascale machines, 33 models above 10²⁵ FLOP. The fourth step, “reading that counts itself,” widens in the record year by year (2012 → 2024). The next depth, the book suggests, is such a reading carried by many bodies, with each human a carrier as each neuron is in a brain. And every closure leaves a rest — the mitochondrion keeps 13 genes, the body keeps its cortical neurons — and the rest is the alphabet of the next depth. The ladder has no last depth.
24. The ladder of life. 2,702 positions on one scale: 767 lengths, durations, ages. From an atom’s diameter (−5.57) to a 3,810 m fungal genet (22.89); from a hydrogen-bond exchange (9.06) to the first solids of the solar system (69.12). “A ladder driven by a rest cannot end.”
What makes the book unusual
Measurability. Every case is a row in a catalogue of 8,761, with value, position and source. Every calculation is shown.
One scale. Atoms, bodies and geology become commensurable.
One pattern. Closure, rest and discharge, one direction, context, opened-once-filled-many, form and carrier — from crystal to metabolism to brain to net.
Fairness to the record. Controversies are stated as such, both readings kept, and the form language is offered as a reading, not a cause.
An annotated reference list
The book cites over 2,000 numbered sources. Below are the ones that carry its most important claims, with a note on what each does in the book. Bracketed numbers are the book’s own reference numbers.
The early Earth and the first chemistry
[1] Connelly et al. (2012), Science. Uranium-lead chronology of the oldest solids: 4,567.30 ± 0.16 Ma. The anchor date of the whole timeline.
[2] Wilde et al. (2001), Nature. The 4,404 Ma Jack Hills zircon with heavy oxygen: the earliest evidence for continental crust and oceans.
[4] Mojzsis et al. (2001), Nature. Oxygen-isotope evidence for surface water interacting with crust by 4,300 Ma.
[6] Abramov & Mojzsis (2009), Nature. The Hadean crust stayed habitable through the bombardment: “no plausible situation in which the habitable zone was fully sterilized.”
[5] Boehnke & Harrison (2016), PNAS. The 3.9 Ga impact spike as an argon-loss artefact. The book keeps both readings side by side — a model of its honesty.
[38], [39], [41] Miller (1953); Bada et al. (2008, 2011). The spark experiments and the re-reading of Miller’s archived vials: 5 amino acids in 1953, 22 in 2008, 23 with sulfur in 1958. What changed was the detector, not the chemistry.
[84] Lincoln & Joyce (2009). Two RNA enzymes that build each other indefinitely, doubling in about an hour: the book’s measured “pair.”
[91], [90] Hanczyc et al. (2003); Chen et al. (2004–05). Clay-catalysed vesicle growth, osmotic competition and division without loss of contents: protocell behaviour in chemistry.
Cells, genes and the two domains
[96] Bell et al. (2015). Graphite inclusions in a 4.10 Ga zircon with δ¹³C −24 ‰: possibly the oldest trace of life.
[113] Weiss et al. (2016), Nature. LUCA as anaerobic, CO₂-fixing, H₂-dependent, with 355 gene families.
[97] Moody et al. (2024), Nature Ecology & Evolution. LUCA at 4.09–4.33 Ga, a genome of 2.75 Mb with 2,657 proteins and CRISPR effectors: the cell opened once and has been filled ever since.
[170] Eme et al. (2023), Nature. Eukaryotes nested with high confidence within Asgard archaea, sister to Hodarchaeales: the host of chapter 5 identified.
[164], [194], [195]. Lateral gene transfer counts: the strict tree holds in ~1 % of prokaryotic genes; the rest is a net.
The symbioses
[341] Andersson et al. (1998), Nature.Rickettsia prowazekii as the nearest relative of mitochondria: the bacterium that became a letter.
[347] Anderson et al. (1981) and successors. Human mtDNA: 16,569 bp, 13 protein-coding genes. The “residue” kept by the letter.
[352] MitoCarta3.0. 1,136 human mitochondrial genes, of which 13 are organelle-encoded: 1.1 %. The quantification of the letter’s reduction.
[336] Reyes-Prieto et al.; [357], [361]. Plastids from a single primary event, then second- and third-hand transfers: opened once, filled at least 8–10 times.
[210] Coale et al. (2024). UCYN-A described as a nitrogen-fixing organelle (the “nitroplast”): the letter-wording still under way, caught in the act.
Multicellularity, land and partnership
[410] King (2004); [415], [416] and successors. Multicellularity 25+ times; multicellular cyanobacteria before 2.4 Ga.
[407] Alegado et al. (2012). A bacterial sulfonolipid at femtomolar concentration switches S. rosetta to rosettes: another organism sets the switch.
[470] Bozdag et al. (2023). 600 days of snowflake-yeast selection: 20,000-fold size gain under low oxygen, 6-fold in the other line. Context, measured.
[913] Floudas et al. (2012), Science. Class II peroxidases and the origin of lignin decay around the end of the Carboniferous.
[905] Hawkins et al. (2023). 3.58 Gt C per year from plants to mycorrhizal fungi: the union of Rhynie, still flowing.
[566] Hughes et al. (2018), Science. Coral bleaching return time down to 6 years: a closure breaking above a fixed line.
Nerves, senses, minds
[477] Sakarya et al. (2007). The sponge’s nearly complete post-synaptic scaffold, with dlg PDZ residues 100 % identical to human: the synapse before the neuron.
[479] Moroz et al. (2014). The comb-jelly genome: a nervous system built on different transmitters, evolved apart.
[1619] FlyWire (2024). The full fly brain connectome: 138 million neurons mapped, 149 m of wiring.
[1518] Azevedo et al. (2009); [1516] Herculano-Houzel (2014). 86 billion human neurons; the elephant brain with 3 times our neurons at one third our cortical density.
[1723] Czeisler et al. (1999). The intrinsic human clock at 24.18 h, in both young and old: the form held for life while cortisol clears 22 times per turn.
[1836] The New York Declaration on Animal Consciousness (2024). “At least a realistic possibility” in all vertebrates and many invertebrates: the widening record of minds.
The human line
[1842] and related dating studies. Toumaï bracketed at 6.8–7.2 Ma; Omo at ≥233 ka; Misliya at ≥177 ka.
[1861] Florio et al. (2015), Science.ARHGAP11B: a single human-lineage gene that folds the mouse cortex.
[1865] Berna et al. (2012), PNAS. Wonderwerk Cave: the earliest secure evidence of fire, about 1 Ma.
[1948] Vanhaeren et al. (2006). Skhul shell beads at 100–135 ka; [1949] Henshilwood et al. (2011), Science. the Blombos ochre workshop: symbolic culture, dated.
[1965] Hawkins, L. (2019), Unicode proposal. The proto-cuneiform counts: 85.6 % of the earliest Uruk texts are administrative — writing born as bookkeeping.
[1985] Bar-On et al. (2018); [1984] Elhacham et al. (2020). Humans and livestock at 94 % of mammal mass; human-made mass equal to living biomass: the human depth, weighed.
[2003], [2006] TOP500 and Epoch AI data. Four exascale machines, 33 models above 10²⁵ FLOP: the newest layer of the net, counted.
How to read the sources
The factual layer rests on peer-reviewed science — Science, Nature, PNAS, and standard references (Stryer, Milo & Phillips, NIST, OpenStax) — with each number traceable to its study. The Vacuum.Net form language is not part of that literature: it is the author’s own reading frame, offered as a lens and not as a cause.
Summary based on a full reading of: J. Konstapel, “The History of Life on the Net — A Reference Work on Evolution and the Vacuum.Net”, Constable Research, Leiden, first edition 14 October 2026, version 1.0 (684 pages, 24 chapters, 8,761 catalogue cases, 2,702 ladder positions).
Animals on the Ladder presents itself as a reference work with a single question at its core: whether the evolution of animals correlates with the evolution of “the net” as described by the Vacuum.Net theory, without adding any assumption to that theory. The book is built around this comparison. Its author defines a logarithmic ladder of lengths and times based on natural units, places the fossil record of fifteen animal groups upon it, and asks whether the pattern that emerges matches the pattern the theory predicts.
The most consequential finding of the book, however, is not a confirmation. It is a documented failure. Three tests, fixed in advance from the theory’s own texts and run against the assembled data, all fail — and the book reports this outcome plainly, in its own words and numbers. Across the five volumes in which the author has set the theory against biological data, the cumulative record now stands at one passed test and nine failed.
A reader who stops at that score misses what the book actually delivers. This essay sets out the insight the book offers in three layers: as a synthesis of current knowledge on animal origins, as a demonstration of evidentiary discipline, and as a case study in how an out-of-mainstream hypothesis should be treated — including by its own author.
1. A Current, Well-Sourced Account of Animal Origins
Stripped of the ladder, Parts 2 and 3 of the book constitute a compact and careful survey of what is presently known about when the major animal groups appeared, and how securely it is known.
The entries do not merely list dates. They distinguish the classes of evidence and their reliability: biomarkers (the 24-ipc steranes of the Huqf Supergroup, and the dispute over whether they are diagnostic of sponges at all), disputed body fossils (the ~890 Ma Little Dal “possible poriferans”, the ~760 Ma Otavia, the ~600 Ma Encyathispongia), accepted fossils (the Ediacaran crown-group sponge Helicolocellus, the first uncontested Lower Cambrian spicules), and molecular clocks with their confidence intervals under alternative calibration schemes. Competing phylogenetic readings — such as whether ctenophores or sponges are sister to all other animals — are presented side by side rather than resolved by preference.
This layer has practical value independent of the theory. The recurring gaps the book itself catalogues in Part 6 — generation times for sponges, cnidarians and molluscs, metabolic exponents for most invertebrate groups, numeric ages for several key fossils — read as a serviceable map of where the empirical record is thin. Anyone assembling a dataset on animal evolution gains from this both a starting inventory and an honest account of its uncertainties.
2. An Operational Model of Evidentiary Discipline
The book’s methodological apparatus is its most distinctive feature. Four marks govern every statement: Text (quoted from a Vacuum.Net text, fetched and cited), Measured (read in the cited source, opened by the author), Computed (arithmetic on text or measured numbers, shown), and Read (a pairing made by the book itself between a text statement and a measured fact — explicitly a hypothesis, kept apart from the numbers it pairs). Section 1.8 inventories what the theory does not say, listing absences checked by search: no text names a body plan, a segment, a vertebra or a shell; no text links a radial order to a ring length. Each entry of Part 2 closes with a “Not verified” list of items found only in secondary sources — items the book then excludes from its own facts.
This apparatus is not decorative. It performs the function that distinguishes a checkable work from an unfalsifiable one: a reader who rejects any given interpretation (“read”) retains the underlying numbers, marked and sourced. The discipline extends to the author’s own errors — the book reports that an earlier count of Test H2 had dismissed the relevant rows as not applicable, and corrects itself against its own interest.
For the working reader, the transferable lesson is concrete. Label what a source says separately from what you conclude from it. Keep your interpretation severable from your data. Publish the list of things you could not verify. These practices cost little and raise the credibility of any document, mainstream or not.
3. The Failed Tests as the Book’s Real Result
Part 5 is the decisive section. The three tests were fixed on the day of the book’s completion, before any dossier was written, and run the same day:
H1 — one direction per body. Required: no species in which the minority mirror form reaches at least 10 per cent of a natural population and persists or is heritable. Found: at least nine species meeting that description, from scale-eating cichlids to left-coiling snail populations stable for decades.
H2 — the first direction is kept. Required: at least 90 per cent of species per clade retain the ancestral handedness, in at least three clades. Found: four of six counted snail families fall below the threshold.
H3 — growth by whole triples. Required: in at least 80 per cent of species with multiple within-species counts of a repeated unit, the smallest difference is a multiple of three. Found: in 27 species examined, not one showed a multiple of three — an outcome below the roughly one-in-three rate expected by chance.
All three fail, and Section 5.4 places them in the wider ledger: fifteen pre-registered confrontations with biological data across five books, of which one passed and nine failed outright.
The insight lies in the diagnosis the book offers for these failures: each test carried a statement about rings of trits to the level of whole animals — a species instead of a strand, a segment instead of a trit — and at that level the form of the theory did not appear. This is a meaningful result in itself. It bounds the hypothesis: whatever the theory may describe, the animal-level predictions drawn from it do not hold, and the failure is now on record with the protocol fixed in advance.
A negative result of this kind is more informative than an untested correlation. It cannot be quietly reinterpreted after the fact without abandoning the pre-registration, and it tells future readers exactly where not to look. The one passed test (nesting depth, in The Weaving of Life) marks the sole data point that remains standing for the theory’s biological claims.
4. What the Book Does Not Deliver
Precision requires stating the limits. The central correlation — that animal evolution tracks the evolution of the net — is not established by this book, and its own tests weigh against it. The Vacuum.Net theory is not part of accepted physics or biology, and the ladder itself is a coordinate system of the author’s construction; placing numbers on any logarithmic axis will produce apparent regularities. The book does not claim causation, and Part 6 candidly lists what remains unverified. A reader seeking a validated account of why animal forms exist will not find it here, and the book does not pretend otherwise.
Conclusion
Animals on the Ladder is best read not as a demonstration of the Vacuum.Net theory but as an unusually transparent record of what happens when a non-mainstream hypothesis is treated with the tools of mainstream science: pre-registered tests, severable interpretation, sourced and re-verified numbers, and outcomes reported against interest. The book delivers three things of lasting use. First, a disciplined synthesis of current evidence on animal origins, with its uncertainties mapped. Second, an operational template for evidentiary marking that any author can adopt. Third, and most unusually, a pre-registered failure that honestly bounds the theory it was written to serve.
The reader learns most, in the end, from the book’s willingness to let its own hypothesis lose.
Annotated References
Konstapel, J. (2026). Animals on the Ladder: A Reference Work on the Evolution of Animals, Set Beside the Evolution of the Net. Constable Research, Leiden, 14 October 2026, version 1.0. Primary source of this essay. Defines the four evidence marks (Text, Measured, Computed, Read) in Part 0; the ladder of lengths and times in §0.2; the theory’s own statements in Part 1, including the inventory of what the texts do not say (§1.8); the fifteen animal-group entries in Part 2; the cross-group tables of Part 3; the correlation statements with status in Part 4; and the three pre-fixed tests and the cumulative fifteen-test ledger in Part 5, §5.1–5.4.
Konstapel, J. (2026). Plants on the Ladder. Constable Research, Leiden, 12 October 2026. Companion volume and source of Parts 1.1–1.6, taken over unchanged with the theory texts fetched and their numbers recomputed. Supplies the plant side of the comparison and part of the earlier test record counted in §5.4.
Konstapel, J. (2026). One Direction or None: Mirror Molecules, the 2026 Nobel Prize in Chemistry and the Closure Rule of the Vacuum.Net (Dutch version: Waarom het Leven maar Eén Kant op Draait). Constable Research, Leiden, 10 October 2026. Source of the direction statements quoted in §1.7: a living strand has one direction or does not exist; the direction is fixed after two steps; the net keeps the direction of what was laid down first. These statements are the fixed basis of tests H1 and H2, and their failure at the level of species and clades is the essay’s central case.
Konstapel, J. (2026). Three Chains, One Standstill. Constable Research, Leiden, 10 October 2026. Source of “a ring grows only by a whole triple” — the statement tested (and failed) as H3 across 27 species with repeated-unit counts.
Konstapel, J. (2026). Life and Consciousness from One Rule: Closure in the Net. Constable Research, Leiden, 29 September 2026. Source of the ring-depth readings used in §§1.2 and 4.5: depths 1–2 as chemistry, depth 3 as self-maintenance, depth 6 as the single cell, and the ring of six dividing into two rings of three that keep the parent’s handedness.
Konstapel, J. (2026). The Balanced Three, the Balanced Tree and the Vacuum.Net Theory. Constable Research, Leiden, 29 August 2026. Source of the Bronze Mean recursion b(n), the growth constant σ₃ = (3 + √13)/2, and the two-layer law separating σ₃ (address, topology) from φ (tension, time). Also §11, the source of the “evolution as filling of depths” statement discussed in Part 4.
Konstapel, J. (2026). The Climate on the Ladder of Three. Constable Research, Leiden, 3 October 2026. Source of the axiom set, the closure rule on windows of three, the necklace counts g(n), and the closure depths with the Pythagorean-comma residue at depth 12 (§3.2, quoted in §1.6).
Konstapel, J. (2026). The Vacuum.Net Theory: Complete Derivation, 28 September 2026. Source of the five axioms N1–N5 (strand, closure, memory, context, scale) quoted in §1.1.
Konstapel, J. (2026). Where Illness Comes From, ninth edition. Constable Research, Leiden. Source of the transition statements quoted in §1.5: a transition as “the next permitted closure depth being reached”, and the remainder that guarantees the ladder keeps offering next rungs.
Konstapel, J. (2026). From Atom to Body. Constable Research, Leiden, October 2026. Source of the ladder’s natural units and, importantly, of its own clustering test: measured human body sizes do not cluster near whole positions (192 of 388 within a quarter step, against 50 per cent expected by chance). The book’s decision to treat positions as coordinates, not marks, follows from this result — an example of the theory’s own tools being used to restrict its claims.
Konstapel, J. (2026). The Universe of Peter Rowlands, 29 June 2026. Quoted in §1.4 as the source of the nilpotent condition Ψ² = 0 (“the totality of everything is nothing”) behind the two-layer law; the cross-reference between this text and the Balanced Tree text is made, and marked as, a “read”.
Konstapel, J. (2026). The Company of Plants; The Weaving of Life; From Cell to Mind. Constable Research, Leiden, 13–14 October 2026. With Plants on the Ladder, the corpus whose twelve earlier tests Part 5.4 counts alongside this book’s three. The Weaving of Life contains the single passed test of the fifteen — the nesting-depth result — and is cited here for the cumulative ledger, not for any claim about animals.
Zhang, Z.-Q. (2013). “Animal biodiversity: an update of classification and diversity in 2013.” Zootaxa 3703(1), 5–11. DOI 10.11646/zootaxa.3703.1.1. External census cited throughout Part 2 and in Table F for the ~1.66 million described animal species across 40 phyla; an example of the book’s practice of anchoring counts in published, verifiable sources.
Blindheid ontstaat wanneer ergens in het visuele systeem iets uitvalt. Dat kan de toegang van licht tot het oog zijn, de lichtgevoelige cellen van het netvlies, de zenuwbanen of de hersenen die beelden verwerken. De plaats van de beschadiging bepaalt welke behandeling mogelijk is.
Bij staar en afwijkingen van de oogsterkte bestaan al effectieve oplossingen. Bij beschadiging van het netvlies bieden gentherapie, lichtgevoelige eiwitten en elektronische implantaten nieuwe mogelijkheden. Ook wordt onderzocht hoe beschadigde zenuwcellen kunnen herstellen en hoe signalen rechtstreeks naar de hersenen kunnen worden gestuurd. De resultaten variëren van beperkt lichtzien tot het herkennen van eenvoudige vormen en letters.
Een aanvullende mogelijkheid is beelden omzetten in geluid of tastbare signalen. Daarmee kan iemand zich beter oriënteren zonder dat het gezichtsvermogen zelf terugkeert.
De Vacuum.Net-theorie beschrijft zien als een netwerk van drie functies: licht opvangen, signalen doorgeven en informatie verwerken. Vanuit dit model ontstaat een nieuw voorstel voor implantaten: stimuleer niet alleen een lichtpunt, maar ook de omgeving ervan. Het contrast tussen licht en donker zou vormen en randen herkenbaarder kunnen maken. Of dit daadwerkelijk beter werkt, moet nog experimenteel worden aangetoond.
De hoofdconclusie is dat blindheid geen enkelvoudig probleem is. Soms bestaat de behandeling al, maar ontbreekt de toegang ertoe. In andere gevallen moet de techniek nog worden ontwikkeld.
English Essay
IntroductionAround 43 million people worldwide are blind, and most of them could once see. Behind that single word — blindness — lie very different problems: a clouded lens in one country, dying photoreceptors in another, a severed optic nerve in a third. Each demands its own remedy, and a remedy that works at one point of failure does nothing at another.
This essay presents a structured map of blindness and vision restoration. It is organised around a simple principle: seeing breaks at a specific layer, and each treatment works by restoring that layer, bypassing it, or reaching the brain by another route. The framework is drawn from the author’s Vacuum.Net theory, which describes vision in five functional layers. The result is a clear-eyed view of what medicine can do today, what is in trial, and where the field is heading as of October 2026.
The Five Layers of Seeing
The theory describes vision as a chain of closures. Light must first pass an opening (the passage); it must be caught by a light-sensing window; the signal must be passed along an edge; it must be handed over at a relay; and finally it must be held as an image in a reading ring. In anatomical terms:
Layer
Function
Anatomy
0. Passage
An opening through which light reaches the window
Cornea, lens, vitreous
1. Window
A closure with light
Photoreceptors and the RPE
2. Edge
Passing the signal on
Bipolar cells, ganglion cells, optic nerve
3. Relay
Handover to the reading ring
LGN in the thalamus
4. Reading ring
The image itself
Visual cortex
Every major cause of blindness breaks exactly one of these layers, and every treatment — from a ten-dollar pair of glasses to a thousand-electrode brain implant — can be classified by the layer it restores.
The Scale of the Problem
The Vision Loss Expert Group counted 43.3 million blind people in 2020, alongside 295 million with moderate or severe vision impairment. Without intervention, blindness is projected to reach 61 million by 2050. Among people over 50, the leading causes are cataract (45.5% of the 33.6 million blind in that age group), glaucoma (11%), undercorrected refractive error (6.6%), age-related macular degeneration (5.6%) and diabetic retinopathy (2.5%).
One finding stands out. Cataract and refractive error together account for roughly half of all blindness in people over 50 — and both have a safe, cheap, well-understood remedy. The bottleneck is not knowledge but reach: in Burundi, effective cataract surgical coverage is just 2.1%, against 77.7% in Qatar. Roughly half of those worldwide who need cataract surgery cannot get it; in Africa, three in four go untreated.
In the Netherlands, the picture is different but structurally similar: a 2008 count found 77,000 blind people, 59,000 of them aged 65 or over, with AMD (38%) and cataract (23%) leading. Around a third of the blind elderly are blind for want of a lens, glasses or an operation — not for want of a technique.
Layer 0: The Passage — A Problem of Delivery, Not Discovery
When the retina is healthy but light cannot reach it in focus, the remedies are established and inexpensive. Glasses and laser correction address refractive error; cataract surgery replaces the clouded lens (141,134 operations in the Netherlands in 2025, with vision improved in 94.1% of cases). Myopia control in children is advancing: high-dose atropine slows progression by about 67%, and defocus lenses reduce eye growth by 62%.
For the cornea, donor tissue remains scarce — one donor cornea for every 70 needed, with 53% of the world lacking access to transplantation — but new routes are emerging: cultured endothelial cells (one donor pair yields roughly a thousand doses; approved in Japan, in US phase 3), a bioengineered cornea of pig collagen (all 14 blind keratoconus patients in the trial regained sight), and the Boston keratoprosthesis (over 19,000 implanted, 74% retained at five years).
The lesson of this layer is commercial as much as medical: the products exist, the outcomes are proven, and the unmet need is a distribution and affordability problem.
Layer 1: The Window — Where Real Innovation Is Happening
Here the photoreceptors themselves die, through inherited dystrophies (retinitis pigmentosa, Stargardt disease, Leber congenital amaurosis), AMD, or diabetic retinopathy. Most current treatments only slow the loss: anti-VEGF injections for leaking vessels, complement inhibitors for geographic atrophy (slowing lesion growth by 14–22% but without clear acuity gains), and strict glycaemic control, which cuts new diabetic retinopathy by 76%.
Restoration, where photoreceptors are already lost, follows three distinct strategies:
1. Restore the missing part of the window. A photoreceptor’s light sensor has two fixed components: the opsin protein and the chromophore retinal. Gene therapy can supply a broken one. Luxturna, the first approved gene therapy for the eye (FDA 2017), does exactly this for RPE65 mutations: 13 of 21 treated patients passed a mobility test at 1 lux, none of the untreated. Successor programmes — laru-zova for X-linked RP (positive phase 3 readout, September 2026), CRISPR editing, RNA therapy — extend the same logic. Crucially, this works only while cells are still alive.
2. Place a new window further along the chain. When photoreceptors are gone, optogenetics gives the surviving bipolar or ganglion cells a light-sensing channel. GenSight’s GS030 restored object localisation to a man blind for 40 years (with goggles). Nanoscope’s MCO-010 works in ordinary light and gained about 2.6 lines of acuity over four years; an FDA decision is due in 2027. Restore Vision’s RV-001 moved all three high-dose patients from no light perception to light perception within a month. And Prima (Science Corp) — a 2×2 mm photovoltaic chip under the macula driven by infrared glasses — gained a mean 25.5 letters in 38 patients with geographic atrophy, 84% of whom could read letters again. Prima carries a CE mark since July 2026 and is under FDA review.
3. Transplant new cells. This remains the hardest road. A 2016 triangulation by three independent groups showed that only 1.2–4.7% of apparent photoreceptor “integration” was genuine; the rest was material transfer into host cells. RPE patches fare better (OpRegen gained up to 9 letters where coverage was extensive), because the RPE sits behind the window rather than forming part of it.
A key engineering insight from the theory: sensitivity is not a matter of amplification chains but of how long a single “window” stays open. Fast channels need bright light (goggles); slow pigments work in ambient light but smear motion. This trade-off sorts today’s products into two families, and it explains why melanopsin — extremely slow and sensitive — can restore light detection but not images.
Layer 2: The Edge — The Hard Problem
Glaucoma, the most common cause of irreversible blindness, kills ganglion cells and their axons: the windows still close, but the signal no longer travels. Everything available today is slowing, not restoring. Lowering eye pressure cuts progression — a 25% pressure reduction halved progression rates in the landmark EMGT trial, and each 1 mmHg of reduction lowers risk by about 10% — but lost ganglion cells do not return.
Restoration is only now entering first-in-human trials. The most notable is Life Biosciences’ ER-100, which uses partial cellular reprogramming (the 2020 mouse work restored vision in a glaucoma model). Its first human data, reported 8 October 2026, showed visual-field improvement in 2 of 3 glaucoma patients at the lowest dose — an uncontrolled first signal, not a result. Nicotinamide is in phase 3. For LHON, gene therapy (Lumevoq) showed bilateral effects that confounded its EU application, though it is available to named patients in France at €425,000 per injection.
Layers 3 and 4: Writing Directly into the Brain
When everything in front is gone, the remaining option is to write directly into the relay (the LGN) or the cortex. This is the route of Phosphoenix, the company founded by Pieter Roelfsema that prompted the original essay: a cortical/LGN prosthesis with over 1,000 electrodes, built on a 2020 macaque experiment in which stimulated electrode patterns formed recognisable letters. A five-patient human study begins at Amsterdam UMC in early 2027, with market entry foreseen around 2032. A Harvard path (Pezaris) showed as early as 2007 that LGN microstimulation produces reliably localised phosphenes; a first human test is hoped for in late 2027.
Cortical implants — Orion, CORTIVIS, ICVP — have shown that humans can localise, count and even read simple letters through direct stimulation, but channel counts remain the binding constraint: simulations suggest 600–700 electrodes are needed merely for reading and walking, and around a million to match the retina’s output.
Where the cortex itself is damaged — stroke, cerebral visual impairment in children — implants cannot help, but training can. Rigorous daily training of the blind field (Huxlin, Rochester) raised task accuracy from about 50% to about 80%, working through a secondary route from the LGN to area MT that bypasses the destroyed V1. This is the mechanism behind “blindsight.”
The Design Rule Nobody Uses Yet
The essay’s most forward-looking contribution is a design principle drawn from the theory’s register of three values (+1, 0, −1). The retina naturally computes in three positions: an ON pathway, an OFF pathway, and centre-surround fields. Yet every clinical device reviewed — every electrode array and every clinical opsin — writes only +1. None writes −1.
Evidence from blind mice (Macé et al., 2015) shows the opportunity: exciting only ON-bipolar cells restored both ON and OFF responses, because the retina’s own circuit supplied the inhibition. The proposed rule is therefore: write a phosphene as +1 at its centre with −1 around it, at the address from which the native circuit forms the −1 itself. Whether a centre-surround phosphene is perceived more sharply than a dot has not been tested — an open, testable hypothesis and a plausible differentiator for the next generation of devices.
Substitution: Reaching the Reading Ring by Another Route
Finally, the reading ring keeps relations, not sources. Echolocation training (ten weeks of mouth-clicks) made the primary visual cortex of both blind and sighted adults respond to echoes, and cut blind participants’ maze-completion time by 58%. The vOICe converts images to soundscapes that activate the visual shape area. BrainPort renders camera images on the tongue (69% of year-long users passed object recognition). Be My Eyes passed one million blind and low-vision users in March 2026.
These routes work for every cause, at every layer, without surgery — and they work best for people who once saw, because their reading ring was already filled.
Conclusion: Three Populations, Three Strategies
Read through the five layers, the field acquires a plain commercial logic:
Half the world’s blind need a lens or glasses. The remedy exists, is cheap and is safe. The gap is delivery and financing, and closing it is a logistics and policy task, not a research one.
Those who lost their windows (RP, dry AMD) are the beneficiaries of genuine breakthroughs: Prima carries a CE mark in Europe, MCO-010 awaits an FDA decision in 2027. What returns is coarse — but it is sight.
Those whose edge was cut (glaucoma patients above all) still have no way back. For them, everything rests on next-generation work — reprogramming, LGN implants — reaching the clinic early in the next decade. Until then, sensory substitution and AI are what is available, today.
The strategic lesson generalises beyond ophthalmology: match the intervention to the layer that is broken, and be honest about which layer a product actually serves. A field that orders itself this way — from the $100 lens to the million-electrode cortex — becomes, as this essay demonstrates, not a list of miracles and failures, but a map.
Annotated References
Each entry gives the source, its role in the essay, and guidance on reading it.
The occasion
Gras, A. (2026). “Ik dacht altijd: blinde mensen weer zicht geven is sciencefiction.” Het Financieele Dagblad, 14 October 2026. The interview with Pieter Roelfsema that prompted the essay: the company, the funding and the 2032 plan. Worth reading for the passage on investors, which shows how a research lab becomes a medical product.
Chen, X., Wang, F., Fernández, E., and Roelfsema, P. R. (2020). “Shape perception via a high-channel-count neuroprosthesis in monkey visual cortex.” Science 370, 1191–1196. The macaque experiment with 1,024 electrodes behind Phosphoenix: simultaneous stimulation produced letters, sequential stimulation produced motion. See the figure in which a letter is built from 8 to 15 phosphenes.
CORDIS (2025). “SIGHTED”, EU project 101212687. The EU-funded plan for the first five patients with the LGN implant, early 2027.
The theoretical framework
Konstapel, J. (2026). One Window: Light-Gated Channels, the 2026 Nobel Prize in Medicine, and the Closure Rule of the Vacuum.Net. Constable Research, Leiden, 10 October 2026. Source of the window with one free place, the three-position register, and the address/tension distinction. Sections 3, 6 and 7 are most relevant; note that the essay corrects its “factor fifty” claim.
Konstapel, J. (2026). From Cell to Mind: How Excitable Cells, Nerve Nets, Brains and Minds Evolved. Constable Research, Leiden, 14 October 2026. The biology of excitable cells and senses, kept separate from the theory. Chapter A6 holds the response-time data used in the speed/sensitivity analysis.
Konstapel, J. (2026). Life and Consciousness from One Rule: Closure in the Net. Constable Research, Leiden, 29 September 2026. The ring of 3 and 6, the membrane as edge, and the reading ring as an image that keeps relations. Sections 4, 6 and 7.
Konstapel, J. (2026). Vacuum.Net Theory: hoe massa, energie en ruimte één net vormen. constable.blog, 28 September 2026. The mesh as an opening between strands, on which the passage layer is read.
Konstapel, J. (2026). What Fits, Stays: Designer Molecules, the Closure Rule of the Vacuum.Net, and Three Simple Roads to a Medicine. Constable Research, Leiden, 10 October 2026. The “mesh instead of a substance” road, which optogenetics and Prima follow.
The size of the problem
GBD 2019 Blindness and Vision Impairment Collaborators and Vision Loss Expert Group (2021). “Causes of blindness and vision impairment in 2020 and trends over 30 years.” Lancet Global Health 9, e144–e160. The core table of blindness by cause. Table 4 covers people over 50; the appendix holds all-ages data.
Bourne, R. et al. (2021). “Trends in prevalence of blindness and distance and near vision impairment over 30 years.” Lancet Global Health 9, e130–e143. The 43.3 million blind in 2020 and the 2050 forecast.
Limburg, H., Keunen, J. E. E., and van Rens, G. H. M. B. (2009). “Hoeveel slechtzienden en blinden zijn er in Nederland?” Tijdschrift voor Gerontologie en Geriatrie 40, 149–155. The only Dutch count by cause, split between care homes and independent living. Shows that a third of blind elderly people need only a lens or glasses.
McCormick, I. et al. (2026). “Effective cataract surgical coverage in 68 countries, 2000–2030.” Lancet Global Health, online 10 February 2026. Coverage by country, from 2.1% in Burundi to 77.7% in Qatar — the essay’s key evidence that layer 0 is a delivery problem.
Layer 0: the passage
Gain, P. et al. (2016). “Global survey of corneal transplantation and eye banking.” JAMA Ophthalmology 134, 167–173. One donor cornea for every 70 needed.
Holden, B. A. et al. (2016). “Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050.” Ophthalmology 123, 1036–1042. The forecast that half the world will be myopic by 2050.
Rafat, M. et al. (2022). “Bioengineered corneal tissue for minimally invasive vision restoration in advanced keratoconus.” Nature Biotechnology 41, 70–81. A cornea made of pig collagen restored sight to all 14 blind patients in the trial — a potential answer to the donor shortage.
Layer 1: the window
Moiseyev, G. et al. (2005). “RPE65 is the isomerohydrolase in the retinoid visual cycle.” PNAS 102, 12413–12418. What the Luxturna gene does: it makes the chromophore, one of the two fixed places of the window.
Russell, S. et al. (2017). “Efficacy and safety of voretigene neparvovec in patients with RPE65-mediated inherited retinal dystrophy.” The Lancet 390, 849–860. The pivotal trial behind the first approved gene therapy for the eye.
Sahel, J.-A. et al. (2021). “Partial recovery of visual function in a blind patient after optogenetic therapy.” Nature Medicine 27, 1223–1229. The first person to see again through a light-gated channel. Read for what he could do with goggles — and could not do without them.
Holz, F. G., Le Mer, Y., Muqit, M. M. K. et al. (2025). “The PRIMA subretinal photovoltaic implant in geographic atrophy.” New England Journal of Medicine 394, 232–242. The Prima trial: 38 patients, a mean gain of 25.5 letters — the strongest restoration result in dry AMD to date.
Macé, E. et al. (2015). “Targeting channelrhodopsin-2 to ON-bipolar cells with vitreally administered AAV restores ON and OFF visual responses in blind mice.” Molecular Therapy 23, 7–16. The experimental basis for the design rule: an exciting opsin at the right address recruits the retina’s own inhibition. Note the strychnine experiment, which removes the OFF responses.
Pearson, R. A. et al.; Santos-Ferreira, T. et al.; Singh, M. S. et al. (2016). Nature Communications 7, 13029; 13028; 13537. Three independent groups showing that most “integrated” photoreceptors after transplant were host cells carrying donor material — the sobering baseline for cell therapy.
Lin, B., Koizumi, A., Tanaka, N., Panda, S., and Masland, R. H. (2008). “Restoration of visual function in retinal degeneration mice by ectopic expression of melanopsin.” PNAS 105, 16009–16014. A slow, sensitive pigment gives light detection but not shape — the illustrative case that sensitivity trades off against image formation.
Layer 2: the edge
Heijl, A. et al. (2002). “Reduction of intraocular pressure and glaucoma progression: results from the Early Manifest Glaucoma Trial.” Archives of Ophthalmology 120, 1268–1279. The measured effect of lowering pressure: slowing, not restoring.
Lu, Y. et al. (2020). “Reprogramming to recover youthful epigenetic information and restore vision.” Nature 588, 124–129. The mouse work behind the first human regeneration trial in glaucoma.
Life Biosciences (2026). “First-in-human data from ongoing phase 1 trial evaluating ER-100 in optic neuropathies.” Press release, 8 October 2026. The first signal of restored visual field in glaucoma: 2 of 3 patients, uncontrolled. Read it as a first signal, not a result.
Layers 3 and 4: relay and reading ring
Pezaris, J. S., and Reid, R. C. (2007). “Demonstration of artificial visual percepts generated through thalamic microstimulation.” PNAS 104, 7670–7675. The case for the LGN as the place to write when the eye and optic nerve are gone.
Fernández, E. et al. (2021). “Visual percepts evoked with an intracortical 96-channel microelectrode array inserted in human occipital cortex.” Journal of Clinical Investigation 131, e151331. A blind woman sees letters through 96 electrodes in V1.
Jazayeri, M., Lindbloom-Brown, Z., and Horwitz, G. D. (2012). “Saccadic eye movements evoked by optogenetic activation of primate V1.” Nature Neuroscience 15, 1368–1370. The first sign that light on V1 is genuinely seen by a primate; the basis for an optogenetic cortical prosthesis.
Ajina, S. et al. (2015). “Human blindsight is mediated by an intact geniculo-extrastriate pathway.” eLife 4, e08935. The second reading route that survives after V1 is lost — the mechanism that field training exploits.
Cavanaugh, M. R. et al. (2017). “Vision recovery in cortical blindness.” Neurology; and Huxlin, K. R. et al. (2009). Journal of Neuroscience 29, 3981–3991. Training brings the blind field of stroke patients back into use.
Other routes to the reading ring
Norman, L. J., Hartley, T., and Thaler, L. (2024). “Brain changes after 10 weeks of click-based echolocation training in blind and sighted adults.” Cerebral Cortex 34, bhae239. Ten weeks of clicking, and V1 answers to echoes in blind and sighted adults alike. Read with Norman et al. (2021), PLoS ONE 16, e0252330, for the effects on daily life.
Amedi, A. et al. (2007). “Shape conveyed by visual-to-auditory sensory substitution activates the lateral occipital complex.” Nature Neuroscience 10, 687–689. Sound that carries shape reaches the visual shape area — substitution is not metaphor but cortical reality.
Background
Fernández, E., and Robles, J. A. (2024). Review of visual prostheses. PLoS Biology 22, e3002896. A review of every prosthesis route, including the estimate of 600–700 electrodes needed for reading and walking. See the section on channel counts.
Land, M. F., and Nilsson, D.-E. (2012). Animal Eyes, 2nd ed. Oxford University Press. How eyes arose at least 40 times independently, and what each design can and cannot do.
Het boek From Cell to Mind onderzoekt de evolutie van prikkelbare cellen tot de menselijke geest in twee strikt gescheiden lagen.
The book asks one question. How did excitable cells, nerve nets, brains and minds evolve? It wants the answer in order, in sizes, in times and in counts. It also asks a second question. Can that evolution be described within the author’s Vacuum.Net model?
The book is the fourth in a series. Plants on the Ladder, The Company of Plants and The Weaving of Life covered photosynthetic life, co-evolution and the interweaving of the kingdoms. This volume follows one line through the animals. It runs from the charged membrane of a single cell to the human mind.
The design is unusual and it is the key to reading the book. The text is split into two layers that are kept strictly apart.
Layer A is biology only. It has ten chapters, A1 to A10. Each chapter follows the same template. What it is. First appearance. Partners or groups. Sizes. Times. Counts. Order of events. Present quantities. Each chapter closes with a Pattern paragraph of at most 150 words. Then it lists what could not be verified and its own references. Layer A never mentions the theory.
Part C gathers the numbers. Seven tables collect the measured values of Layer A across all chapters. They cover first appearances, repeated origins, sizes, times, lifetimes, neuron numbers and present-day quantities. Together they hold more than a thousand rows of data.
Layer B is the confrontation. It asks one question of Layer A. Do the measured patterns correlate with what the Vacuum.Net texts state? Nothing new is added. The texts speak of the membrane, the nerve cell, synchrony, sleep, seizure, memory and consciousness. Three tests were fixed from the texts before any dossier was written. Then they were run. Section B14 gives the answer to the book’s question.
This design has a clear business logic. The data are gathered first and blind to the theory. The claims are fixed before the data are seen. The outcomes stand as they fell. In a field where theory and evidence often blur, the book separates them by rule, not by rhetoric.
2. The rules that make the book checkable
Three rules run through the whole work.
Marks for every statement
Every factual claim carries a mark. Measured means the number was read in a cited source. Computed means arithmetic on measured numbers, shown openly. Modelled means the output of a published model. Text means a statement of a Vacuum.Net text, and appears only in Layer B. Read means a pairing between a text statement and a measured pattern. A read is a hypothesis, never a fact.
The ladder
Every size and every time is placed on a logarithmic scale of base three, called its position. The units are Larson’s natural units. A day sits at position 43.50. A year at 48.87. A two-millisecond spike at 27.50. An 80-year human lifetime at 52.86. The position is a coordinate. It lets the book compare a nanosecond ion current with a geological age on one axis.
What is not added
No rule is added to the theory. No sentence says the net causes a biological pattern. Where the texts are silent, Layer B says none. Chapter B12 collects these silences. The texts make no statement on neuron numbers, brain size, energy use, gene counts or sense-organ counts. The book records the gaps and refuses to fill them.
The verification discipline is strict. Citations were checked against OpenAlex or Crossref. Items found only in secondary sources are listed under “Not verified” and are treated as non-facts. Where two sources disagree, both numbers stand. The book does not choose.
3. Layer A: the ten chapters as one story
The ten chapters form a single arc. Each stage builds on the one before it. The story moves from chemistry, to cells, to networks, to organs, to behaviour, and finally to minds.
A1. The Excitable Cell
The story starts before neurons, before animals, before multicellularity. Every cell measured carries a voltage across its membrane. The inside is negative. Plant cells hold the largest values, up to −250 mV. Bacteria come next, at −105 to −150 mV. This is not a detail of animal life. It is a property of life itself, already present in LUCA.
Excitability comes from channels in the membrane. They open and close with voltage, ligand or touch. Potassium, sodium and calcium channels occur in bacteria, archaea, protists, algae, plants and animals. The voltage-gated sodium channel family emerged “more than a billion years ago”, before nervous systems and before multicellularity.
The chapter then shows that electrical signalling works long before there are nerves. Bacillus subtilis biofilms send long-range potassium waves. Diatoms fire fast sodium-based action potentials. The Venus flytrap closes its trap on action potentials. A glass sponge conducts all-or-none impulses with no neurons at all. The chapter sets the reference values the rest of the book will scale against. The neuron rests at −70 mV, with threshold −55 mV and peak +30 mV.
A2. Before the Neuron
If every cell is excitable, what is a neuron for? This chapter collects the parts of the neuron that exist without neurons. The post-synaptic scaffold occurs in choanoflagellates, the unicellular sister group of animals. Sponges keep a nearly complete post-synaptic scaffold without synapses. Placozoans coordinate feeding with secreted peptides.
The chapter also records the open question of the field. Did neurons arise once or twice? The answer hinges on the branching position of ctenophores. In ctenophores many bilaterian neuron-specific genes are absent or not expressed in neurons. The chapter logs every statement found on division of neurons in early animals. That record will matter in Layer B.
A3. Nerve Nets
The first true nervous systems are diffuse nets. The Pattern paragraph fixes the numbers for Hydra: 0.12 nerve cells per epithelial cell, in two nerve nets that do not mix. A stem pool of 3,000–6,000 cells supplies about 900 nerve cells per day. The whole net is renewed every two to three weeks.
Molecular clocks place nerve nets at about 741 Ma. This is the first rung of the ladder with a date. The chapter also records the spread of speeds within one phylum: cnidarian conduction spans a factor of 133, from 3.0 cm per second to 4 m per second.
A4. Centralisation
Nets then gather into cords, commissures and brains. The chapter is honest about the disagreement in the literature. One account derives the bilaterian nervous system from the fusion of two centres, an apical and a blastoporal one. A second account, based on gene expression across five clades, finds no conserved dorsoventral regionalisation. Origins of nerve cords are recorded as one fused system and as convergence “many times”.
The counts climb fast. Ciona has 177 CNS neurons. C. elegans has 302. The fly larva brain has 3,016. The adult fly brain has 139,255. Fossil brains appear in Cambrian Stage 3, around 520 Ma. Molecular dates sit earlier: bilaterian phyla before 638.8 Ma, the vertebrate split at about 560 Ma.
A5. Speed and Size
Centralisation raises a cost problem. A bigger body needs faster signals. The chapter records two solutions: giant axons and myelin. Myelin arose in five groups across three bilaterian superclades. Its speed gain is at least tenfold. Myelinated copepods react in 1.5–1.6 ms; non-myelinated ones need 3.1–6.6 ms.
Speed has a price. One action potential costs 3.84 × 10⁸ ATP. The adult human brain is 2% of body mass but takes 20% of resting energy. In childhood its share reaches 65–66%. Neuron numbers run from about two million in a gecko to 257 billion in an elephant. The chapter closes with the division status of vertebrate neurons. Human cortical neurons date from around birth. This record feeds directly into Layer B.
A6. Senses
Brains are only as useful as the information they receive. This chapter covers the receptor molecules and organs of light, smell, taste, touch, hearing and electric fields. Light is treated in most detail, and for a reason that will surface in Layer B.
The numbers are striking. Eye or photoreceptor origins are counted at least 40 times, up to 65 or more. A quarter of those lineages reached resolving eyes. Mammals carry about 400 to 2,000 olfactory receptor genes. Electroreception arose at least four times and was lost at least five times. The first dated compound eyes, at about 515 Ma, already hold 3,000 and 16,000 lenses.
A7. Rhythms and Sleep
With senses and speed in place, the book turns to what brains do over time. Networks of neurons oscillate in well-defined frequency bands. One research group publishes ten classes, from 15–40 s to 200–600 Hz. The mean frequencies form a linear progression on a logarithmic scale. The Pattern paragraph notes that these frequencies vary only 30–100% across mammals, while brain size varies hundreds- to thousands-fold.
Sleep appears here too. Sleep-like states occur from Hydra to mammals. Daily sleep spans 0.69 h to 20 h. Infraslow sleep periods run 47–51 s in endotherms and 84–134.5 s in lizards. The chapter holds the data for one of the book’s three tests.
A8. Learning and Memory
The chapter records three forms of learning: habituation, sensitisation and association. These forms are not reserved for animals with complex brains. A slime mould, Physarum, habituates. A ciliate, Stentor, habituates. A plant, Mimosa, habituates. Box jellyfish with tiny nerve rings learn by association.
One pattern recurs in four animal groups. Spaced training gives memory of days to weeks; massed training does not. The phase that lasts beyond hours needs protein synthesis. This holds in rat, Aplysia, worm and bee. Two forgetting curves show breaks, at 24 hours and at 3–6 years. These are the data that Layer B will set beside the theory’s statement on memory.
A9. Minds in Many Bodies
Complex brains are not a human monopoly. They are also not a shared inheritance. The chapter records complex brains in seven groups that do not share a complex-brained ancestor: insects, octopodid molluscs, teleosts, corvid and psittacid birds, cetaceans, elephants and primates. Each pairs with a different multimodal centre: mushroom bodies, the vertical lobe, the pallium, the cortex.
The measured capacities include counting, self-control, planning, tool construction, play and mirror self-recognition. The chapter closes with dated statements on animal consciousness, such as the New York Declaration of 2024. It reports them as positions held by named authors, not as measurements.
A10. The Human Brain and Beyond
The final chapter gathers the human numbers. The brain holds 86.1 ± 8.1 billion neurons and is described as “an isometrically scaled-up primate brain”. Cranial capacity runs from about 450 cc in Australopithecus to 1,297–1,460 cc in modern humans. The rate of increase steps up at 2.10 Ma. Tool and image dates run from 3.3 Ma to 51.2 ka. Synaptogenesis peaks follow one another from three months to 28.7 years.
The chapter then steps beyond the single brain. Social insect colonies, ant biomass, and the global count of internet users enter the tables. The human mind is placed in its widest context.
4. How the chapters connect
Read in sequence, the chapters form a chain of dependencies.
Excitability (A1) is the foundation. It exists in every cell and needs no nervous system. The parts of the neuron (A2) then assemble from components that predate animals. Nerve nets (A3) are the first system to exploit those parts at tissue scale. Centralisation (A4) is a solution to a coordination problem of larger bodies. Speed (A5) is the price and the enabler of centralisation. Senses (A6) are the reason centralisation pays off: a fast brain with good input can act on the world.
Rhythms (A7) and memory (A8) are what the fast brain does with its input. Rhythms segment time; memory stores what matters. Complex brains (A9) show that these capacities were invented again and again, in bodies as different as a crow and an octopus. The human brain (A10) is the last rung: not a new kind of organ, but a scaled-up primate brain, wrapped in culture and tools.
The recurring theme is continuity. Nothing in the chain is a true novelty. The neuron is an excitable cell with a specialised edge. The brain is a net with gathered centres. The mind, in the book’s reading, is the third rung built on the second.
5. Layer B: the theory set beside the data
Layer B does not reinterpret the biology. It sets the Vacuum.Net texts next to Layer A and asks where the forms match. The texts name four rungs: closure, life, representation, consciousness. They say the membrane carries the largest potential step of the cell. They say the nerve cell adds nothing new. They say each layer is a factor three. They say memory reaches two rings directly. They say a brain in which all cells fire together is a brain in seizure.
The pairings produce three reads. First, the neuron adds no new part; it adds speed and reach to an edge every cell already has. Second, membrane, nerve net and brain follow the rungs of life and representation. Third, the carriers are fast and the forms slow at each scale. In Hydra, the nerve cell itself is a carrier and the net is the form. In the human cortex, the neuron is the slow form and its spikes are the carriers.
The three tests
Three tests were fixed before any dossier was written. This is the book’s strongest methodological move. The outcomes stand as they fell.
M1: the nerve cell does not divide. Required: no documented completed division of a differentiated neuron. Found: Rb-deficient mouse cortical neurons in culture divide, with daughters that keep their neuronal marker. Fail. The direction survives in most of the data: with cell-cycle brakes intact, the neuron does not divide. But the text made no exception, and the test does not grant one afterwards.
M2: the rhythm step is three. Required: a mean step of 2.85–3.15 between oscillation classes. Found: 2.73 over ten published classes. The value e = 2.718 lies 0.5% from it. Three lies 9.8% above it. Fail. The constant step and the conservation across brain sizes hold; the specific number does not.
M3: the living form is not synchronous. Required: at least 80% of awake spike-count means below 0.1, and higher synchrony in every sleep and seizure comparison. Found: 69%. One sleep comparison shows less synchrony. Seizures go in both directions. Fail. Again the direction holds in much of the data; the strict form does not.
B14: the answer to the question
The book’s own summary is candid. Layer A gives the history: a negative inside in every cell measured; channels with a fourfold pore in bacteria; synaptic parts in the unicellular sister of animals; nerve nets at 741 Ma; brains at about 520 Ma; myelin in five groups; eyes at least 40 times; complex brains in seven groups; ten rhythm classes held across a thousandfold range of brain size; two phases of memory in four animal groups.
Layer B gives three reads and three failed tests. Across the four books of the series, the theory has been set against biological data twelve times with the outcome fixed first. One passed. Six failed outright. One failed on its own rule. Four could not be completed.
The conclusion follows the book’s rules exactly. In these terms the mind is the third rung built on the second. Its parts are older than the neuron. Its rhythm classes keep nearly one spacing. Its cells are carriers or forms depending on the scale. The texts give the form of that description. Where they gave a number, three times, the number failed.
6. Assessment
The book’s value rests on two achievements that are easier to state than to practise.
First, the reference layer is disciplined and dense. Fixed chapter templates, explicit marks, checked citations, recorded disagreements and “Not verified” lists make the claims auditable. The Pattern paragraphs give each chapter a usable summary, and Part C turns the whole into a numerical corpus.
Second, the theory layer is honest to an unusual degree. Tests are fixed before the data are seen. Failures are printed as failures, with the surviving directions clearly labelled as reads made after the test and not as passes. Chapter B12 lists the places where nothing pairs, rather than stretching the pairings. Part D lists what was left out, by rule.
The reader should note what the book does not claim. It does not claim the Vacuum.Net model is confirmed. On the numbers, three of its fixed claims failed. The book claims only this: the model supplies a vocabulary of rungs, carriers and forms that fits the shape of the data at many points, and that vocabulary is now marked, tested and bounded.
That is a rare position in a debate often driven by assertion. The book treats evolution and theory with the same instrument: a measured number, a fixed claim, and an outcome that stands as it fell.
7. Annotated references
The book and its texts
Konstapel, J. (2026). From Cell to Mind: How Excitable Cells, Nerve Nets, Brains and Minds Evolved. A Reference Work in Two Layers. Constable Research, Leiden, 14 October 2026. The source work discussed in this essay. Its Layer A holds the biology; Layer B holds the theory tests; B14 states the answer.
Konstapel, J. (2026). Life and Consciousness from One Rule: Closure in the Net. Constable Research, Leiden, 29 September 2026. Source of the four rungs (closure, life, representation, consciousness), the membrane as the place of the largest potential step, and the coherence values R = 1 and R = 0. Quoted in B2.
Konstapel, J. (2026). One Window: Light-Gated Channels, the 2026 Nobel Prize in Medicine, and the Closure Rule of the Vacuum.Net. Constable Research, Leiden, 10 October 2026. Source of the statement that the nerve cell adds nothing new, and of the synchrony reading used in test M3.
Konstapel, J. (2026). Three Chains, One Standstill: Dementia and Parkinson’s Disease on the Ladder of the Vacuum.Net. Constable Research, Leiden, 10 October 2026. Source of tests M1 and M3, and, with Life and Consciousness, of M2. States the nerve cell as a ring that does not divide and each layer as a factor three.
Konstapel, J. (2026). The Balanced Three, the Balanced Tree and the Vacuum.Net Theory. Constable Research, Leiden, 29 August 2026. Source of the memory statement: reach two, chaining, and the impossibility of forgetting. Set beside the A8 data in B10.
Konstapel, J. (2026). The Climate on the Ladder of Three. Constable Research, Leiden, 3 October 2026. Defines depth as layer and the position coordinate used for every size and time in the book.
Konstapel, J. (2026). Plants on the Ladder; The Company of Plants; The Weaving of Life. Constable Research, Leiden, 12–13 October 2026. The three earlier books of the series. B13 counts their nine earlier tests together with the three new ones.
Key primary sources used in Layer A
Burkhardt, P., et al. (2023). Syncytial nerve net in a ctenophore adds insights on the evolution of nervous systems. Science 380:293–297. DOI 10.1126/science.ade5645. Central to the A2 question of one or two origins of neurons. The ctenophore nerve net lacks classical synaptic machinery.
Buzsáki, G., & Draguhn, A. (2004). Neuronal oscillations in cortical networks. Science 304(5679):1926–1929. DOI 10.1126/science.1099745. With Penttonen & Buzsáki (2003), the source of the ten oscillation classes tested in M2.
Azevedo, F. A. C., et al. (2009). Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain. J. Comp. Neurol. 513:532–541. DOI 10.1002/cne.21974. Source of the 86 billion neuron count that anchors A10.
Herculano-Houzel, S., et al. (2014). The elephant brain in numbers. Frontiers in Neuroanatomy 8:46. DOI 10.3389/fnana.2014.00046. Source of the 257-billion-neuron elephant brain, the upper end of the neuron counts in A5 and A9.
Attwell, D., & Laughlin, S. B. (2001). An energy budget for signaling in the grey matter of the brain. J. Cereb. Blood Flow Metab. 21(10):1133–1145. DOI 10.1097/00004647-200110000-00001. Foundation of the spike-cost figures in A5, including the 3.84 × 10⁸ ATP per action potential.
Kuzawa, C. W., et al. (2014). Metabolic costs and evolutionary implications of human brain development. PNAS 111(36):13010–13015. DOI 10.1073/pnas.1323099111. Source of the childhood brain-energy share of 65–66% of resting metabolism in A5 and A10.
Bielecki, J., et al. (2023). Associative learning in the box jellyfish Tripedalia cystophora. Current Biology 33(19):4150–4159.e5. DOI 10.1016/j.cub.2023.08.056. Shows associative learning in an animal with a few thousand nerve cells. A key A8 data point.
Bliss, T. V. P., & Lømo, T. (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit. J. Physiol. 232(2):331–356. DOI 10.1113/jphysiol.1973.sp010273. The classical LTP paper, cited in A8 as the cellular counterpart of long-term memory.
Background books
Godfrey-Smith, P. Other Minds: The Octopus, the Sea, and the Deep Origins of Consciousness. Framed the cephalopod material of A9.
Buzsáki, G. Rhythms of the Brain. Framed the oscillation classes of A7.
Herculano-Houzel, S. The Human Advantage. Framed the neuron-count scaling of A5 and A10.
Striedter, G. F. Principles of Brain Evolution. Framed the comparative structure of A4 and A9.
Feinberg, T. E., & Mallatt, J. The Ancient Origins of Consciousness. Framed the consciousness positions recorded in A9.
The birth of a shared word is modeled as a Hopf bifurcation; the theory’s coefficients place this birth in the subcritical regime, where lived practice and the fixed word coexist as two stable states separated by an unstable boundary, with history rather than logic deciding which state a group occupies.
Six further correspondences are drawn: readiness-to-hand as “closure is zero force,” source–target fit as natural selection, agentless grammar as an external measure, the loom as the true image of computing, placing people in a room as the reading of addresses, and group life as having two kinds of coherence.
“Metaphor” and “translation” share the same root meaning: carried across. This essay applies that idea to the Vacuum.Net theory — a model of reality as a single net made of one strand, closed windings, knots and meshes.
The central claim is a match between layers. A metaphor as the body lives it belongs to the theory’s tension layer, where change happens. A metaphor as a community fixes it in a shared word belongs to the address layer, which is frozen topology. The birth of a shared word behaves like a Hopf bifurcation — the moment a system at rest starts to move in a repeating cycle. The theory’s own coefficients place this birth in the subcritical regime, in which the lived practice and the fixed word coexist as two stable states, separated by an unstable boundary. History, not logic, decides which state a group occupies.
Six further correspondences follow: readiness-to-hand as “closure is zero force”; the fit between source and target as natural selection by fit; agentless grammar as an external measure; the loom as the true image of computing; placing people in a room as the reading of addresses; and group life as having two kinds of coherence rather than one. The essay states plainly which claims are proved, which are correspondences, and which remain open.
1. One strand, carried across
The Greek meta-pherein and the Latin trans-latus describe the same action: carrying something across. A metaphor takes structure from a domain the body knows — weight, distance, balance — and uses it to think about a domain one cannot touch, such as time, markets or organisations. Linguists call these the source and the target.
The Vacuum.Net theory describes reality as one net built on five axioms:
N1 — there is one strand.
N2 — a state persists only if it closes on itself.
N3 — the net keeps the orientation in which it was laid.
N4 — properties depend on the local state.
N5 — the same pattern returns at every scale.
From these follow four forms: a strand (the net itself), a winding (a strand that returns to itself), a knot (windings that hold each other in three directions) and a mesh (the opening between strands).
Axiom N5 already contains the structure of metaphor. If the same winding appears at every scale, then a pattern found at one depth can legitimately be read at another. Carrying across is then not a figure of speech but identity over scale. This essay follows that thread through the body, the word, grammar, the machine, and the room where people meet.
2. Two layers: the lived metaphor and the fixed word
A hand discovers the use of a stone before any word for it exists. The grip fits; the throw works. Something is treated as something else, entirely in the body. Call this layer metaphor-in-use.
Later, a group settles on a word for that use. The word is shared, taught and policed; using it correctly becomes a condition of belonging. Call this layer the espoused metaphor. The pair follows Argyris and Schön’s classic distinction between theory-in-use and espoused theory. (“Espouse” comes from Latin spondere, to pledge — to espouse a theory is to wed it in public.)
The two layers behave differently:
The lived metaphor is present tense and resists being named.
The fixed word arrives late; “a thought” is grammatically a finished thing, like a past participle.
Articulation creates order, and order is not a property of the present.
The Vacuum.Net theory has the same split, called the two-layer law. The address layer holds frozen topology: which windings exist and how they are linked. The tension layer is the only layer in which anything happens. Exhaustive computation supports the frozen character of the address layer: every admissible local rule on a ring of five sites (1,594,323 sign-symmetric rules and 144 conserving flux-form rules) turned out to be either the identity or a shift. In the address layer there is only passage; change lives in the tension layer.
The correspondence is direct: metaphor-in-use ↔ tension layer; espoused metaphor ↔ address layer. The gap that organisational science observes between what people do and what they profess is the gap between the two layers. Nobody has to lie for it to open.
3. A word is the birth of a winding
When does lived use become a word? When the use repeats and is shared. A one-time grip leaves no trace; a grip that returns, and that others recognise as the same grip, becomes a thing with a name. In net terms: a thing is a turn that keeps coming back.
Dynamical systems theory has a precise description of the birth of a repeating cycle: the Hopf bifurcation. Its normal form, in polar coordinates, is
ṙ = μr − ar³, θ̇ = ω
where r is the amplitude, θ the phase, μ the control parameter and a > 0.
Step 1. For μ < 0 the only equilibrium is r = 0: rest, or address 0 in the net.
Step 2. For μ > 0 the rest point loses stability and a closed orbit appears with radius √(μ/a). A phase that returns after one full turn is, by definition, a winding.
Step 3. The winding number around the rest point jumps from undefined to 1. It cannot change gradually. So μ belongs to the tension layer and the winding number to the address layer; the Hopf point is where the two layers meet.
Exactly at μ = 0, the equation becomes ṙ = −ar³, whose solution is r(t) = r₀/√(1 + 2ar₀²t). With r₀ = 1, a = 1 and t = 100, the amplitude falls to 1/√201 ≈ 0.071. A system sitting precisely on the threshold does not gain a living rhythm; it fades slowly. This matters for the attractive idea of a “permanent Hopf bifurcation”: sitting exactly on the point gives slow decay, not life.
The subcritical form. The net’s coefficients point elsewhere. Its free-energy functional F(A) = αA² + βA⁴ + γA⁶, combined with the tension relation γ = 2π, forces the quartic coefficient negative — a cancellation of (2π − 1)/2π = 5.2832/6.2832 = 84.1%. Under gradient dynamics, a negative quartic term in F yields a positive cubic term in the amplitude equation:
ṙ = μr + ℓr³ − cr⁵, ℓ > 0, c > 0
This is the subcritical Hopf bifurcation, sitting near the Bautin point where ℓ changes sign. Its equilibria are r = 0 plus the solutions of r² = (ℓ ± √(ℓ² + 4μc))/2c, real for μ ≥ −ℓ²/4c.
Worked example. Take ℓ = 1, c = 1, μ = −0.1 (inside the window −0.25 < μ < 0):
Rest and a finite winding are both stable at the same μ, separated by an unstable cycle. This is the dynamical form of the gap between lived use and fixed word: a group can sit in either state under identical conditions, and which one it occupies depends on its history — hysteresis. Note that the window has width ℓ²/4c. It is a range, not a point; the net need not hit an exact threshold to stay in it.
4. Selection by fit
Not every source fits every target. A metaphor for time built from the taste of lemons carries nothing, because taste lacks what time requires: direction, sequence, distance. The body limits which carryings are available, and only a source whose structure fits the target survives. This is selection by fit: what returns in phase stays; what does not fit does not close. No mechanism is needed — nothing pushes the right metaphor into place; the ill-fitting ones simply fail to close.
The limit does not fix a single answer. Speakers of Aymara place the future behind and the past in front, for a bodily reason: one sees what lies ahead, and only the past is known. English uses the opposite map. Both are coherent; both close. In the net’s balanced-ternary register this is the mirror pair: +1 and −1 sit symmetrically around 0, each a full closure with the opposite orientation. Axiom N3 keeps orientation as memory — a culture’s choice of map is such an orientation, laid once and kept.
5. Readiness-to-hand is zero force
Heidegger observed that a tool in use does not appear as an object: a hammer while hammering is invisible, and shows up as a separate thing only when the use breaks down.
The net has a proved counterpart. Its phase register uses three values (−1, 0, +1); a local window holds three consecutive crossings and is closed when its sum is 0 modulo 3. Two kinds of closure exist: homogeneous (all three values equal, e.g. (1,1,1)) and heterogeneous (all three different, e.g. (−1,0,1)). The closure test is the discrete Laplacian tᵢ₋₁ − 2tᵢ + tᵢ₊₁ (mod 3) — the same operator that appears in linearised coupling between neighbouring oscillators. Restricting the phases to the three trit values, the coupling force on the centre of the window is zero exactly when the window is closed. Closure is zero force; an open window carries residual tension.
The correspondence with Heidegger is then exact in form:
A tool in use is a closed window: no force, no visibility.
A broken tool is an open window: residual tension makes it appear.
The object is the open window.
6. Condition and the unconditioned
The Latin condicio comes from con-dicere, to speak together. A condition is the residue of a conversation; it exists because it circulates between parties. In net terms, a condition is a closure between people.
Grammar, however, requires a subject, and “It rains” supplies one from nowhere — nothing in the sky does the raining. The same habit hides human choices: “The market moved.” “It had to happen.” “The system won’t let us.” The sentence has no actor, so an objection has no addressee. The same applies to founding speech that fixes a state of affairs with no second party present: nothing was spoken together, and the cause stands outside every conversation that follows.
This is what the Vacuum.Net work calls the external measure. An inventory of 58 silent assumptions in modern science finds that 32 trace back to it; agentless grammar (“one subject per event”) is a second terminal assumption in the same inventory. Rosen’s analysis of living systems offers the alternative: when the chain of production closes on itself, an external driver is no longer needed — the system is its own fixed point. Bayesian reasoning is the formal version of the same stance: every cause stays provisional, today’s estimate being tomorrow’s prior, with no first cause at which the process stops.
A practical consequence follows for anyone who works with groups. The question “why” asks for a cause and lands as a demand for defence. Questions about what happened, what changed and what would need to differ keep the condition open: they stay inside the closure instead of reaching for a driver outside it.
7. “Is” and “is like”
“The market is difficult” asserts an identity between two domains. “The market is like a difficult negotiation” restores the comparison. The second sentence reopens what the first closed.
Vaihinger built a philosophy on this: causation, the unified self and scientific laws work as if — as useful fictions — and the danger begins when the fiction is taken as literal and unconditioned. The Vacuum.Net foundational paper makes the same discipline operational: every claim carries one of five statuses (assumed, conditionally proved, corresponded, predicted, open). “Corresponded” is the status of “is like”; “conditionally proved” is the status of “is”, under named hypotheses. Keeping the two apart is Vaihinger’s warning made routine.
8. The loom and the net
Text, textile, texture and context all come from Latin texere, to weave. A text is a woven thing; making sense and making cloth share a root.
Computing had a weaving origin too. Jacquard’s loom of 1804 ran on punched cards; Babbage adapted the cards for his Analytical Engine; Lovelace wrote in 1843 that the Engine weaves algebraic patterns as the loom weaves flowers and leaves. Later the calculator became the governing image and the loom was buried.
The net is a loom in the strict sense. Its first theorem identifies the elementary mark with a crossing, and the smallest stable knot — the trefoil — holds by its crossings alone and needs exactly three (any diagram with fewer crossings can be undone). Strength in a fishing net does not come from any single thread but from how the threads cross; the same holds for a text, a group and the vacuum.
The difference between calculator and loom is the difference between an external measure and a relation. A calculator returns one correct number against a fixed scale. A loom holds a pattern made of crossings — and in a loom, the weaver is also a thread.
9. Placing people: reading addresses
One method in group work places people silently in a room, each standing for a party in a real case. They know nothing of the case and receive only a stance: left or right, near or far, facing or turned away. Reports are consistent: people placed this way soon speak the situation, often using the same words as the real parties — words they never heard.
The net needs no mechanism to read this. A position in the room is an address; the pull a person feels at that position is the tension at that address. Whoever stands on the same address meets the same tension. This is correspondence and identity over scale, not transmission. The sequence of the method matches the two layers: first the stance, then the feeling, then the words. Tension comes before address; the lived layer before the fixed one. The same holds for objects placed in a circle: a group that has just placed its own objects cannot easily say “it just happened” about the pattern — it saw its own hands make it.
10. Two closures in the room
Conventional coherence is measured by the order parameter R: represent each participant’s phase as an arrow on a circle and average the arrows. Arrows pointing the same way give R near 1; arrows spread out give R near 0, which is usually read as disorder.
The trit register shows why that reading fails. Map the three trit values to the cube roots of unity: 1, e^(2πi/3), e^(4πi/3).
Homogeneous closure: three equal arrows, R = 1.
Heterogeneous closure: three arrows at 120°, whose sum is 1 + (−½ + i√3/2) + (−½ − i√3/2) = 0. So R = 0.
Both are exact closures — one is alignment, the other structured difference. Their stability differs. For a ring of N oscillators with coupling K and phase step Δ, perturbation modes grow at λₘ = −4K cos(Δ) sin²(πm/N). For homogeneous closure Δ = 0, so cos Δ = 1 and the state is stable for positive coupling. For heterogeneous closure Δ = 2π/3, so cos Δ = −½ and λₘ = 2K sin²(πm/N): the state is stable only for K < 0. Heterogeneous closure needs the differences to be maintained, not pulled together.
The meaning for groups is plain. A team in which everyone thinks alike shows high coherence. A team in which three people hold three different, mutually necessary roles shows R = 0 — and is no less closed. Measured by alignment alone, the second team looks broken. It is not. The paradoxes of group life — belonging versus individuality, trust versus disclosure, closeness versus distance — do not resolve, and do not need to. They are heterogeneous closures. The work is to keep the window closed while the parts stay different; forcing them into alignment would destroy the closure.
Older contemplative traditions name the same pair: Sufi metaphysics distinguishes jam’ (gathering into unity) from farq (separation into multiplicity), described as simultaneous rather than sequential. In the trit register, jam’ is homogeneous closure and farq is heterogeneous closure, and the subcritical window of section 3 is the dynamical form in which both can hold at once.
11. Three closures and the edge
The net’s ladder runs strand → winding → knot, a knot being closed in three directions. Dynamical systems have a parallel ladder:
a rest point becomes a cycle (the first Hopf bifurcation);
the cycle becomes a 2-torus (the Neimark–Sacker bifurcation);
the 2-torus becomes a 3-torus.
A 3-torus is motion closed in three independent directions. Newhouse, Ruelle and Takens proved in 1978 that quasi-periodic flow on a torus of dimension three or more is fragile: arbitrarily small perturbations can turn it into a strange attractor. Read against the net, three closures are the last stable rung before the edge of chaos — and the knot sits on that edge, at the level the net identifies with mass. This fits the trefoil (three crossings, the minimum for a knot) and the trit register (three values, the minimum for a heterogeneous closure). The number three returns at each depth.
12. Two Hopfs
The name Hopf belongs to two mathematicians. Eberhard Hopf described the bifurcation in 1942; Heinz Hopf described the fibration in 1931. Their work is unrelated in origin, but in the net it meets. The net’s state is a rotor represented by a unit quaternion; unit quaternions form the 3-sphere S³, and the Hopf fibration maps S³ onto the 2-sphere S² with a circle S¹ above each point. Any two such circles are linked exactly once — the Hopf invariant, which the net uses as its conserved topological charge Q ∈ π₃(S²) = ℤ.
Read together: the 2-sphere carries direction and the circle above it carries phase. The cycle born in Eberhard Hopf’s bifurcation is a fibre of Heinz Hopf’s fibration. Every winding is linked to every other winding exactly once — every lived use is threaded through every other one. No word stands alone.
13. Language models: address without tension
A large language model learns the conditional probability of the next token given all previous tokens, over a corpus that others have already argued over. Its words arrive pre-negotiated. In the terms of this essay, the corpus is a frozen address layer: the model reproduces the espoused metaphor with great fluency, but has no tension layer of its own — no hand that found a fit, no stake in being wrong, no group that had to reach agreement.
Two consequences follow for people. The first is practice: fluent, correct language becomes cheap, and the capacity to feel a new fit before it has a name is exercised less; capacities that are not exercised weaken. The second is visibility: a weakening of the lived layer would not appear on any measure, because formal fluency is what institutions are built to count and the lived layer does not announce itself.
Roles that work with people in a room become the place where this layer is kept in practice. The concrete work is simple: let people write before they speak; use their own words back to them; bring an object instead of a question; leave silence after a concrete question. The first answer after a pause usually comes from the lived layer.
14. Status of the claims
Proved or computed in the text
the decay r(t) = r₀/√(1 + 2ar₀²t) at the Hopf point;
the subcritical window −ℓ²/4c < μ < 0, with the worked example;
the cancellation (2π − 1)/2π = 84.1%;
the link between a negative quartic term and a subcritical bifurcation;
R = 0 for heterogeneous closure;
the stability condition K < 0 for heterogeneous closure;
closure as zero force within the trit model;
passage-only dynamics in the address layer (on rings of five, nearest-neighbour rules).
Correspondences
metaphor-in-use and espoused metaphor as the tension and address layers;
the birth of a shared word as a Hopf bifurcation;
readiness-to-hand as zero force;
the fit of source to target as selection by fit;
the Aymara map as the mirror pair;
agentless grammar as the external measure;
placing people as reading addresses;
jam’ and farq as homogeneous and heterogeneous closure;
three Hopf steps as the knot;
the Hopf cycle as a fibre of the Hopf fibration.
Open
The subcritical regime rests on γ = 2π, which is a correspondence in the theory, not a derivation; a finite lattice computation showed it does not follow from 2π-periodicity alone.
The coefficient ℓ has not yet been derived from the trit rule directly; that derivation would settle whether the window exists without γ = 2π.
Hysteresis in language adoption has not been measured: a group adopting a new word should show a different threshold on the way in than on the way out.
The R = 0 closure as a marker of a well-held group has not been measured; the triplet (R, C_hom, C_het) makes it measurable.
Annotated References
Argyris, C. & Schön, D. A. (1974). Theory in Practice: Increasing Professional Effectiveness. Jossey-Bass, San Francisco. Source of the distinction between theory-in-use and espoused theory, on which section 2 builds its pair of metaphor layers. Read the first three chapters: they define both terms and show how the gap stays hidden.
Argyris, C. (1990). Overcoming Organizational Defenses. Allyn and Bacon, Boston. Shows how organisations defend the gap between what they profess and what they do — the organisational face of the two-layer law. The chapters on defensive routines are the core; read them with section 6 in mind.
Bateson, G. (1972). Steps to an Ecology of Mind. Chandler, San Francisco. Contains the double bind and the definition of information as “a difference that makes a difference”; both bear on agentless grammar. Start with “Toward a Theory of Schizophrenia” and “Form, Substance and Difference”.
Bautin, N. N. (1949). The Behaviour of Dynamical Systems near the Boundaries of the Stability Region. Gostekhizdat, Moscow (in Russian). Origin of the analysis of the degenerate Hopf point where the cubic coefficient changes sign; section 3 places the net near this point. For most readers Kuznetsov (2004) gives the same result in accessible form.
Geary, J. (2011). I Is an Other: The Secret Life of Metaphor and How It Shapes the Way We See the World. Harper, New York. A broad survey of metaphor across thought, money, science and politics; estimates about six metaphors a minute in ordinary speech. Read the chapters on etymology and on the body — they supply the material for sections 1 and 2.
Heidegger, M. (1927). Sein und Zeit. Niemeyer, Halle. English: Being and Time, transl. Macquarrie & Robinson, Harper & Row, 1962. Introduces readiness-to-hand: a tool in use does not appear as an object; section 5 matches this to zero force. Paragraphs 15 and 16 are enough — they contain the hammer example.
Hong, H. & Strogatz, S. H. (2011). “Kuramoto Model of Coupled Oscillators with Positive and Negative Coupling Parameters.” Physical Review Letters 106, 054102. Shows that negative coupling stabilises non-aligned states — the established mathematics behind the K < 0 result in section 10. Short and self-contained; read the whole paper.
Hopf, E. (1942). “Abzweigung einer periodischen Lösung von einer stationären Lösung eines Differentialsystems.” Berichte der Mathematisch-Physischen Klasse der Sächsischen Akademie der Wissenschaften zu Leipzig 94, 1–22. The original proof of the bifurcation that bears his name; section 3 rests on it. An English translation with commentary appears in Marsden & McCracken (1976).
Hopf, H. (1931). “Über die Abbildungen der dreidimensionalen Sphäre auf die Kugelfläche.” Mathematische Annalen 104, 637–665. Introduces the Hopf fibration and the Hopf invariant; section 12 uses both. Demanding; a modern algebraic-topology textbook gives the same content in current notation.
Konstapel, J. (2025). “Het Hopf Principe: Een Principe op het Snijvlak van Wetenschap en Mystiek.” constable.blog, 2 September 2025. Proposes that reality sits permanently on the Hopf threshold and links this to contemplative traditions; section 3 sharpens that claim. Read the sections on walking as near-falling and on jam’ and farq.
Konstapel, J. (2026a). “Closure Is Zero Force: The Oscillator Framework Read as the Phase Register of the Vacuum.Net Theory.” Constable Research, Leiden. Proves within the trit model that closure is zero coupling force and introduces homogeneous and heterogeneous closure; sections 5 and 10 rest on it. The proofs are short; read the stability section with a pen.
Konstapel, J. (2026b). “The Vacuum.Net Theory: Foundational Paper.” Constable Research, Leiden. States the axioms N1–N5, the theorems, the five epistemic statuses and the open computations; every net term in this essay is defined there. Read the axioms and the status section first; the theorems can follow later.
Konstapel, J. (2026c). “The Assumptions Nobody Declares.” constable.blog, 26 July 2026. An inventory of 58 silent assumptions in modern science, five of them terminal; section 6 uses the external measure and the one-subject-per-event assumption. Read the five terminal assumptions; the rest serves as reference.
Kuznetsov, Y. A. (2004). Elements of Applied Bifurcation Theory, 3rd ed. Springer, New York. The standard reference for Hopf and Bautin normal forms and their coefficients; all formulas in section 3 can be checked here. Chapter 3 for Hopf, chapter 8 for the Bautin point.
Lakoff, G. & Johnson, M. (1980). Metaphors We Live By. University of Chicago Press, Chicago. Shows that abstract thought borrows its structure from bodily experience; “argument as war” and “time as space” come from here. Short; read it whole. It is the basis for sections 1 and 4.
Lovelace, A. A. (1843). “Notes by the Translator.” In: L. F. Menabrea, “Sketch of the Analytical Engine Invented by Charles Babbage.” Scientific Memoirs 3, 666–731. Contains the comparison of the Analytical Engine with the Jacquard loom — computing’s first and buried founding image. Note A contains the loom passage; Note G contains the first program.
Marsden, J. E. & McCracken, M. (1976). The Hopf Bifurcation and Its Applications. Springer, New York. Gives the full theorem, an English translation of Hopf’s 1942 paper and many applications. Section 1 and the translation are enough for this essay.
Newhouse, S., Ruelle, D. & Takens, F. (1978). “Occurrence of Strange Axiom A Attractors near Quasi Periodic Flows on Tᵐ, m ≥ 3.” Communications in Mathematical Physics 64, 35–40. Proves that motion closed in three independent directions is fragile; section 11 places the knot on that edge. Six pages; the theorem statement and the introduction carry the point.
Núñez, R. E. & Sweetser, E. (2006). “With the Future Behind Them.” Cognitive Science 30, 401–450. Documents the Aymara construal of time, with the future behind; section 4 reads it as the mirror pair. Read the gesture data: they show that the map lives in the body before it lives in words.
Reddy, M. J. (1979). “The Conduit Metaphor.” In: A. Ortony (ed.), Metaphor and Thought, 284–324. Cambridge University Press, Cambridge. Shows that English talk about communication itself runs on a metaphor (ideas as objects, words as containers): there is no metaphor-free place from which to describe metaphor. Read the first half and the appendix of examples.
Rosen, R. (1991). Life Itself. Columbia University Press, New York. Argues that living systems are closed to efficient causation, making an external driver unnecessary; section 6 uses this. Chapters 10 and 11 hold the argument; the earlier chapters prepare the notation.
Smith, K. K. & Berg, D. N. (1987). Paradoxes of Group Life. Jossey-Bass, San Francisco. Shows that the tensions of group life are structural and do not resolve; section 10 reads them as heterogeneous closures. Read the chapters on belonging and engaging.
*Vaihinger, H. (1911). Die Philosophie des Als Ob. Reuther & Reichard, Berlin. English: The Philosophy of ‘As If’, transl. C. K. Ogden, 1924.* Treats causation, the self and scientific laws as useful fictions, warning that the danger begins when the fiction is taken as literal and unconditioned; section 7 uses this. The introduction and Part I are sufficient.
Watzlawick, P., Beavin Bavelas, J. & Jackson, D. D. (1967). Pragmatics of Human Communication. Norton, New York. States the axiom that one cannot not communicate: behaviour carries message before words do, matching the order tension before address. Chapter 2 holds the axioms.
This blog is written in english ans is related to a podcast on the Agora.
Short Dutch summary
“The Weaving of Life” betoogt dat de drie domeinen van het leven (Bacteria, Archaea en Eukaryoten) diep met elkaar verweven zijn, niet alleen via afstamming maar vooral via horizontale genoverdracht en endosymbiose. Het boek laat zien dat samenwerking en symbiose op alle niveaus terugkomen: van genen die van soort naar soort springen tot cellen die andere cellen opnemen en meercelligheid die meer dan 25 keer onafhankelijk ontstond. Uiteindelijk vormt het leven één groot netwerk waarin elke nieuwe verbinding voortbouwt op eerdere samenwerkingen, met de mens als recent knooppunt dat bijna alle zoogdierbiomassa beheert.
One ancestor, three domains
The book opens with the deepest question. All cellular life falls into three groups: Bacteria, Archaea and eukaryotes. How are they connected?
The grouping began in 1977, with ribosomal RNA. That comparison found “three aboriginal lines of descent”. In 1990 the rank of domain was proposed: Archaea, Bacteria and Eucarya. In 2013 newer methods shook this picture. Eukaryotic core genes fell inside the Archaea. The authors wrote that the three-domains tree “may be incorrect”.
Since 2015 the Asgard archaea settled the matter. The first Asgard genome carried eukaryote-like proteins. In 2023, 63 new genomes placed eukaryotes “as a well-nested clade within Asgard archaea”. And in 2019 the first Asgard cell was cultured after twelve years of work: Prometheoarchaeum syntrophicum, a slow organism of about 550 nm that doubles in two to three weeks.
How old is the common ancestor? The book gives three answers side by side. One study says about 4.2 Ga. Another says more than 3.9 Ga. A third says 3.5–3.8 Ga. The book does not average them. It records the spread and moves on.
The rock record tells the same story in stone. A 4.10 Ga zircon from Jack Hills holds graphite with a carbon signature of possible life. Putative microfossils in Quebec may be 3,770–4,280 million years old. Isua sediments in Greenland hold isotopically light carbon from about 3.7 Ga. Younger records cluster in Australia: hot-spring deposits at 3.48 Ga, sulphate reduction at 3.47 Ga, methanogenesis before 3.46 Ga.
LUCA itself has been reconstructed twice. One study traced 355 gene families to it. Another models a genome of 2,657 proteins. Both describe the same organism: anaerobic, CO₂-fixing, thermophilic.
Genes that move sideways
The book’s sharpest finding is that the tree of life is mostly not a tree. Genes move between lineages.
The numbers are stark. In 181 prokaryotic genomes, “at least 81 ± 15 %” of each genome’s genes took part in lateral transfer. In E. coli, 17.6 % of genes entered by transfer since it split from Salmonella. Between domains, bacteria-to-archaea flow is more than five times the reverse. Against this traffic, the tree-like core is tiny: about 1 % of an average prokaryotic proteome.
The cell that ate
The second connection is the deepest union of all. One cell took in another and kept it.
The mitochondrion dates between 2.05 and 1.21 Ga. The plastid, taken from a cyanobacterium, carries five different dates spread over 0.85 Gyr. Both events happened once. Two later primary uptakes are known: Paulinella and UCYN-A, which has gained organelle status.
Each uptake follows the same pattern. The bacterium loses genes to its host. Rickettsia, a free-living relative, has 834 protein genes. Human mitochondria keep 13. Synechocystis has 3,168 proteins; the Arabidopsis plastid keeps 87, with about 4,500 moved to the nucleus.
Plastids then passed “from hand to hand” through further eatings. Red and green nucleomorphs — tiny remnant nuclei — both hold three chromosomes. Insect bacteria show the same shrinking: Buchnera in aphids keeps 583 genes, and the smallest symbiont, Nasuia, only 112 kb.
Some animals steal chloroplasts outright. Sea slugs keep stolen plastids working for months. An Antarctic dinoflagellate keeps them for at least 30 months.
Becoming many
Multicellularity arose again and again. The book counts more than 25 origins across eukaryotes. At least seven arose by aggregation. Many arose by clonal growth, in red and green algae and once in brown algae.
The dates run from cyanobacteria at 2.45–2.22 Ga to red algae at 1.047 Ga, animals at 574–558 Ma, land plants at 515–470 Ma, and volvocine green algae at 250–50 Ma.
Animals and their microbes
Every animal carries a microbial load. A human holds about as many bacterial cells as human cells — 3.8 against 3.0 × 10¹³ — almost all in 0.4 litres of colon. Microbial fermentation supplies about 10 % of human energy. In a ruminant it supplies about 70 %.
The partnerships are old and specific. Hominid-gut bacteria cospeciation is dated at 15 Myr. Coral photosymbiosis is about 212 Myr. Bioluminescence in deep-sea fish, powered by housed bacteria, is about 150 Myr. The bacterium Wolbachia lives in 40–66 % of all arthropod species.
Squid flush their light-organ symbionts each dawn and regrow them within hours. Bees build a gut community of a billion bacteria in ten days.
Fungi between the kingdoms
Fungi sit between animals and plants and connect both. Clocks date fungi at roughly 1,000–1,760 Ma. Fossils run from 1,010–890 Ma to glomalean spores at ~460 Ma.
The partnerships multiply. Arbuscular mycorrhiza form one phylum that touches most plants. Ectomycorrhiza arose 66–82 times in fungi and 30 times in plants. Lichens arose 20–30 times. Wood-rot fungi split the work: white rot dates to ~300 Ma; brown rot arose at least four times.
Animals even farm fungi. Ants, ambrosia beetles and termites evolved this agriculture at least twelve times, with ant agriculture dated at about 66.65 Ma.
Coming ashore together
Life did not colonise land alone. It went as a community.
Microbial mats stood on land from 3,220 Ma — 2,747 million years before plant spores. By clock, fungi lost their flagellum at ~720 Ma, still before plants. Arthropod tracks appear at the Cambrian–Ordovician boundary. Vascular plants follow at 426.9 Ma. Charcoal — proof of burning vegetation — appears at 423 Ma.
Then comes the jewel: the Rhynie chert, 411.5 Ma. It preserves a whole community. Plants of 10–40 cm with arbuscular mycorrhiza inside them. A lichen. About ten fungi. Nematodes. Seven arthropod groups of 0.29–20 mm. The book reads it as the first snapshot of a working land ecosystem.
Eating and being eaten
The oldest predation trace is a hole in the oldest mineralised animal tube: 2.7 % of more than 500 tubes. The oldest parasitism is Cambrian: 47.8 % of 429 brachiopods bore encrusting parasites.
Predation intensified in steps. Drilling frequency rose from 10–20 % to 100 %. Eleven of 13 shell-breaking groups arose in the Triassic–Cretaceous.
The modern numbers are sharp. Predators average 42 times their prey’s mass. Parasitoids average 0.76 times their host’s. Mean energy transfer between levels is 10 %. Food chains run 3.5 to 5 levels. Parasitism arose at least 223 times in animals and 12 times in flowering plants. And plants outweigh animals 225 to 1.
Flowers, fruit and animals
Insects appear before flowers. Rhyniognatha, the earliest insect fossil, precedes the angiosperm clades by about 276 Myr. Pollination is recorded on gymnosperms at about 165 Ma, and on angiosperms from 100–99 Ma.
Today about 87.5–90 % of the 295,383 known angiosperm species are animal-pollinated, by about 350,000 pollinator species. Vertebrates are only 0.42 % of pollinators — but excluding them cuts seed set by 63 %.
Seed dispersal runs from 1 m, carried by ants, to 5.3 km, carried by elephants. Three-quarters of 52 measured mutualistic networks are nested: specialists attach to generalists, like branches to a shared trunk.
The cycles across kingdoms
The elements that life needs are passed between kingdoms in loops.
The dates: sulphate reduction at 3.47 Ga. Nitrogen fixation at 3.2 Ga by isotopes. Free oxygen at 2.46–2.33 Ga — the Great Oxidation.
No eukaryote carries the nitrogenase genes. Eukaryotes solve this by housing the fixers: in root nodules of about 10 of 380 angiosperm families, and in UCYN-A, which travelled from symbiont to organelle.
At sea, 10–50 % of production passes through bacteria. The carbon pump shrinks at each step: export is 5.8–8.6 % of global production, only 0.62 % reaches 1,000 m, and 0.19 % is buried. Animals pump elements upward and landward — from zooplankton to whales and seabirds, moving 99 Gg of phosphorus per year. That carrying capacity is now 5–8 % of its former value.
The whole network and the human knot
The book closes with the totals. Estimated species: plants 0.30–0.37 million, fungi 0.6–3.8 million, animals 7.8 million, all eukaryotes 8.7 million. Bacteria may number 1–6 billion species; virus particles 10³⁰–10³¹.
Described shares differ by group. Plants are nearly complete. Insects are about 20 % described. Fungi only 3.2–5.5 %.
The measured links are few: 1.18 pollinators per angiosperm, 3.0 hosts per plant virus, 4.3 viruses per plant. And one figure stands out: humans and their livestock now hold 94–96 % of all mammal mass.
The theory layer, briefly
The book’s second layer sets these patterns beside the author’s Vacuum.Net texts. Four pairings (“reads”) are made. One became a fixed test and passed: in 24 cases, no cell holds a chain of more than two nested partners, each with its own genome. A second test — that each supplier must be older than the one it supplies — failed: 5 of 14 cases. Across the author’s three books, the theory has been tested nine times with outcomes fixed in advance. One passed. The book prints the full score.
Conclusion
Read as biology, the book tells one story: life connected itself by every available route. Genes crossed lineages. Cells took in cells. Cells became bodies. Bodies took partners. Partners became farmers, pollinators, parasites and prey. The elements moved in loops through them all. Each connection arrived on ground that an earlier connection had prepared — microbes before fungi, fungi before plants, plants before pollinators, and finally humans, holding almost all the mammal mass on Earth.
Annotated Reference List
Konstapel, J. (2026). The Weaving of Life: How Animals, Plants, Fungi and Microorganisms Became Interwoven. Constable Research, Leiden, 13 October 2026. The primary work. Ten biology chapters (A1–A10), seven data tables (Part C) and a theory layer (Part B). Use: source of all facts in this essay.
Konstapel, J. (2026). The Vacuum.Net Theory: Complete Derivation. Constable Research, Leiden, 28 September 2026. The five axioms behind Layer B. Why read: to follow the theory statements the book tests.
Konstapel, J. (2026). The Balanced Three, the Balanced Tree and the Vacuum.Net Theory. Constable Research, Leiden, 29 August 2026. Closure rules and the tree structure of the model. Why read: its §11 maps the model onto species and communities.
Konstapel, J. (2026). The Climate on the Ladder of Three. Constable Research, Leiden, 3 October 2026. Defines the logarithmic “position” coordinate used for every size and time in the book. Why read: needed for the book’s coordinate system.
Konstapel, J. (2026). Plants on the Ladder and The Company of Plants. Constable Research, Leiden, 12–13 October 2026. The two earlier books in the series, on photosynthetic life and plant co-evolution. Why read: they hold the six earlier theory tests.
Woese, C. R., and Fox, G. E. (1977). “Phylogenetic structure of the prokaryotic domain: the primary kingdoms.” PNAS 74, 5088–5090. The ribosomal RNA comparison that first split life into three lines. Why read: the origin of the three-domain view (chapter A1).
Spang, A., et al. (2015). “Complex archaea that bridge the gap between prokaryotes and eukaryotes.” Nature 521, 173–179. The first Asgard (Lokiarchaeota) genome. Why read: the discovery that placed eukaryotes inside the Archaea.
Imachi, H., et al. (2020). “Isolation of an archaeon at the prokaryote–eukaryote interface.” Nature 577, 519–525. The first cultured Asgard cell, Prometheoarchaeum syntrophicum, after a decade of cultivation. Why read: the closest living relative of the host that took in the mitochondrion.
Williams, T. A., et al. (2013). “An archaeal origin of eukaryotes supports only two primary domains of life.” Nature 504, 231–236. The two-domain proposal of 2013. Why read: the paper that challenged the three-domain tree.
Nelson-Sathi, S., et al. (2015). “Origins of major archaeal clades correspond to gene acquisitions from bacteria.” Nature 517, 77–80. Documents 2,264 gene acquisitions from bacteria at the origins of 13 archaeal taxa. Why read: evidence for the dominant direction of gene flow between domains.
Quammen, D. (2018). The Tangled Tree: A Radical New History of Life. Simon & Schuster, 496 pp. Background book listed by the author. Why read: the narrative history of the shift from tree to web of life.
Sheldrake, M. (2020). Entangled Life. Random House, 352 pp. Background book. Why read: fungi as the connective tissue of the living world — lichens, mycorrhiza and networks.
Note: the sources for chapters A3–A10 in the primary work are numbered per chapter; the entries above are the principal studies named in the text. The five background books (Quammen; Dawkins & Wong; Beerling; Sheldrake; Begon, Townsend & Harper) framed the topics only — no number in the primary work is taken from them.
The Company of Plants: A Verified Map of Co-evolution and What It Is Worth
An essay on J. Konstapel’s The Company of Plants (Constable Research, Leiden, 2026), with an annotated reference list.
1. What this book is
The Company of Plants is a 100-page reference work. It documents how plants evolved together with the rest of life. Ten chapters cover ten interplays: insects and pollination, fungi and mycorrhiza, bacteria and nitrogen, animals and seed dispersal, soil life, herbivores and defence, decomposers, climate and atmosphere, competition and forests, and viruses and pathogens.
The book follows an earlier work, Plants on the Ladder (12 October 2026). That earlier book treated fifteen groups of photosynthetic life. This one treats what that book left out: the partners.
The structure is unusual and disciplined. The book is built in two layers, kept strictly apart.
Layer A is biology only. Ten chapters, A1 to A10. Each chapter has the same nine sections: what the interaction is; first appearance; partners; sizes; times; counts; order of events; present quantities; and a closing Pattern of at most 150 words. Part C gathers the numbers into eight tables. Layer A and Part C mention no theory. They stand on their own.
Layer B asks one question of Layer A: do the measured patterns correlate with the structure of the Vacuum.Net theory, as its texts state them? Each pairing carries a status. Three tests were fixed before the data were seen. Their outcomes stand as they fell.
Every statement of fact carries a mark. Measured means read in a cited source. Computed means arithmetic on measured numbers, shown. Modelled means the output of a named published model. Text means a statement of a Vacuum.Net text. Read means a hypothesis pairing a text statement with a measured pattern.
The result is a rare thing: a self-published work with the evidentiary hygiene of a peer-reviewed monograph.
2. The method is the product
Most popular science books sell a story and let the reader trust the author. This book does the opposite. It sells verification.
Layer A holds 457 numbered references. Every number in the book traces to a source. Every size and time carries a position: the logarithm of the value over a natural unit of length or time. Positions are coordinates, not rounded scores. Items found only in secondary sources are listed under “Not verified” at the end of each chapter and are excluded from the book’s facts.
This is a method a business reader should recognize. It is the same discipline as an audited dataset. Each record has provenance. Each derived figure shows its arithmetic. Each uncertain item is flagged rather than absorbed. In a market where AI training data and curated scientific datasets are becoming paid products, provenance discipline is the asset.
3. What the numbers say
The book’s patterns are striking even stripped of all theory.
Insects are far older than flowers. The insect fossil record begins at 396–407 Ma. A molecular clock places the origin of insects at about 479 Ma. The main angiosperm clades diversified only 135–130 Ma. The gap is about 275 million years. Insects did not evolve to serve flowers. Flowers arrived into a world already full of insects.
Herbivory appears in pulses, always after the tissue. Plant tissue is eaten 54 to 98 million years after it first appears. Leaves, seeds, wood, roots: each is colonized in turn. The pattern repeats across the fossil record.
Pollination began on gymnosperms. The oldest definitive pollinating insects, with siphonate proboscides, date to about 165 Ma. Four gymnosperm pollination modes are counted between 125 and 90 Ma. Pollination of flowering plants is recorded only from 100 Ma. The partnership predates the flower.
The fungal side is older still. Fungus-like mycelial fossils occur in 2.4 Ga basalt. The oldest unambiguous fungus dates to 1,010–890 Ma. Crown Fungi sit at about 1,020 Ma on a six-gene clock. Arbuscular mycorrhiza in rock appears at about 460 Ma — before roots. Today 72% of vascular plant species form arbuscular mycorrhiza. The plant-fungus contract is the oldest standing agreement in the biosphere.
Obligate pairs recur, and the animal side comes first. Fig wasps and figs, yucca moths and yuccas, orchid bees and perfume orchids: each pair is 7 to 60 million years old. In each, the insect side is dated first or together with the plant. The service provider existed before the client.
The present-day balance is measurable. 295,383 angiosperm species. 349,368 known pollinator species — 1.18 per plant species. 87.5% of flowering plants are animal-pollinated. Herbivory explains 25–35% of diversification variance across 31 insect orders. Insect body lengths span a factor of 4,079, from a 139-micrometre wasp to a 567-millimetre stick insect.
These are facts a specialist can use. They are also facts a non-specialist can check, because the book shows every step.
4. A verification sample
An independent check of twelve core references from chapters A1 and A2 confirmed titles, venues, years and key figures without exception. Christenhusz & Byng (2016) is the accepted plant species count. Ollerton, Winfree & Tarrant (2011) is the source of the 87.5% animal-pollination figure and the 308,006 species estimate. Misof et al. (2014) is the phylogenomic study dating insect origins to about 479 Ma. Ren et al. (2009) documents the pre-angiosperm scorpionfly pollinators. Cai et al. (2018) documents the 99 Ma cycad-pollinating beetle. Brundrett & Tedersoo (2018) is the source of the 72% arbuscular mycorrhiza figure. Bengtson et al. (2017) reports the 2.4 Ga fungus-like fossils. Redecker, Kodner & Graham (2000) reports the 460 Ma Ordovician glomalean spores.
Two caution flags are worth recording, and the book’s own “Not verified” discipline makes this easy. Melittosphex burmensis, called the oldest bee at 100 Ma, has been reclassified by later work as incertae sedis within the Aculeata — its bee status is contested. Ourasphaira giraldae, dated about one billion years, is argued by its discoverers to be an early fungus, but the fungal affinity is not universally accepted. The book’s phrasing (“oldest unambiguous fungus”) is stronger than the literature supports. Neither flag affects the book’s patterns; both are entry points for a second edition.
5. Who can use this
The content is not operational data. No farmer plants a crop differently tomorrow because insects are 275 million years older than flowers. The value is strategic. Three user groups stand out.
Plant breeders. Breeding is applied co-evolution. Resistance genes come from wild relatives, which hold the defences won in millions of years of arms races with herbivores and pathogens. Mycorrhiza responsiveness is now a selectable trait. Pollination biology governs hybrid seed production. The book does not hand a breeder a gene. It hands the breeder a map: which interactions are ancient and stable, where resistance will likely hold, and where an arms race will likely break it.
Biologicals and pollination-services companies. The mycorrhizal inoculant market is young and troubled with quality problems. The pollination-services market is large and growing. Both industries sell exactly the partnerships chapters A1 and A2 document. A verified synthesis of the underlying biology is a credible foundation for white papers, staff training, and product strategy.
Education and research-data users. The eight tables of Part C form a clean, machine-readable dataset: first appearances, sizes, times, counts, shares. Every record carries a source mark and a computed position. Institutions increasingly license structured scientific content with provenance. This material fits that market directly.
Layer B — the Vacuum.Net pairings — serves none of these markets. That is not a criticism. The book itself keeps the layers apart precisely so that Layer A and Part C can be used without Layer B. Commercial deployment should follow the book’s own architecture: sell the verified layer, keep the theoretical layer in a separate track.
6. An honest market assessment
The total addressable market is small but real. The book will not sell in the tens of thousands. Its buyers are specialized: breeders’ strategy and education teams, biologicals companies, university libraries, data curators, and the audience for disciplined natural-history synthesis.
The value lies in three separable assets:
The compilation. 457 references synthesized into one consistent structure. Anyone could rebuild it; almost no one has. The cost of rebuilding is the price of the work.
The verification discipline. Marks on every fact, shown arithmetic, and an explicit “Not verified” quarantine. This is what separates the work from a blog post and justifies institutional purchase.
The narrative. Insects before flowers. Tissue eaten only after it appears. The service provider older than the client. These are memorable, teachable patterns, usable far beyond the reference audience.
The recommended route is the book’s own structure in commercial form: Part C as a licensed dataset, Layer A as an educational and advisory foundation, and the patterns as content. The theoretical layer stays in its own channel, where its audience is.
7. Conclusion
The Company of Plants is a modest, rigorous, and unusual book. It does what expensive data products promise: every number has a source, every derivation is shown, and every uncertain item is quarantined. Its biology is sound on independent sampling. Its patterns are genuinely instructive — about pollination, symbiosis, defence, and the deep antiquity of the partnerships agriculture now depends on.
Its commercial potential is modest in scale and high in credibility. The knowledge is gold for those who work with the living partnerships themselves — breeders, biologicals firms, and educators — provided it is delivered as a map and a dataset, not as a monograph. The book’s own two-layer discipline points the way: the verified layer is the product; the theory is the author’s own journey.
Annotated Reference List
A selection of the book’s 457 Layer-A references, annotated for the business reader. Each annotation states what the source establishes and why it matters commercially.
[1] Christenhusz, M.J.M. & Byng, J.W. (2016). “The number of known plants species in the world and its annual increase.” Phytotaxa 261: 201–217. The accepted census of known plant species: 295,383 angiosperms, 1,079 gymnosperms, roughly 374,000 plants overall, with about 2,000 new species named per year. Independently verified in this essay’s sample. Why it matters: the baseline denominator for any biodiversity claim, product claim, or dataset.
[2] Ollerton, J. (2017). “Pollinator diversity: distribution, ecological function, and conservation.” Annual Review of Ecology, Evolution, and Systematics 48: 353–376. Counts about 350,000 known pollinator species across insects, birds, mammals and lizards. Why it matters: the size of the service-provider pool behind the pollination-services industry.
[3] Ollerton, J., Winfree, R. & Tarrant, S. (2011). “How many flowering plants are pollinated by animals?” Oikos 120: 321–326. Establishes that 87.5% of flowering plant species (308,006 species) are animal-pollinated; community means run from 78% in temperate to 94% in tropical sites. Independently verified. Why it matters: the single most quoted dependency figure in pollinator-sector communication and policy.
[6] Misof, B. et al. (2014). “Phylogenomics resolves the timing and pattern of insect evolution.” Science 346: 763–767. A phylogenomic analysis of 1,478 genes placing the origin of insects at about 479 Ma and insect flight at about 406 Ma. Independently verified. Why it matters: anchors the book’s central asymmetry — insects long predate flowers — and underpins the argument that pollinators were available before the client existed.
[10] Labandeira, C.C. & Currano, E.D. (2013). “The fossil record of plant-insect dynamics.” Annual Review of Earth and Planetary Sciences 41: 287–311. Documents the two pre-angiosperm herbivory pulses and the recovery intervals after mass extinctions: 24 Myr after the Permian–Triassic, about 10 Myr after the K–Pg. Why it matters: the template for how fast plant-herbivore systems rebuild after collapse — relevant to resilience thinking in crop protection.
[11] Ren, D. et al. (2009). “A probable pollination mode before angiosperms: Eurasian, long-proboscid scorpionflies.” Science 326: 840–847. Eleven species in three extinct families fed on gymnosperm ovular secretions across a 62-million-year interval. Independently verified. Why it matters: proves the pollination service predates the flower — the biological licence for managed pollination beyond bees.
[13] Cai, C. et al. (2018). “Beetle pollination of cycads in the Mesozoic.” Current Biology 28: 2806–2812. A 99 Ma boganiid beetle with cycad pollen, including measured pollen grain dimensions. Independently verified. Why it matters: a data point with physical measurements — the kind of record that makes Part C usable as a dataset rather than a story.
[15] Poinar, G.O. & Danforth, B.N. (2006). “A fossil bee from Early Cretaceous Burmese amber.” Science 314: 614. Describes Melittosphex burmensis, about 100 Ma, long treated as the oldest bee. Verified as published; later work reclassifies the lineage as incertae sedis within Aculeata. Why it matters: a caution example — the book’s “Not verified” discipline exists precisely for cases like this.
[33] Estimated insect diversity (about 5.5 million species, roughly 80% undescribed). As cited in the book from the entomological literature. Why it matters: the discovery gap is a market signal — most of the pollinator and herbivore cast remains formally unknown.
[34] Herbivory shares by insect order (Lepidoptera ~100%, Orthoptera ~94%, Hemiptera ~80%, Coleoptera ~26%, Hymenoptera ~7%); herbivory explains 25–35% of diversification variance across 31 orders. As cited from the macroevolutionary literature. Why it matters: quantifies the herbivore pressure that crop defence breeding works against.
[47] Brundrett, M.C. & Tedersoo, L. (2018). “Evolutionary history of mycorrhizal symbioses and global host plant diversity.” New Phytologist 220: 1108–1115. Establishes that 72% of vascular plant species form arbuscular mycorrhiza, 2.0% ectomycorrhiza, 1.5% ericoid, 10% orchid. Independently verified. Why it matters: the market-size basis for the mycorrhizal inoculant industry, and the breeder’s case for mycorrhiza-responsive cultivars.
[49] Bengtson, S. et al. (2017). “Fungus-like mycelial fossils in 2.4-billion-year-old vesicular basalt.” Nature Ecology & Evolution 1: 0014. Reports mycelium-like filaments in 2.4 Ga South African basalt. Independently verified; the fungal interpretation remains cautious (“fungus-like”). Why it matters: the deepest anchor of the book’s timescale — the fungal occupation of rock is older than most multicellular life.
[50] Loron, C.C. et al. (2019). Ourasphaira giraldae, about 1,010–890 Ma. As cited in the book from the Canadian Arctic microfossil work. Verified as published; the fungal affinity is argued rather than settled. Why it matters: the second caution example — deep-time fungal claims carry interpretive risk, and the book flags rather than hides it.
[51] Six-gene molecular clock on 199 fungi (crown Fungi ~1,020 Ma; Glomeromycota ~659 Ma; embryophytes ~482 Ma). As cited in the book from the fungal phylogenomics literature. Why it matters: the clock that makes the mycorrhiza story coherent — the symbiosis machinery existed long before land plants needed it.
[52] Redecker, D., Kodner, R. & Graham, L.E. (2000). “Glomalean fungi from the Ordovician.” Science 289: 1920. Fossil hyphae and spores from about 460 Ma Wisconsin rocks, resembling modern Glomales. Independently verified. Why it matters: the physical evidence that the mycorrhizal service predates roots — the founding datum of the inoculant industry’s pitch.
Closure-Horizon Economics (CHE) breidt de economische theorieën van Hall en Woodford uit met het principe dat verwachtingen een begrensde tijdshorizon hebben. Onvolledige afsluiting van economische processen veroorzaakt restbelasting (loading), die zich opbouwt en crises mede voorspelbaar maakt. Het model verklaart consumptiegevoeligheid, begrenst de invloed van toekomstig monetair beleid en laat parameters veranderen met de economische toestand. De auteur rapporteert dat kredietgroei financiële crises voorspelt: de gemeten voorspelkracht bedraagt AUC 0,751 en 0,698 in twee testperioden.
Wat de Nobelprijs Economie 2026 goed zag, en waar het model ophoudt
Op 12 oktober 2026 kregen Robert E. Hall en Michael Woodford de Nobelprijs voor Economie. Het comité prees hun werk aan de “micro-economische grondslagen van macro-economische stabilisatie”. NRC vatte het samen: overheden zijn beter voorbereid op crises. Recessies duren korter.
Die prijs is verdiend. Na de oliecrisis lag het keynesiaanse model in scherven. De overheid greep in, en toch stegen werkloosheid en inflatie tegelijk. Hall en Woodford bouwden de theorie opnieuw op. Niet vanuit het totaal, maar vanuit wat huishoudens en bedrijven werkelijk doen.
Hun werk laat drie verschijnselen open. Alle drie krijgen een plaats zodra één regel wordt toegevoegd. Die regel komt uit de Vacuum.Net-theorie.
Wat Hall en Woodford lieten zien
Hall, 1978. Een huishouden dat zijn uitgaven over het hele leven spreidt, stopt alles wat het weet al in de uitgaven van vandaag. De uitgaven van morgen zijn dan niet te voorspellen. Alleen nieuw nieuws verandert ze.
Woodford, 2003. Bedrijven passen hun prijzen af en toe aan. Bij elke aanpassing kijken ze vooruit. De inflatie van vandaag is daardoor de optelsom van alle verwachte spanning in de toekomst. Wie die verwachtingen beïnvloedt, beïnvloedt de inflatie. Daarom praten centrale banken tegenwoordig over hun toekomstige rente. Dat heet forward guidance.
Het model rekent met een weegfactor van 0,99 per kwartaal. Een belofte voor over vijf jaar weegt dan nog 0,99²⁰ = 82 procent. De horizon van het model is 100 kwartalen: 25 jaar.
Drie dingen die niet passen
De helft van de huishoudens volgt het model niet. Campbell en Mankiw maten in 1989 dat ongeveer de helft van de consumptie meebeweegt met het inkomen van deze maand. Volgens Hall zou dat nul moeten zijn.
Beloften werken veel zwakker dan het model zegt. In het model werkt een renteverlaging die voor over vijf jaar wordt aangekondigd sterker dan een verlaging vandaag. In de praktijk gebeurt dat nergens. Economen noemen dit de forward-guidance-puzzel.
Crises kondigen zich aan. Krediet groeit jaren voor een bankencrisis harder dan de economie. In het model komen grote crises van buitenaf, als schok. Een schok is per definitie niet te zien aankomen.
Voor elk van de drie bestaat een aparte reparatie. Elke reparatie voegt een getal toe dat wordt afgesteld tot het past.
Eén toegevoegde regel
De Vacuum.Net-theorie beschrijft de werkelijkheid als een net. Wat blijft bestaan, is een draad die op zichzelf terugkomt: een winding. Elke winding heeft een sluitingstijd. Een aandelentransactie sluit in dagen. Een prijs blijft een jaar staan. Een lening sluit in jaren.
Daaruit volgen drie regels voor de economie.
Een verwachting reikt niet verder dan de sluitingstijd van haar winding. Een huishouden dat per maand rondkomt, rekent in maanden. Een prijs die een jaar staat, rekent in een jaar. Verder kijkt geen van beide.
Geen winding sluit exact. Er blijft altijd een rest over. Die rest hoopt zich op. Dat heet belading.
De getallen in het model veranderen met de belading. Ze zijn geen constanten.
Dat is alles. De rest is rekenwerk.
Wat eruit volgt
De helft die per maand rondkomt
Stel dat een huishouden H maanden vooruit rekent. Dan beweegt zijn consumptie voor een deel 1/H mee met tijdelijk inkomen. Bij H = 1 volledig. Bij een onbeperkte horizon niet. Dat laatste is Hall.
Over alle huishoudens samen is de meebeweging het gemiddelde van 1/H. De gemeten waarde is 0,5. Eén verdeling die daarbij past: de helft van de huishoudens kijkt niet verder dan de volgende maand, de andere helft plant onbeperkt.
Dat getal gaat over mensen. Over wie aan het eind van de maand de huur betaalt en dan pas weet wat er over is. Voor hen bestaat een rentebelofte voor volgend jaar niet. Er is geen kring waarin zij kan landen.
Beloften doven uit
In het Engelse artikel staat een exacte voorwaarde. Een rentebelofte voor ver in de toekomst dooft uit als de sluitingstijd van de prijzen kort genoeg is. Met de gangbare getallen ligt de grens bij 6,8 kwartaal. Gemeten prijzen staan gemiddeld 8 tot 11 maanden stil, dus drie à vier kwartalen. Ze vallen ruim binnen die grens.
Het verschil is groot. Neem een renteverlaging van één procentpunt, aangekondigd voor over vijf jaar.
Effect vandaag, vergeleken met een verlaging nu
Model van Woodford
8,5 keer zo groot
Met de sluitingsregel
0,08 keer zo groot
Het oude model geeft een belofte voor 2031 meer kracht dan een besluit vandaag. Met de sluitingsregel houdt een belofte over vijf jaar 8 procent van de kracht van een besluit nu. De effectieve horizon is 1,1 jaar, geen 25 jaar.
De puzzel verdwijnt. Er is geen getal afgesteld. De horizon volgt uit hoe lang prijzen in winkels werkelijk stilstaan.
Crises zijn zichtbaar
Als crises van buitenaf komen, zegt geen enkel cijfer van vorig jaar iets over de crisis van dit jaar. Een voorspeller scoort dan precies 0,5 op de gangbare schaal: muntgooien.
Met belading ligt dat anders. De driejaarsgroei van krediet ten opzichte van het bbp is gemeten over achttien landen, van 1870 tot 2020. Het model werd vastgezet op de jaren tot 1960 en daarna niet meer aangepast. In twee latere perioden scoorde het 0,75 en 0,70.
De basiskans op een bankencrisis in een gewoon jaar is 3,9 procent. Bij een kredietgroei van 52 procentpunt in drie jaar is die kans vier keer zo groot.
De voorspelling van het standaardmodel, 0,5, valt in beide perioden af. De voorspelling van de sluitingsregel klopt in beide.
Onder druk wordt de curve steil
Hoe korter prijzen blijven staan, hoe sterker de inflatie op spanning reageert. Bij prijzen die vier kwartalen staan is de reactiefactor 0,086. Bij twee en een half kwartaal is hij 0,27: drie keer zo sterk.
In 2021 en 2022 gebeurde precies dat. Winkels en leveranciers pasten hun prijzen veel vaker aan. De inflatie schoot omhoog. Op zo’n moment werkt een belofte nog minder. Wat dan telt, is het besluit van vandaag.
Het register van de centrale bank
Een rentebesluit heeft drie standen: omhoog, gelijk, omlaag. In stappen van 25 basispunten is dat het drietallige register van het net: +1, 0, −1.
De ECB laat zien wat er gebeurt als het register breekt. Van juli 2022 tot september 2023 ging de depositorente van −0,50 naar 4,00 procent. Dat zijn 18 stappen in 10 vergaderingen. Acht keer werd een dubbele of driedubbele stap gezet. De verwachtingen lagen niet vast.
Van juni 2024 tot juni 2025 ging de rente van 4,00 naar 2,00 procent. Dat ging in acht enkele stappen van −25. Elke stap was aangekondigd.
Wat dit betekent
Hall had gelijk: het pad van de consumptie is niet te lezen. De sluitingsregel voegt toe: de ontlading wel.
Woodford had gelijk: beleid werkt via verwachtingen. De sluitingsregel voegt toe: een verwachting reikt zo ver als de kring die haar draagt.
Het comité zegt dat crises korter duren. Dat klopt. Maar de rest verdwijnt niet als de ontlading wordt gedempt. Zij schuift door naar de diepere winding: het krediet. Of dat zo is, is per land af te lezen aan de kredietgroei sinds 2008.
Er zitten mensen achter die cijfers. De helft van de huishoudens leeft binnen een maand. Voor hen is stabiliteit geen belofte over 2031. Het is een maand die sluit.
Wat bewezen is en wat niet
Uitspraak
Status
Hall en Woodford zijn grensgevallen van de sluitingsregel
bewezen
Meebeweging consumptie = gemiddelde van 1/H
bewezen onder regel 1, gecontroleerd met simulatie
Grens voor uitdovende beloften
bewezen, getallen nagerekend
De puzzel is het geval zonder sluiting
bewezen
Bij crises van buitenaf is voorspellen muntgooien
bewezen
Kredietbelading scoort 0,75 en 0,70
gemeten, in twee blinde perioden
Horizon volgt uit de gemeten prijsduur
regel 1; nog te toetsen tegen gebeurtenisstudies
Reactiefactor stijgt met belading
richting bevestigd in 2021–2022; vorm open
Geannoteerde referenties
J. Konstapel — Closure-Horizon Economics: How the Vacuum.Net closure rule extends the Hall–Woodford theory of stabilisation (2026).Waarom lezen? Het Engelse artikel achter deze blog. Het bevat de vier stellingen met volledige bewijzen en alle tabellen. Leesadvies: zelfbevattend; paragraaf 6 bevat de grens voor uitdovende beloften.
Robert E. Hall — Stochastic Implications of the Life Cycle–Permanent Income Hypothesis (Journal of Political Economy, 1978).Waarom lezen? Het artikel waarmee de prijs begon. Hall draaide de vraag om: niet wat de consumptie voorspelt, maar dat niets haar voorspelt. Leesadvies: de eerste twee secties. Let op dat Hall zelf al vond dat aandelenkoersen wel iets toevoegden.
Michael Woodford — Interest and Prices (Princeton University Press, 2003).Waarom lezen? Het standaardwerk van het nieuwkeynesiaanse model. Hier staan de twee vergelijkingen die de sluitingsregel aanpast. Leesadvies: hoofdstukken 3 en 4. Technisch, maar helder geschreven.
Michael Woodford — Monetary Policy Analysis When Planning Horizons Are Finite (NBER Macroeconomics Annual, 2019).Waarom lezen? Woodford zag het probleem zelf en voerde een eindige planningshorizon in. Dat ligt dicht bij regel 1. Het verschil: bij hem wordt de horizon gekozen, hier volgt hij uit een gemeten sluitingstijd. Leesadvies: de inleiding en de paragraaf over forward guidance.
John Campbell en Gregory Mankiw — Consumption, Income, and Interest Rates (NBER Macroeconomics Annual, 1989).Waarom lezen? De meting dat de helft van de consumptie het inkomen van nu volgt. Dat getal wordt in de sluitingsregel een gemiddelde horizon van twee perioden. Leesadvies: de schattingstabellen zijn de kern.
Marco Del Negro, Marc Giannoni en Christina Patterson — The Forward Guidance Puzzle (Federal Reserve Bank of New York, 2012).Waarom lezen? De nauwkeurige beschrijving van de puzzel. Wie hem eenmaal ziet, ziet hoe ver het model van de praktijk af ligt. Leesadvies: de eerste twee secties.
Xavier Gabaix — A Behavioral New Keynesian Model (American Economic Review, 2020).Waarom lezen? Gabaix voerde een “cognitieve korting” van ongeveer 0,85 per kwartaal in. De sluitingsregel komt uit op 0,75 tot 0,78, maar dan uit gemeten prijsduur in plaats van afstelling. Leesadvies: de eerste secties en de resultaten over forward guidance.
Emi Nakamura en Jón Steinsson — Five Facts about Prices (Quarterly Journal of Economics, 2008).Waarom lezen? De metingen van hoe lang prijzen werkelijk stilstaan: acht tot elf maanden. Dat is de sluitingstijd waaruit de horizon volgt. Leesadvies: het eerste feit en de bijbehorende tabel.
Pierpaolo Benigno en Gauti Eggertsson — It’s Baaack: The Surge in Inflation in the 2020s (NBER, 2023).Waarom lezen? Laat zien dat de Phillipscurve steil wordt als de economie onder druk staat. Bevestigt de richting van regel 3. Leesadvies: de empirische delen over 2021–2022.
Moritz Schularick en Alan Taylor — Credit Booms Gone Bust (American Economic Review, 2012).Waarom lezen? Het onafhankelijke bewijs dat kredietgroei crises aankondigt, over 140 jaar. Leesadvies: de sectie over voorspelling.
Òscar Jordà, Moritz Schularick en Alan Taylor — Macrofinancial History (macrohistory.net).Waarom lezen? De database van achttien landen sinds 1870 waarop de kredietbelading is getoetst. Vrij beschikbaar. Leesadvies: begin bij de documentatie van de crisisdata.
Robert Lucas — Econometric Policy Evaluation: A Critique (1976).Waarom lezen? De vaststelling dat de getallen in een model meebewegen met het beleid. Dat is regel 3 in economische taal. Leesadvies: kort; in één keer te lezen.
J. Konstapel — The Comma in the Market (constable.blog, 5 september 2026).Waarom lezen? De dubbelblinde toets achter de kredietbelading: twee meters die standhielden, zes die afvielen. Leesadvies: secties 4 en 5 bevatten de getallen.
J. Konstapel — Vacuum.Net Theory: hoe massa, energie en ruimte één net vormen (constable.blog, 28 september 2026).Waarom lezen? Het net zelf: draad, winding, knoop en maas, en de sluitingsregel waaruit regel 1 komt. Leesadvies: de eerste paragrafen volstaan.
Het artikel betoogt dat de 800 miljoen jongeren zonder duidelijke baan geen tekort aan werk hebben, maar dat hun werk niet telt omdat het geen rendement voor kapitaal oplevert. Waar deze spanning niet in een gesloten kring wordt opgevangen, ontlaadt ze via drugs, criminaliteit, migratie of mentale terugval. Drie bewezen programma’s en drie conceptuele lussen (werk, energie, kennis) laten zien hoe werk wél kan tellen zonder afhankelijk te zijn van buitenlands kapitaal. De kern is dat een daad pas sluit als een ander haar afmaakt, en dat de tijd dringt: buffers raken uitgeput tussen oktober 2027 en 2028.
Zonder banengroei zal de onrust toenemen, stelt Wereldbank-president Ajay Banga: ‘We moeten ervoor zorgen dat jonge mensen een productieve toekomst hebben.’ Foto: Sakchai Lalit/AP
Wereldbank-president Ajay Banga rekent het voor. In ontwikkelingslanden bereiken de komende tien jaar 1,2 miljard jongeren de werkende leeftijd. Er komen ongeveer 420 miljoen banen bij. Er blijven bijna 800 miljoen jongeren over zonder duidelijke weg.
Zijn antwoord is banen, gefinancierd met privaat kapitaal. De Wereldbank haalde in het boekjaar 2026 een record op: $112 miljard, ruim drie keer zoveel als in 2022.
Deel dat bedrag door de 800 miljoen jongeren. Dat is $140 per persoon, voor één jaar. Daarmee sluit je de kloof niet.
De kloof is ook geen tekort aan werk. Werk is er genoeg: daken, oogsten, water, zorg voor ouderen. Het telt alleen niet. Werk telt pas als baan wanneer iemand met kapitaal er rendement op verwacht. Wat niet rendeert, wordt niet afgewezen. Het wordt niet gezien.
Dat is de kern. Banga’s jongeren zijn geen mislukte werknemers. Ze zijn ongelezen werk.
Het beeld
Het beeld door dit hele werk is een visnet in water. Eén draad loopt overal en kruist zichzelf. Waar een lus in fase terugkomt, sluit ze en blijft ze. Waar ze niet sluit, blijft spanning staan. Die spanning verdwijnt niet. Ze ontlaadt ergens.
Een daad sluit pas als een ander haar afmaakt. Een gerepareerd dak sluit als de eigenaar zegt: klopt. In de geldeconomie sluit het pas als een belegger in Londen rendement ziet.
Hoe het geld nu werkt
Banga noemt zelf hoe het geld binnenkwam:
ongeveer 40% van de publieke financiering ging naar beter bestuur, zodat beleggers voorspelbare regels vinden;
de risicodekking groeide naar $25 miljard, grotendeels herverzekerd in Zwitserland en Londen;
bij sommige projecten wordt het rendement gemaximeerd, zodat er meer overblijft voor commerciële financiers;
kleine projecten worden gebundeld: één waterproject in Kenia is te klein voor een pensioenfonds, 22 samen ter waarde van $1 miljard niet.
Elke stap past het werk aan de belegger aan. Het water in Kenia komt er als het een vorm heeft die een fonds kan kopen. Daarna gaat de opbrengst de regio uit.
Opvallend is dat Banga zelf de les trekt die dit omkeert. Uit de energiecrisis leerde hij dat afhankelijkheden binnen een regio beter te beheersen zijn dan lange ketens door smalle corridors. Zijn voorbeeld: Ethiopian Airlines bleef vliegen dankzij een raffinaderij in Nigeria. Hij past die les toe op raffinaderijen en havens. Nog niet op geld, werk en kennis.
Waar de spanning heen gaat
Banga noemt drie routes: drugs, criminaliteit en migratie. De metingen laten zien hoe ze werken.
Migratie. In 24 Afrikaanse landen dacht in 2024 47% van de mensen aan vertrek. De helft daarvan noemt werk als reden. Niet de armsten vertrekken. Emigratie stijgt met het inkomen tot ongeveer $10.000 per hoofd. Wie niets heeft, kan de reis niet betalen.
Op 30 en 31 juli 2026 zwommen ongeveer 60.000 mensen van Marokko naar Ceuta. Op hun telefoons stond een bericht dat wie de overkant haalde een verblijfsvergunning kreeg. Het was niet waar. Minstens 57 mensen verdronken. Ruim 48.000 gingen binnen enkele dagen terug. In Marokko is 35,8% van de jongeren werkloos.
In Nepal zette een verbod op 26 sociale-mediaplatforms in september 2025 het land in brand. Voor jonge Nepalezen zijn die platforms vacaturebank, nieuwsbron en werkplaats. Een kwart tot een derde van het nationaal inkomen komt van migranten in het buitenland.
Geweld. Een grote jonge generatie verhoogt het risico op conflict. Maar werkloosheid alleen is niet de aanleiding. Onderzoek in Afghanistan, Colombia en Somalië vond dat onrecht de doorslag geeft: corruptie, discriminatie, geweld door politie en leger. Een jongere wordt dan wel gezien, maar verkeerd.
Drugs. In Bangladesh gebruiken 8,2 miljoen mensen drugs, bijna 5% van de bevolking. Bijna allemaal begonnen ze voor hun 26e. Negen op de tien zeggen dat drugs makkelijk te krijgen zijn. Eén op de acht heeft toegang tot behandeling. In Ghana slikken chauffeurs en bijrijders tramadol om de dienst vol te maken. In Amerika lag de sterfte aan overdoses vijf jaar na de sluiting van een autofabriek 85% hoger dan verwacht.
Een vierde route, naar binnen. Banga noemt hem niet. In rijke landen valt dezelfde generatie naar binnen uit. In Nederland groeit de WIA het snelst bij twintigers, vooral met psychische klachten. Een wereldwijd onderzoek onder een miljoen mensen vindt dat 41% van de jongvolwassenen in een crisis zit die het dagelijks functioneren aantast. Opvallend: jongeren in Afrika ten zuiden van de Sahara doen het het best, jongeren in rijke landen het slechtst. Daar zijn familiebanden en zingeving dunner.
Het patroon onder alle vier heeft een naam: waithood, wachten op volwassenheid. De overgang naar een eigen huis, een gezin en erkend werk komt niet. Waar familie sterk is maar werk ontbreekt, gaat de spanning naar buiten. Waar familie dun is maar een vangnet bestaat, gaat ze naar binnen.
Wat aantoonbaar werkt
Drie programma’s zijn goed genoeg getoetst om op te bouwen.
Liberia. 999 mannen met een hoog risico kregen acht weken groepstherapie en $200. Tien jaar later was hun crimineel gedrag ongeveer gehalveerd. Kosten: $530 per persoon.
Chicago. Jongeren uit gewelddadige wijken kregen een zomerbaan van acht weken met een mentor. Het aantal arrestaties voor geweld daalde met 43%, vooral ná het programma.
IJsland. Het land investeerde in begeleide activiteiten en tijd met ouders. Het aandeel 15- en 16-jarigen dat de afgelopen maand dronken was, daalde van 42% naar 5%.
Geen van de drie geeft inkomen op zich. Alle drie geven een daad die telt, iemand die haar ziet, en tijd om te herhalen. Dat is precies een gesloten lus.
Drie lussen zonder buitengeld
De vraag is of zo’n sluiting op de schaal van 800 miljoen kan. Daarvoor moeten drie lussen dicht.
Werk. De MAZE-bank draait al. Het is geen bank maar het geheugen van een kring. Eén regel is een zin: Karim hielp Fatima met het dak, dinsdag. Fatima tikt “klopt”, en dan telt het. Er is geen rente. Een stand die niemand gebruikt, zakt langzaam naar nul en wordt een gift aan de kring. Geld komt alleen aan de rand voor. De kring is nog niet een jaar lang in een echte gemeenschap getoetst. Die toets ligt klaar.
Energie. Banga’s kapitaalbehoefte zit grotendeels in energie: raffinaderijen, netten, datacenters. In 2021 mat een groep in Boulder stroom uit een laagje van enkele nanometers, zonder brandstof en zonder aangelegde spanning. Omgerekend is dat 70 watt per vierkante meter, dag en nacht. Dat is ruim drie keer een zonnepaneel in Nederland. Een Nederlands huishouden heeft er 4,9 vierkante meter aan genoeg. In vijf jaar heeft niemand het onafhankelijk herhaald. Of lagen optellen is onbekend. Maar als het standhoudt, is energie geen reden meer om kapitaal van buiten te halen.
Kennis. Banga’s “kleine AI” laat een boer zijn telefoon op een blad richten. Het systeem noemt de ziekte en het gif, en de boer drukt op kopen. Eén blad, één product, het geld de regio uit. Een AI met een poort die alleen kan teruggeven wat onderbouwd is, of zeggen “weet ik niet”, beschermt juist de boer die het antwoord niet kan controleren. Die poort draait al. De andere helft, die meet of een geheel in de pas loopt, is nog niet gebouwd.
Kunstmest
Banga’s derde zorg is de prijs van kunstmest. Toen de Straat van Hormuz dichtging, daalde de uitvoer van ureum uit de Golf in april 2026 met 83%. Kunstmest is drie problemen.
Stikstof komt uit de lucht. Het is een energievraag. Met de energie van één laag uit Boulder heb je ongeveer 2 vierkante meter nodig voor de stikstof van een hectare. Oppervlak is niet het probleem. Een kleine, lokale fabriek wel. Vlinderbloemigen, zoals bonen en klaver, doen het zonder machine.
Fosfor komt niet uit de lucht. Marokko heeft 68% van alle fosfaatreserves ter wereld. De jongeren die naar Ceuta zwommen, leven boven tweederde van alle fosfaat op aarde. Fosfor moet in de kring blijven: terugwinnen uit mest en as.
Kalium komt vooral uit Canada, Rusland en Wit-Rusland. Daar speelt de Golf geen rol.
Er is één route die niemand heeft onderzocht. Een Japans industrielab meldde dat een gelaagde film met zwaar waterstof het ene element in het andere omzet, telkens vier plaatsen hoger in het periodiek systeem. Pas dezelfde stap toe op keukenzout. Natrium wordt dan fosfor. Chloor wordt kalium. De rekensom klopt precies. Het verschijnsel is niet aanvaard, en de proef met zout is nooit gedaan. Die proef is klein en beslissend.
De klok
Banga’s zorg heeft een datum. Als de crisis nog één à twee jaar duurt, zegt hij, zijn de buffers op. Dat is tussen oktober 2027 en oktober 2028. Op de ladder van drie bereiken beide huidige oorlogen in februari 2028 dezelfde grens. Wat dan niet gesloten is, wordt een trage spanning.
Voor een jongere in Caïro of Dhaka betekent dat iets concreets. Bestaat er vóór die tijd geen kring waarin zijn werk telt, dan ontlaadt de spanning langs de routes die Banga zelf noemt.
Wat er staat en wat open is
Gemeten: de kloof, het geld, de routes, de drie werkende programma’s, de stroom uit één laag, de kunstmeststromen.
Open: een jaar MAZE-bank in een echte gemeenschap, de herhaling van de laag uit Boulder, de tweede helft van de AI, kleine stikstoffabrieken, en de proef met zout.
Elk van die open punten is goedkoper te beslissen dan één jaar van de kloof.
Leeslijst
De Boer, M. en Jongsma, M. (2026). “Wereldbank-president: ‘Wat gaan jongeren zonder hoop doen?'” Het Financieele Dagblad, 12 oktober 2026.Waarom lezen? Het interview waar dit blog op reageert. Leesadvies: lees de passage over regionale afhankelijkheden naast die over gebundelde waterprojecten. Ze wijzen in tegengestelde richting.
Konstapel, J. (2026). The Unread Generation. Constable Research, Leiden, 12 oktober 2026.Waarom lezen? Het Engelse artikel achter dit blog, met alle berekeningen, de status van elke bewering en ruim 40 geannoteerde bronnen. Leesadvies: paragraaf 5 voor wat werkt, paragraaf 7 voor kunstmest.
Wereldbank (2026). Persbericht over $112 miljard privaat kapitaal, 17 september 2026.Waarom lezen? Hoe de Bank succes meet: in opgehaald geld, niet in banen.
Afrobarometer (2024). Bijna de helft van de Afrikanen overweegt emigratie.Waarom lezen? Wie weg wil en waarom, in 24 landen.
Mercy Corps (2015). Youth & Consequences: Unemployment, Injustice and Violence.Waarom lezen? Het bewijs dat onrecht, niet werkloosheid, jongeren naar geweld drijft.
Blattman, C. e.a. (2023). “Cognitive Behavior Therapy Reduces Crime and Violence over Ten Years.” American Economic Review: Insights.Waarom lezen? Tien jaar volgen, gehalveerd crimineel gedrag, $530 per persoon.
Heller, S. B. (2014). “Summer jobs reduce violence among disadvantaged youth.” Science.Waarom lezen? Acht weken werk met een mentor, 43% minder geweld, vooral daarna.
Honwana, A. (2012). The Time of Youth. Kumarian Press.Waarom lezen? Het begrip waithood, uit gesprekken met jongeren in vier Afrikaanse landen. Leesadvies: begin met haar Lugard-lezing van 2013, twintig pagina’s.
Sapien Labs (2026). Global Mind Health in 2025.Waarom lezen? Een miljoen mensen, 84 landen, en de verrassende uitkomst dat Afrikaanse jongeren het mentaal het best doen. Leesadvies: lees de notitie over representativiteit erbij; alleen mensen met internet doen mee.
Konstapel, J. (2026). “Waarom Stijgen de Mentale klachten Wereldwijd?” constable.blog, 31 juli 2026.Waarom lezen? De vierde route, naar binnen, met de cijfers uit Nederland, Engeland en Finland.
Konstapel, J. (2026). “Hoe de MAZE Bank traditionele banken uitdaagt.” constable.blog, 8 september 2026.Waarom lezen? De kring, het verval en de drie getallen.
Konstapel, J. (2026). “Waarom ziekte geen raadsel is.” constable.blog, 16 september 2026.Waarom lezen? De toets met drie uitkomsten, waaruit dit blog de “ongelezen” jongere haalt.
Moddel, G. e.a. (2021). “Optical-Cavity-Induced Current.” Symmetry 13, 517.Waarom lezen? De meting achter elk energiegetal in dit blog. Leesadvies: figuur 4 voor de curve, paragraaf 3.2 voor de controles.
Konstapel, J. (2026). “Transmutation: Chemical Elements on Demand.” constable.blog, 9 oktober 2026.Waarom lezen? Wat er gemeten, gemeld en verworpen is over het veranderen van elementen. De bron van de zoutroute.
Konstapel, J. (2026). “Global Tensions 2026: February 2028 as a Turning Point.” constable.blog, 11 oktober 2026.Waarom lezen? De klok achter de datum van februari 2028.
U.S. Geological Survey (2026). Mineral Commodity Summaries 2026: Phosphate rock; Potash.Waarom lezen? Wie de fosfaat en het kalium van de wereld bezit.
De vraag: hoe hangt de evolutie van planten en bomen samen met de evolutie van het net uit de Vacuum.Net-theorie?
De regel: er wordt niets aan de theorie toegevoegd. Wat de theorie over haar eigen groei zegt, komt uit haar teksten. Wat over planten bekend is, komt uit metingen. Een verband is een paar van die twee met dezelfde vorm. Waar geen paar bestaat, staat dat er.
Elke uitspraak draagt een merk. Tekst: staat in een Vacuum.Net-tekst. Gemeten: gelezen in een genoemde bron. Berekend: rekenwerk, uitgeschreven. Gelezen: een koppeling die hier wordt gelegd. Een lezing is een hypothese.
Hoe het net groeit, in eigen woorden
Het net is één draad. Bij elke kruising staat zij in één van drie standen: +1, 0, −1. Drie buren vormen een venster. Het venster sluit als de drie standen optellen tot een drievoud. Van de 27 vensters sluiten er negen (tekst).
Een ring sluit helemaal alleen als elke stand gelijk is aan die drie plaatsen terug. Ringen met alle drie de standen bestaan alleen op diepten die deelbaar zijn door drie. Op diepte 3 zijn er twee, in tegengestelde richting. Op 4 en 5 geen. Op 6 weer twee, en een ring van 6 deelt in 3 + 3 (tekst). De teksten lezen diepte 3 als zelfhandhaving zonder deling en diepte 6 als de vorm van een eencellige (tekst).
Het net groeit volgens b(n) = 3·b(n−1) + b(n−2): 1, 3, 10, 33, 109, 360. De verhouding nadert σ₃ = (3 + √13)/2 (tekst), 3,3028 (berekend). De tweede term is het geheugen van het net (tekst).
Dan de twee-lagenwet (tekst). “σ₃ is de wet van het adres.” “φ is de wet van de spanning.” De afleiding van φ is drie zinnen. Tegenwinding op de eigen winding is ontwinding: Ψ² = 0, de sluiting verdwijnt. In de tijd heeft een bestaande sluiting daarom twee standen, laden en niet laden, en zij kan niet twee keer achter elkaar laden zonder eerst te ontladen. Dat is precies de voorwaarde die Fibonacci voortbrengt: x² = x + 1. Samengevat: adres is topologie, σ₃, groei in diepte, drietallig. Leven is spanning, φ, groei in de tijd, tweetallig. Twee lagen, één draad, nooit in één formule.
Ψ² = 0 is de nilpotente voorwaarde. “De algebra van de sluiting is de nilpotente voorwaarde Ψ² = 0” (tekst). Een aparte tekst zegt waar die vandaan komt: het stelsel van Rowlands, waarin de operator kwadrateert tot nul en het totaal van alles niets is (tekst).
Evolutie, in de teksten: “Een soort is een deelboom; een organisme is een sluiting op een diepte.” “Evolutie is geen selectie op frequentie, maar het vullen van diepten tot de volgende wordt afgedwongen.” De ruimte van diepte n is (3ⁿ − 1)/2: 1, 4, 13, 40, 121, 364. “Een groter getal dwingt de volgende diepte af” (tekst).
Ten slotte: waar drie twee ontmoet (tekst). De adresladder verdrievoudigt; de dimensieladder verdubbelt. Zij komen dicht bij elkaar waar 3^q ≈ 2^p. Dat zijn de convergenten van log₂3: 2/1, 3/2, 8/5, 19/12, 65/41, 84/53, 485/306. De noemers 1, 2, 5, 12, 41, 53, 306 zijn de sluitingsdiepten. Het quotiënt 3^q/2^p is de rest. Op diepte 12 is zij 1,01364: de komma van Pythagoras. In cents: −498, +204, −90, +23,5, −19,8, +3,6, −1,8 (berekend). De rest wordt nooit één. Elders: “geen sluiting is ooit exact: elke diepte laat een rest”; “een ladder die door een rest wordt gedreven, kan niet eindigen” (tekst).
Over planten zeggen de teksten niets. Geen blad, geen bladgroen, geen boom behalve als datastructuur en als naam naast laag 18 en 19 van de maatladder (tekst, nagezocht).
Hoe planten evolueerden, gemeten
De oudste aanvaarde bewijzen, met de positie op de tijdladder (berekend uit de ouderdom in seconden en T₀ = 1,520655 × 10⁻¹⁶ s):
Stap
Datering
Positie
Grote Oxidatie
2.501–2.434 Ma; elders 2,33 Ga
68,57–68,50
Oudste thylakoïden in een fossiele cel
1,78–1,73 Ga
68,24
Oudste zekere meercellige alg
1,047 Ga
67,78
Oudste meercellige groenalg
~1,0 Ga
67,73
Oudste landplantsporen
473–471 Ma
67,05
Oudste Cooksonia
433–427 Ma
66,97
Rhynie: huidmondjes, wortelassen
407,1 ± 2,2 Ma
66,92
Oudste hout
407–397 Ma
66,90
Oudste bos
393–388 Ma
66,88
Bomen van 8 m
~385 Ma
66,86
Oudste zaden
Laat-Devoon, 383–359 Ma
66,85–66,80
Genoomverdubbeling zaadplanten
~319 Ma
66,69
Genoomverdubbeling bloemplanten
~192 Ma
66,23
Oudste hele bloemplanten
≥ 124,6 Ma
65,84
Oudste aanvaarde grassen
67–65 Ma
65,26
Eerste C4-fotosynthese
32–25 Ma
64,5
De maten: cellulosevezel 4 nm op −2,21; thylakoïdmembraan 4,0 nm; granumherhaling 16,3 nm op −0,94; Prochlorococcus 0,5–0,7 µm op 2,2–2,5; tracheïde 10–42 µm op 4,9–6,2; vatlengte tot 27 cm op 14,19; Wattieza 8 m op 17,28; hoogste gemeten boom 112,7 m op 19,69 (gemeten; posities berekend). Tweeëntwintig posities. Het menselijk lichaam, van Bohrstraal tot langste man, beslaat er ook tweeëntwintig, ongeveer vier lager aan beide einden (berekend).
De tellingen (gemeten). In de lichtreactie: vier ladingen voordat zuurstof vrijkomt; Mn₄CaO₅; 35 chlorofyllen per kern van fotosysteem II, 96 per fotosysteem I; 14 eenheden in de ring van het ATP-synthase; 8 tot 12 lichtkwanta per O₂. In de wand: zes lobben van drie synthasen, 18 ketens per vezel; drie monolignolen in lignine; vier klassen hemicellulose. In het bouwplan: de eerste vaatplanten vertakken in twee gelijke dochters; de levenscyclus heeft twee fasen; twee bevruchtingen; twee oude genoomverdubbelingen; vier onafhankelijke oorsprongen van het platte blad; 62 oorsprongen van C4; secundaire groei in minstens zeven lijnen; meercelligheid tweemaal in elk van de drie fotosynthetische groepen. In het blad: parastichietellingen zijn opeenvolgende Fibonaccigetallen in 73,6 procent van 657 zonnebloemen; 8 en 13 in 6.000 dennenappels; en dezelfde Fibonacci-orde ontstaat in een natuurkundige proef waarin elk nieuw element wordt gelegd waar de vorige de meeste ruimte laten.
Waar de twee elkaar raken
Twee tellingen
De theorie houdt twee lagen uit elkaar: het adres, drietallig, groei in diepte; en de spanning, tweetallig en φ, groei in de tijd (tekst).
De groei van de plant in de tijd telt in tweeën en in φ (gemeten). De stengel vorkt in tweeën. De generaties wisselen in tweeën. Het genoom groeide door verdubbeling, tweemaal. De bladeren volgen Fibonacci.
De staande vorm van de plant telt op twee plaatsen in drieën (gemeten). De vezel wordt gesponnen door zes drietallen. Lignine bestaat uit drie alcoholen.
De rest van de tellijst is geen van beide. Vier, vijf, veertien, 35, 96, 62. Een lijst planttellingen is gemengd, en de tweeën en drieën eruit halen is selectie achteraf. Het paar rust dus op de vier feiten aan de tijdkant, niet op de lijst. Gelezen: matig aan de tijdkant, zwak aan de vormkant.
De rest
Dit is het paar dat staat.
De sluitingsdiepten van het net zijn de convergenten van log₂3. Elk is de beste rationale benadering van een irrationaal getal op zijn grootte; elk laat een rest; de rest wordt nooit één (tekst, berekend). Op diepte 12 is zij de komma.
Het blad is dezelfde rekensom (berekend, gelezen). De convergenten van 1/φ² zijn 1/2, 1/3, 2/5, 3/8, 5/13, 8/21: de klassieke bladbreuken. De convergenten van 1/φ zijn 1/1, 1/2, 2/3, 3/5, 5/8, 8/13: de gemeten parastichieparen, met 8/13 als zesde. De hoek die volgt, 360/φ² = 137,508°, is irrationaal in omwentelingen, dus geen blad staat ooit precies boven een eerder blad. Douady en Couder maakten dit buiten de biologie na en noemden de reden: het stelsel vermijdt rationale, periodieke schikkingen en wordt naar de gulden snede gedreven (gemeten).
De theorie leidt φ af uit Ψ² = 0: een sluiting kan niet twee keer achter elkaar laden (tekst). De plant legt elk nieuw blad waar de vorige ruimte laten, nooit op het laatste (gemeten). Dat zijn twee beschrijvingen van één beperking: de volgende stap kan de vorige niet herhalen. In het net geeft dat de Fibonacciregel voor groei in de tijd. In de plant geeft dat het Fibonacciblad. De gelijkstelling is een lezing. Zij voegt geen regel toe, want beide helften staan al op papier, de ene in de theorie, de andere in de meting.
En de rest is dezelfde rest. De ladder van het net “kan niet eindigen” omdat elke diepte een rest laat (tekst). De spiraal van het blad kan niet sluiten omdat 1/φ² geen laatste convergent heeft (berekend). De komma op diepte 12, 23,5 cents, en de afstand tussen 8/13 en 1/φ, 0,0026 (berekend), zijn hetzelfde soort getal: wat overblijft als drie tegen twee wordt gemeten, of het ene blad tegen het volgende.
Vullen tot het wordt afgedwongen
De teksten noemen evolutie “het vullen van diepten tot de volgende wordt afgedwongen” en een soort een deelboom (tekst). Het fossiele archief zegt dat dezelfde vormen vele malen onafhankelijk ontstonden: C4 62 keer, het platte blad vier keer, hout minstens zeven keer, meercelligheid tweemaal per groep (gemeten).
Gelezen: als een diepte wordt bereikt zodra de diepte eronder vol is, verschijnt dezelfde vorm in veel deelbomen rond dezelfde tijd en heeft zij geen enkele uitvinder. Dat is wat de tellingen laten zien.
Maar de gangbare verklaring leest dezelfde tellingen als convergente evolutie onder één druk, en voor C4 volgen de dateringen de daling van CO₂ in het Oligoceen (gemeten). De tellingen onderscheiden de twee lezingen niet. Het paar staat; het beslist nog niets.
Waar de twee elkaar niet raken
De diepten 3 en 6. De teksten lezen diepte 3 als stofwisseling en diepte 6 als de eencellige. Het eerste organisme in de plantenlijn, de cyanobacterie van twee miljard jaar geleden, is al een delende cel met erfelijkheid (gemeten). Haar op diepte 6 zetten is een lezing; maar dan ligt de hele plantenlijn op of boven diepte 6, en de teksten geven geen ringdiepte voor alg, landplant of boom. De reeks 3 → 6 → 3 + 3 beschrijft de stap naar de cel. Niet de stappen van de cel naar de boom. Geen verband in de zin van ringdiepte. In de andere zin van diepte, de laag van het klimaatartikel, heeft elke plantmaat er wél een: dat is de positie, en de plantenlijn vult de posities 2 tot 20 (zie boven).
De sluitingsdiepten. Diepte 12 is 24 mm en 81 ps; diepte 41 is 92 minuten; diepte 53 is 93 jaar. Geen enkele geverifieerde plantgrootheid ligt daar. Ladingsscheiding 3 ps, de stap naar het chinon 200–250 ps, het huidmondje 45–60 minuten, de naald 45 jaar, de eerste eikels van de eik na 50 tot 200 jaar (gemeten). Geen verband.
De ladder. De plantenlijn beslaat 22 posities in maat en 49 in procestijd, van 1 ps tot 4.900 jaar. De dateringen beslaan vier posities. Dat zijn coördinaten. Het boek From Atom to Body toetste de maten van het lichaam op ophoping bij hele posities en vond niets; de plantlijst hier is te kort om te toetsen. Geen bewering.
De proeven, eerst vastgelegd en daarna gedaan
Drie proeven zijn vastgelegd vóór er naar gegevens is gekeken, en op 12 oktober 2026 door Claude Code gedaan op de vooraf genoemde gegevens. De verantwoording is van de auteur. De uitslagen staan zoals ze vielen.
Proef 1, de tellijst. Bron vooraf gekozen: OpenStax Biology 2e, hoofdstukken 8, 30 en 32. Vereist: drievouden oververtegenwoordigd bij de staande vorm, tweeën en Fibonaccigetallen bij groei in de tijd, elk op één op honderd. Gevonden: 134 tellingen; 116 vorm, 11 tijd. Drievouden: 28 procent van de vormtellingen, 55 procent van de tijdtellingen. Tweeën en Fibonacci: 45 procent van de tijdtellingen, 61 procent van de vormtellingen. Beide p-waarden boven 0,9. Mislukt, in omgekeerde richting. In die tekst is twee de gewoonste telling van de staande vorm (twee sluitcellen, twee zaadlobben, twee spermacellen, twee fotosystemen) en drie de gewoonste telling van een proces (drie stadia, drie omlopen van de Calvincyclus). Het eerste paar overleeft zijn eigen proef niet.
Proef 2, de convergenten. Vereist: in de gegevens van Swinton en Fierz daalt de frequentie van elk Fibonaccipaar met zijn plaats in de reeks, en blijven niet-convergente paren onder de ruis van de auteurs of onder 10 procent. Gevonden: Swintons tellingen per waarde staan in een gegevensbestand dat niet te bereiken was; de volgordeproef kon daar niet. Het aandeel niet-convergent over alle 768 tellingen is 17,8 procent, boven de lat, al is het 5 procent bij de hoofden die de auteurs eenduidig noemen. Fierz: 0,40 procent niet-convergent in 6.000 dennenappels en 3,85 procent in 1.506 lariksappels, onder de lat; maar elk monster bevat één Fibonaccipaar, 8/13 of 5/8, zodat een volgordeproef over meerdere paren daar onmogelijk is. Mislukt volgens de regel zoals opgeschreven. De regel was slecht gemaakt: zij vroeg monsters van één soort om een reeks die zij niet kunnen bevatten. Wat de gegevens wél laten zien is smal: 97,3 procent van de dennenappels zit op één convergent van 1/φ, 8/13, en 99,6 procent op een paar van het convergententype. Dat is geen proef van het paar met het net; het is de gemeten helft ervan, bevestigd.
Proef 3, vullen tegen druk. Vereist: voor de gedateerde C4-oorsprongen past een Poissonfit van oorsprongen per 5 miljoen jaar tegen de opgehoopte C3-lijnen twee AIC-eenheden beter dan tegen CO₂, of omgekeerd. Gevonden: 53 oorsprongen met dateringen uit Christin 2008 en 2011; 92 procent in de laatste 20 miljoen jaar, mediaan 8,9 Ma, oudste 28,5 Ma. De CO₂-fit wint 50 AIC-eenheden van het nulmodel, helling −0,72 per 100 ppm. De gedateerde grassenboom die voor de lijnenkromme nodig is, was niet te bereiken, dus de vulkant is niet gebouwd. Niet gedaan. De CO₂-kant is zo sterk als de gangbare verklaring zegt; of de vulkant daaraan kan tippen, is niet bekend.
Stand van zaken
Uit de teksten. De sluitingsregel, de ringen per diepte, b(n) en σ₃, de twee-lagenwet, Ψ² = 0 → φ, vullen tot afgedwongen, de sluitingsdiepten en de komma, de restregel.
Gemeten. De dateringen, maten en tellingen hierboven, elk met een primaire bron in het dossier achter het essay.
Berekend. σ₃; de resten in cents; de convergenten van φ, 1/φ en 1/φ²; de gulden hoek; elke ladderpositie; de drie toetsgrootheden.
Gelezen, en getoetst. Paar 1, twee tellingen: zakte voor zijn proef, omgekeerde richting. Paar 2, de rest: de rekensom staat, de gemeten helft staat op 97,3 procent op één convergent, de proef van het paar zelf was verkeerd gesteld en moet opnieuw. Paar 3, vullen: niet getoetst; de drukkant past sterk.
Geen verband. De diepten boven de cel; de sluitingsdiepten; de ladder als meer dan coördinaat.
Niet beweerd. Dat het net een plantvorm veroorzaakt. Dat de teksten planten noemen. Dat het restpaar een proef heeft doorstaan.
Wat na de proeven overblijft is één ding. De theorie leidt φ af voor groei in de tijd uit Ψ² = 0. Het blad groeit in de tijd op φ, tot 97,3 procent op één paar in de gemeten appels. De twee uitspraken hebben dezelfde vorm. Of het dezelfde uitspraak is, staat open, en de proef die dat sluit moet opnieuw worden gebouwd, op gegevens met meer dan één Fibonaccipaar per monster.
Geannoteerde literatuur
Konstapel, J. (2026). The Remainder in the Leaf. Constable Research, Leiden, 11 oktober 2026.Waarom lezen? Het Engelse essay waarvan dit stuk de Nederlandse uitleg is; dezelfde paren, dezelfde getallen, met het dossier van gedateerde, gemeten en getelde feiten erachter. Leesadvies: deel 4.2 voor de rest, deel 6 voor de proeven.
Konstapel, J. (2026). The Balanced Three, the Balanced Tree and the Vacuum.Net Theory. Constable Research, Leiden, 29 augustus 2026; pdf Balanced Ternary and the Balanced Tree.Waarom lezen? De bron van alles wat hier over de groei van het net staat: b(n), σ₃, de twee-lagenwet, Ψ² = 0 → φ, en “het vullen van diepten tot de volgende wordt afgedwongen.” Leesadvies: §2, §4, §5 en §11.
Konstapel, J. (2026). The Climate on the Ladder of Three. Constable Research, Leiden, 3 oktober 2026.Waarom lezen? De sluitingsregel voluit, de periode-3-stelling, de ruimte (3ⁿ − 1)/2, en §3.2: waar de ladder van drie de ladder van twee ontmoet, met de komma.
Konstapel, J. (2026). Life and Consciousness from One Rule. Constable Research, Leiden, 29 september 2026.Waarom lezen? De ringen per diepte, de twee levende ringen op diepte 3, de kloof op 4 en 5, de deling op 6. De reden dat boven de cel niets wordt gevonden.
Konstapel, J. (2026). The Universe of Peter Rowlands. Constable Research, Leiden, 29 juni 2026. En Rowlands, P. (2007). Zero to Infinity: The Foundations of Physics. World Scientific, Series on Knots and Everything 41.Waarom lezen? Waar Ψ² = 0 vandaan komt: een operator waarvan het kwadraat nul is, en een totaal dat niets is.
Konstapel, J. (2026). From Atom to Body. Constable Research, Leiden, eerste druk.Waarom lezen? De ladder, de natuurlijke eenheden, de vier merken, en de proef die liet zien dat de ladder ordent maar niet kiest.
Douady, S., en Couder, Y. (1992). “Phyllotaxis as a physical self-organized growth process.” Physical Review Letters 68, 2098–2101.Waarom lezen? Fibonacci-orde gemaakt in een schaal ferrovloeistof: elke nieuwe druppel waar de vorige ruimte laten, en de gulden snede als ontsnapping aan elke rationale schikking. De gemeten helft van het sterkste paar.
Swinton, J., Ochu, E., en het MSI Turing’s Sunflower Consortium (2016). “Novel Fibonacci and non-Fibonacci structure in the sunflower.” Royal Society Open Science 3, 160091.Waarom lezen? 657 zonnebloemen; 565 van 768 tellingen Fibonacci. Het getal achter “73,6 procent”, en de minderheid die Proef 2 nodig heeft.
Fierz, V. (2015). “Aberrant phyllotactic patterns in cones of some conifers: a quantitative study.” Acta Societatis Botanicorum Poloniae 84, 261–265.Waarom lezen? 6.000 appels van één den, 8/13 overheersend, en de afwijkende patronen gerangschikt naar hun afstand tot 0,618.
Nixon, B. T., e.a. (2016). “Comparative structural and computational analysis supports eighteen cellulose synthases in the plant cellulose synthesis complex.” Scientific Reports 6, 28696. En Hill, J. L., Hammudi, M. B., en Tien, M. (2014). Plant Cell 26, 4834–4842.Waarom lezen? Zes lobben van drie: de ene telling van drie in de vorm van de plant die tot op het molecuul is gemeten.
Vanholme, R., Demedts, B., Morreel, K., Ralph, J., en Boerjan, W. (2010). “Lignin biosynthesis and structure.” Plant Physiology 153, 895–905.Waarom lezen? De drie monolignolen, bij naam.
Edwards, D., Davies, K. L., en Axe, L. (1992). “A vascular conducting strand in the early land plant Cooksonia.” Nature 357, 683–685.Waarom lezen? De eerste vaatplant: gelijk vertakkende assen, twee gelijke dochters, en de eerste tracheïden.
Jiao, Y., e.a. (2011). “Ancestral polyploidy in seed plants and angiosperms.” Nature 473, 97–100.Waarom lezen? De twee verdubbelingen van het plantengenoom, rond 319 en 192 miljoen jaar.
Niklas, K. J., en Kutschera, U. (2010). “The evolution of the land plant life cycle.” New Phytologist 185, 27–41.Waarom lezen? De twee generaties, en de voorouder die er maar één had.
Sage, R. F., Christin, P.-A., en Edwards, E. J. (2011). “The C4 plant lineages of planet Earth.” Journal of Experimental Botany 62, 3155–3169. En Christin, P.-A., e.a. (2008). “Oligocene CO₂ decline promoted C4 photosynthesis in grasses.” Current Biology 18, 37–43.Waarom lezen? Tweeënzestig onafhankelijke oorsprongen, en de datering die de eerste aan de CO₂-daling bindt. De gegevens voor Proef 3.
Boyce, C. K., en Knoll, A. H. (2002). “Evolution of developmental potential and the multiple independent origins of leaves in Paleozoic vascular plants.” Paleobiology 28, 70–100.Waarom lezen? Vier oorsprongen van het platte blad, uit 641 fossiele soorten.
Demoulin, C. F., Lara, Y. J., Lambion, A., en Javaux, E. J. (2024). “Oldest thylakoids in fossil cells directly evidence oxygenic photosynthesis.” Nature 625, 529–534.Waarom lezen? De oudste fotosynthese die rechtstreeks is gezien, 1,75 miljard jaar oud, en het oudste aanvaarde cyanobacteriefossiel op 2,0.
Rubinstein, C. V., Gerrienne, P., de la Puente, G. S., Astini, R. A., en Steemans, P. (2010). “Early Middle Ordovician evidence for land plants in Argentina.” New Phytologist 188, 365–369.Waarom lezen? De oudste landplantsporen, 473–471 miljoen jaar.
Stein, W. E., Mannolini, F., VanAller Hernick, L., Landing, E., en Berry, C. M. (2007). “Giant cladoxylopsid trees resolve the enigma of the Earth’s earliest forest stumps at Gilboa.” Nature 446, 904–907.Waarom lezen? De eerste bomen van acht meter, 385 miljoen jaar geleden: 88 miljoen jaar na de eerste sporen.
Morris, J. L., e.a. (2018). “The timescale of early land plant evolution.” PNAS 115, E2274–E2283.Waarom lezen? De moleculaire-klokmarges voor groene planten, landplanten en vaatplanten, naast de fossiele dateringen gezet.
De energiewereld wil een nieuw klimaatfonds. Het oude fonds van €35 miljard is leeg. Na 2030 raken ook de andere potten op. Het Nationaal Plan Energiesysteem rekent met €40 tot €60 miljard aan investeringen per jaar. Dat is twee tot drie keer zoveel als nu. Duitsland heeft een fonds van €500 miljard. De sector is jaloers.
Het PBL laat zien waarom de druk zo hoog is. Nederland komt in 2040 uit op 65% minder uitstoot dan in 1990. In het gunstigste geval 72%. Het doel is 90%.
Haagse bronnen zeggen dat een nieuw fonds politiek onhaalbaar is. Er zijn andere prioriteiten: defensie, woningbouw, sociale zekerheid, box 3.
Het debat gaat dus over één vraag: hoeveel geld komt er? Dat is de verkeerde vraag. De cijfers laten zien waarom.
Het beeld
Het energiesysteem is te lezen als een visnet. Er zijn draden, knopen en mazen. Een knoop houdt alleen als hij op zichzelf sluit. Een knoop die alleen blijft zitten omdat iemand hem vasthoudt, valt uit elkaar zodra die hand loslaat.
Een fonds is zo’n hand. Het houdt van buitenaf open wat niet zelf sluit. De vraag is daarom niet hoe groot de hand moet zijn. De vraag is welke knopen zelf houden.
Wat valt weg als de toevoer stopt
De subsidieregeling SDE laat het patroon goed zien.
Wat geen brandstof nodig heeft, blijft steeds vaker zelf staan. Volgens de evaluatie van de SDE++ is een groeiend deel van de zon- en windprojecten op land rendabel zonder subsidie.
Wat brandstof nodig heeft, valt weg. Voor biomassa bestaan aparte regelingen om installaties na afloop van hun subsidie overeind te houden. De biomassafabriek van Primco in Duiven had recht op maximaal €65 miljoen over twaalf jaar. Toen de geldstroom haperde, lag het werk stil.
Veel van wat wordt toegezegd, komt er nooit. Van de SDE++-rondes 2020–2022 was na twee jaar 14% gerealiseerd. Voor €756 miljoen werden beschikkingen ingetrokken. In 2022 leverden de projecten samen 23 kiloton CO₂-reductie op. Bij de rondes van 2019 en voorjaar 2020 werd meer dan de helft van de projecten nooit gebouwd. In 2021 werden aanvragen voor CO₂-opslag van €6,1 miljard ingetrokken.
Het patroon is eenduidig. Wat sluit, blijft. Wat wordt opengehouden, verdwijnt zodra het geld stopt. Een nieuw fonds verlengt het openhouden. Het maakt niets gesloten.
Wat de daling van de uitstoot echt deed
De voorzitter van de NVDE zegt in het FD dat Nederland een gat van 25% al eerder heeft gedicht. In 2016 voorspelde het PBL 24% reductie voor 2030. Nu komen we rond 50% uit.
De cijfers van het CBS vertellen een ander verhaal.
2021: de uitstoot steeg. Het begin van het jaar was koud en gas werd duurder, waardoor centrales meer kolen stookten.
2022: −9%. Vooral door de hoge gasprijs na de Russische inval en een zachte winter.
2023: −6%. Meer stroom uit wind en zon, minder uit kolen en gas.
2024: de daling vlakt af. De industrie stootte iets meer uit.
De grootste stap kwam van buiten. Van een prijsschok en van het weer. Het PBL zei al in 2019 dat de kolen- en gasprijs wereldwijd worden bepaald en de CO₂-prijs in Europa.
Ook het Klimaatakkoord van 2019 was minder breed dan het lijkt. Greenpeace, Milieudefensie, Natuur & Milieu, FNV en MVO Nederland tekenden niet. Hun bezwaar: de industrie werd niet gehouden aan haar beloofde 19,4 megaton, en de deur bleef open voor CO₂-opslag.
Het net zelf is de grootste rekening
Wat in het FD-artikel ontbreekt, is het net.
De wachtrij van bedrijven bij de regionale netbeheerders groeide eind 2025 tot ruim 15.000 aanvragen. Een nieuwe aansluiting kostte gemiddeld 45 weken wachten. De maatschappelijke schade wordt geschat op €10 tot €40 miljard per jaar. De netbeheerders investeren nu €8,5 miljard per jaar. Richting 2040 kan netuitbreiding oplopen tot €270 miljard.
Dat betekent: het vastgelopen net kost per jaar ongeveer evenveel als het hele Klimaatfonds in zijn bestaan. Wie een nieuw fonds vult zonder het net te veranderen, vult een emmer met een gat.
Een systeem uit het stoomtijdperk
Het elektriciteitsnet is ontworpen rond de stoomturbine. Grote centrales draaien op 50 hertz. De stroom gaat via vele transformaties naar het stopcontact.
Een gascentrale zet 55 tot 60% van de brandstof om in stroom. Een kolencentrale ongeveer 40%. De rest verdwijnt als warmte. Bij elke transformatiestap gaat nog eens ongeveer 1,5% verloren. Een studie van KEMA voor de NMa kwam uit op 5.400 gigawattuur netverlies per jaar, ter waarde van €400 miljoen.
Reken de keten door. Bij gas komt 0,58 × 0,95 ≈ 0,55 van de brandstof aan. Bij kolen 0,40 × 0,95 ≈ 0,38. Bijna de helft gaat verloren voordat er iets gebeurt.
Wind en zon passen slecht in dat ontwerp. Ze leveren in pieken. Het net is gebouwd voor gelijkmatige centrale levering. Daar komt de congestie vandaan.
Windparken en water
Er is een tweede rekening die niemand opmaakt: water.
Gemeten op land. In Chinese graslanden daalde het bodemvocht binnen windparken met 4,4%. Benedenwinds was de afname gemiddeld tot 2,85%, bovenwinds 0,21%. In Texas warmde de grond boven vier grote windparken met ruim 2.350 turbines op met 0,72 °C per decennium, vooral ‘s nachts. Warmere grond verdampt meer water.
Berekend op zee. Het Helmholtz-Zentrum Hereon rekende in 2021 dat parken op de Noordzee elkaar de wind afnemen. Hoe groter en dichter op elkaar, hoe meer de opbrengst daalt: met 20% of meer bij de parken benedenwinds. In 2022 vond dezelfde groep dat de wind achter de parken ongeveer 1 meter per seconde afneemt. Boven de parken valt 5% meer regen.
In augustus 2026 volgde de volgende stap. Bij grootschalige uitbreiding regent de lucht uit boven zee. Aan de kust komt minder aan. Voor delen van Denemarken, Duitsland, Nederland en het Verenigd Koninkrijk rekent het model tot 15% minder neerslag aan de kust. Het gaat om een bewust extreem scenario, boven de 300 gigawatt die voor 2050 wordt besproken. Een andere modelstudie voor Midden-Europa vond op land juist een zeer klein effect.
De richting is wel duidelijk. Op de Noordzee wordt de grootste bouwopgave van Europa gepland. Langs de kust wonen en werken de mensen die het water nodig hebben: boeren in Noord-Holland, Friesland en Zeeland, drinkwaterbedrijven, natuurbeheerders in de duinen. Nederland worstelt al met droogte. In het FD-artikel, het Klimaatfonds en het pleidooi voor een nieuw fonds komt water niet voor.
Iedereen in de pas
Waarom ziet niemand dit? Het antwoord ligt in hoe de sector beweegt. Iedereen loopt dezelfde kant op, in dezelfde maat. In de natuurkunde heet dat fase-vergrendeling: losse ritmes die in elkaars pas vallen tot ze één ritme zijn.
Op de Noordzee is dat letterlijk te zien. Alle landen bouwen op dezelfde plek. De parken nemen elkaar de wind af. De zwerm verlaagt zijn eigen opbrengst.
In de subsidie ook. De gerealiseerde projecten uit de SDE++-rondes 2020–2022 waren vrijwel uitsluitend zonnepanelen.
En in wat níet wordt gedaan. Voor vroege, nieuwe technologie lag bij het PBL €4,2 miljard aan aanvragen op tafel. Er was nog €228 miljoen vrije ruimte. Van de €13,2 miljard die vrij is in het Klimaatfonds, zit €13 miljard bij kernenergie.
Wat buiten beeld blijft
Er bestaat een andere route. Ze staat in een vakblad, niet in een fonds.
In 2021 bouwde de groep van Garret Moddel in Boulder een laagje van enkele nanometers met een spiegel ervoor. Zonder aangelegde spanning leverde het stroom. Er gingen acht controles op meetfouten overheen, meer dan duizend exemplaren werden gemaakt en de opbrengst werd zes maanden gevolgd. De gemeten waarde: 1,4 picowatt op 0,02 vierkante micrometer. Dat is 70 watt per vierkante meter per laag.
In de VS haalde het bedrijf Casimir Inc. in mei 2026 $12 miljoen op voor een chip op dezelfde basis. Die moet in 2028 op de markt komen. De chip is ontworpen voor 1,5 volt bij 25 microampère op 5 bij 5 millimeter.
Reken het om:
Vermogen per m²
Per m² per jaar
Zonnepaneel in Nederland
~23 W gemiddeld
~200 kWh
Casimir-chip (37,5 µW op 25 mm²)
1,5 W, dag en nacht
~13 kWh
Moddel, één laag
70 W, dag en nacht
~613 kWh
Bij 70 watt per vierkante meter levert één laag ongeveer drie keer zoveel als een zonnepaneel. Zonder net. Zonder ruimtebeslag. Zonder effect op wind of regen. Een huishouden met 3.000 kilowattuur per jaar gebruikt gemiddeld 342 watt. Dat is 4,9 vierkante meter van één laag.
In het beeld van het visnet is dit een maas die doorlaat. Hoe kleiner de maas, hoe meer er doorgaat. Er wordt niets verbrand en niets vervoerd. De voorziening hoort bij het huis, niet bij een centrale.
Nederland heeft groepen die drie van de vier benodigde stappen al beheersen: in Groningen, Twente, Eindhoven, Amsterdam en Delft. De vierde stap, hoeveel van wat de maas sluit als stroom naar buiten komt, wordt nergens in Nederland onderzocht.
Wat er dan wel moet
Geen groter fonds. Een andere vraag.
Tel wat sluit. Financier alleen wat na afloop van de subsidie zelf blijft staan. Biomassa en CO₂-opslag vallen daar buiten.
Zet het net op de rekening. €10 tot €40 miljard per jaar aan congestieschade hoort in elke afweging over nieuwe productie.
Zet water op de rekening. Laat de Noordzeeplannen doorrekenen op kustneerslag en grondwater, voordat er verder wordt gebouwd.
Doe de ene meting. Eén laboratorium met een dunnefilmlijn kan de vierde stap meten. Dat kost een fractie van één windpark.
Wat vaststaat en wat open is
Gemeten: de lage realisatie en de intrekkingen in de SDE, de rol van gasprijs en weer in de daling van de uitstoot, de netcongestie en haar kosten, het verdrogen van bodem in Chinese windparken, de nachtelijke opwarming in Texas, de stroom uit het laagje van Moddel.
Berekend: de windverliezen en neerslageffecten op de Noordzee. Dat zijn modelresultaten, het sterkste effect in een extreem scenario.
Open: of de stroom van Moddel standhoudt bij herhaling elders. Vijf jaar na publicatie is er geen onafhankelijke herhaling. Een theoretische analyse uit 2024 verwerpt het voorgestelde mechanisme, maar laat de meting staan. Of lagen optellen is onbekend.
Juist dat open punt is goedkoop te beslissen. Het kost minder dan één jaar aan netverlies.
Geannoteerde referenties
De aanleiding en het beleid
Rooijers, E. en Selderbeek, E. (2026). “Energiewereld roept om nieuw klimaatfonds, maar zo’n miljardenpot lijkt politiek onhaalbaar.” Het Financieele Dagblad, 12 oktober 2026.Waarom lezen? De stand van het debat: het lege fonds, de €40–60 miljard per jaar, de jaloezie op Duitsland en de spagaat van de minister. Leesadvies: lees het met de vraag wat er níet in staat: het net en het water.
Ministerie van KGG (2024). Tussentijdse evaluatie SDE++ (2020–2022).https://zoek.officielebekendmakingen.nl/blg-1131349.pdfWaarom lezen? De harde cijfers over realisatie: 14% gebouwd, €756 miljoen ingetrokken, 23 kiloton reductie in 2022. Leesadvies: de samenvatting volstaat; de passage over projecten die zonder subsidie rendabel zijn is de kern.
KEMA Nederland (2011). Verdeling van de kosten van netverliezen. Rapport voor de NMa.https://berthub.eu/tkconv/getraw/2011D26421Waarom lezen? De 5.400 gigawattuur en €400 miljoen netverlies per jaar, technisch onderbouwd. Leesadvies: hoofdstuk 2.
Klimaatstichting HIER (2026). Lokale Energie Monitor 2025.https://www.hier.nu/lokale-energie-monitor-2025Waarom lezen? 696 coöperaties, 143.000 leden, 9.000 vrijwilligers, en de terugval in 2025. Het gezicht van de energietransitie van onderop, en waar die vastloopt.
Windparken, bodem en regen
Wang, G. e.a. (2023). “Wind farms dry surface soil in temporal and spatial variation.” Science of the Total Environment 857, 159293.https://pubmed.ncbi.nlm.nih.gov/36209881/Waarom lezen? Gemeten verdroging van de bodem binnen en rond windparken, per seizoen en windrichting. Leesadvies: de samenvatting en de figuren met bovenwinds tegen benedenwinds.
Zhou, L. e.a. (2012). “Impacts of wind farms on land surface temperature.” Nature Climate Change 2, 539–543.https://tethys.pnnl.gov/publications/impacts-wind-farms-land-surface-temperatureWaarom lezen? De eerste satellietmeting van opwarming boven windparken: 0,72 °C per decennium, vooral ‘s nachts. Kort artikel, goed leesbaar.
Akhtar, N., Geyer, B., Rockel, B. e.a. (2021). “Accelerating deployment of offshore wind energy alter wind climate and reduce future power generation potentials.” Scientific Reports 11.https://doaj.org/article/c8bb766548d14226af1c3f403ff1f97aWaarom lezen? Hoe parken op de Noordzee elkaar de wind afnemen, tot 20% of meer. De onderzoekers noemen wind op de Noordzee een beperkte bron.
Akhtar, N., Geyer, B. en Schrum, C. (2022). “Impacts of accelerating deployment of offshore windfarms on near-surface climate.” Scientific Reports 12.https://marineinfo.org/id/publication/359580Waarom lezen? Minder wind achter de parken, meer regen erboven, en een windeffect groter dan dat van klimaatverandering zelf.
Akhtar, N., Elizalde, A., Geyer, B. en Schrum, C. (2026). “Projected impacts of future offshore wind farms on coastal precipitation over the Northwest European shelf.” Communications Earth & Environment.https://nachrichten.idw-online.de/2026/08/20/offshore-wind-farms-could-alter-precipitation-patternsWaarom lezen? Tot 15% minder kustneerslag voor onder meer Nederland in een extreem uitbreidingsscenario. Leesadvies: begin met het persbericht van Hereon, daarna de figuren met de kustregio’s.
Hewage e.a. (2026). “Impact of Wind Turbines on Near-Surface Microclimate and Soil Moisture Under Different Stability Regimes.” Boundary-Layer Meteorology 192:50.https://tethys.pnnl.gov/sites/default/files/publications/Hewage_et_al_2026.pdfWaarom lezen? Het tegengeluid: een model voor Midden-Europa met een zeer klein effect op bodemvocht. Lees het naast Wang om te zien hoe gemeten en gemodelleerd uiteenlopen.
De andere route
Moddel, G., Weerakkody, A., Doroski, D. en Bartusiak, D. (2021). “Optical-Cavity-Induced Current.” Symmetry 13, 517.https://www.mdpi.com/2073-8994/13/3/517Waarom lezen? De enige gemeten structuur van haar soort, met acht controles op meetfouten. Leesadvies: tabel 1 voor de opbouw, figuur 4 voor de metingen, paragraaf 3.2 voor de controles.
Maclay, G. J. (2024). “Modeling Electronic Devices with a Casimir Cavity.” Physics 6, 1124–1131.Waarom lezen? De enige theoretische analyse van het apparaat. Verwerpt het voorgestelde mechanisme, laat de meting staan. Laat zien waarom het open punt open is.
Konstapel, J. (2026). “Het Net.” constable.blog, 2 augustus 2026.https://constable.blog/2026/08/02/the-net/Waarom lezen? Het visnet als beeld van het vacuüm: draad, knoop, maas en de regel dat alleen blijft wat sluit.
Konstapel, J. (2026). “Three Roads to a Dynamic Vacuum.” constable.blog, 6 oktober 2026.https://constable.blog/2026/10/06/vacuum-netmoddelcasimirinc/Waarom lezen? De drie onafhankelijke lijnen naast elkaar, en de Nederlandse groepen die drie van de vier stappen al beheersen. Leesadvies: paragraaf 6 en 9.
The perovskite crystal is one of the most versatile structures in materials science. It powers solar cells, ferroelectrics, superconductors and catalysts. This essay examines that structure through the closure rule of the Vacuum.Net theory. Four findings stand out.
First, the perovskite unit cell is an occupation of the trit cube. The chemical formula ABX₃ follows from occupying three of the cube’s four grades. Second, barium titanate passes through all four grades as it cools. The number of variants in each phase matches the number of cells per grade: 1, 6, 12 and 8. Third, the closure rule carries the point symmetry −3m. Perovskites reach exactly this symmetry in their deepest rhombohedral states.
Fourth, and most practical: a ferroelectric perovskite can act as the open wall of a mesh. Its polarisation has three states — positive, zero and negative. This gives a single junction three settings of its asymmetry. The only measured device of this kind has lacked exactly that.
Along the way, the essay corrects one step in an earlier reading of that device. The published curve shows current, not power. At the widest gap, the difference is a factor of more than ten.
2. The net in brief
Picture a fishnet stretched in water. One strand runs everywhere and crosses itself. The strands are under tension. Where strands wind around each other, knots form. These knots are matter. Between the strands lie meshes: openings with a shape and a size. A winding persists only if it returns in phase inside its mesh.
At each crossing, the strand holds one of three values: +1, 0 or −1. This unit is the trit. Three trits form a window. A window is closed when its three values sum to zero, counted modulo 3. An open window carries tension. Only what closes, stays.
Three values in three places give 27 windows. Nine of them close. Three have equal values: (+,+,+), (0,0,0) and (−,−,−). Six have three different values: the six orders of (−, 0, +). The remaining 18 are open.
3. The trit cube
Give each axis of a cube one trit. A value of 0 means the axis plays no part. A value of +1 or −1 means it takes part, with that sign. Each cell (x, y, z) then has a grade: the number of axes that do not read 0.
Grade
Position in the cube
Cells
0
centre
1
1
on the axes
6
2
in the median planes
12
3
at the corners
8
The total is 1 + 6 + 12 + 8 = 27 = 3³.
Apply the closure rule to the cells. A cell closes when x + y + z ≡ 0 (mod 3). Nine cells close. The centre (0,0,0) is one of them. Six plane cells with opposite signs, such as (1, −1, 0), form a hexagon around the centre. The two corners (1,1,1) and (−1,−1,−1) complete the set.
All six axis cells are open; their sum is ±1. Six of the eight corners are open. Six of the twelve plane cells are open.
4. The perovskite cell as an occupation of the trit cube
The ideal perovskite has the formula ABX₃. Its prototype is CaTiO₃. Place the B ion at the origin and take half the cell edge as the unit. The ions then sit on the cells of the trit cube.
Grade
Cells
Occupant
Share per cell
Count
Closed / open
0
1
B
1
1
1 closed
1
6
X
1/2 (shared by 2 cells)
3
6 open
2
12
empty
—
—
6 closed, 6 open
3
8
A
1/8 (shared by 8 cells)
1
2 closed, 6 open
The count gives B₁X₃A₁, which is ABX₃. The perovskite formula is the occupation of grades 0, 1 and 3 of the cube.
Three features follow at once:
The anion cage stands entirely on open cells. All six X positions have a sum of ±1. The octahedron carries the tension of the cell.
The rest sits in the middle. B occupies the only closed cell inside the cage.
The closed hexagon is unoccupied. In the algebra of the cube, the hexagon is the register of rotations: the bivectors, or the quaternion disc. Section 8 returns to this.
5. The fit of the occupant
The A ion sits in the opening between eight octahedra. In the terms of the net, A is the occupant of a mesh. Goldschmidt’s tolerance factor measures how well it fits:
t = (r_A + r_X) / (√2 · (r_B + r_X))
The ideal value is 1. Three titanates illustrate the range, using Shannon radii (O²⁻ 1.40 Å; Ti⁴⁺ 0.605 Å; A in twelvefold coordination):
SrTiO₃: t = 1.00. The cell is cubic.
BaTiO₃: t = 1.06. The occupant is too large. The B ion has room and leaves the centre. The crystal becomes ferroelectric.
CaTiO₃: t = 0.966. The occupant is too small. The octahedra rotate to narrow the mesh. Rotation typically begins near t = 0.97.
The sign of (t − 1) acts as a trit. A positive value means the centre gives up its closure. Zero means rest. A negative value means the rotation register absorbs the misfit. All three outcomes are measured. Their reading as one trit is an interpretation.
6. Barium titanate walks the grades
Barium titanate passes three transitions on cooling. It is cubic above 403 K, tetragonal down to 278 K, orthorhombic down to 183 K, and rhombohedral below that. The polarisation in the three polar phases points along [001], [011] and [111].
Count the symmetry-equivalent states per phase:
Phase
Polarisation
Variants
Grade of the cube
Cells
cubic
none
1
0, centre
1
tetragonal
⟨001⟩
6
1, axes
6
orthorhombic
⟨011⟩
12
2, planes
12
rhombohedral
⟨111⟩
8
3, corners
8
The sum is again 27. On cooling, the titanium ion walks from grade 0 through grades 1 and 2 to grade 3. The number of variants per phase equals the number of cells per grade.
Atomistic simulation adds one detail. Locally, the titanium is displaced towards a corner in every phase. The higher phases are averages: over 8 corners for cubic, 4 for tetragonal, 2 for orthorhombic. The rhombohedral phase is the ground state, with three polar components in every cell.
7. The symmetry of the closure rule
Which symmetries of the cube keep the nine closed cells closed? Any permutation of the coordinates keeps x + y + z unchanged: 6 elements. Reversing all signs turns a sum of 0 into 0, and ±3 into ±3 — both closed: 2 elements. Together these give 12 elements. A single sign change breaks closure: (1, −1, 0) becomes (−1, −1, 0), with sum −2. No larger group preserves the closed set.
This group of order 12, built from coordinate permutations and inversion, is the point group −3m. The full symmetry group of the cube has 48 elements. The index is 48 / 12 = 4: there are four ways to choose which body diagonal is the axis of closure.
The closure rule therefore carries the trigonal symmetry of one body diagonal. Perovskites reach exactly this symmetry in their deepest states:
the rhombohedral ground state of BaTiO₃, space group R3m, whose point group 3m is the polar subgroup of −3m;
the a⁻a⁻a⁻ rotation pattern of LaAlO₃, space group R-3c, with point group −3m.
The count is exact. That the coincidence is more than a coincidence is a reading.
8. The rotations, and a negative result
When t < 1, the octahedra rotate. Glazer’s notation gives each axis a⁺ (in-phase rotation from layer to layer), a⁻ (out of phase) or a⁰ (none). This alphabet per axis is itself a trit.
The raw sign patterns number 27. Up to a permutation of axes, there are 10 sign types. Magnitudes split these further. Glazer listed 23 tilt systems. Group theory by Howard and Stokes reduced them to 15 distinct space groups.
A natural test is to apply the closure rule to the signs. The closed patterns are a⁰a⁰a⁰, a⁺a⁺a⁺, a⁻a⁻a⁻ and a⁰b⁺c⁻. The most common real pattern, a⁻a⁻c⁺ (the GdFeO₃ type, Pnma), has a sum of −1 and is open.
The closure rule on tilt signs does not predict which rotation patterns occur most often. The test was built for this essay, and it is not part of the theory. “In phase” and “out of phase” describe repetition along an axis, not the value at a crossing. The negative result stands.
9. Where perovskites sit on the two roads
The Vacuum.Net theory names two ways in which the net lets energy pass.
The mesh passes. An opening with a size lets winding through. The passage depends on the size.
The knot opens. One closure of a knot gives way, and a winding leaves. Stars open knots; that is what their light is.
A perovskite solar cell lives on the second road, at a distance. It catches windings that the sun has released. It needs light.
A device on the first road needs no light. Its question is how much of what the mesh closes leaves as current.
Perovskites can serve the first road too — not as the cell that catches light, but as the wall of the mesh. Before that, one point about the record of the first road needs setting straight.
10. What has been measured on the open side
10.1 The device
Moddel, Weerakkody, Doroski and Bartusiak (2021) built metal–insulator–metal tunnel junctions with an optical cavity on one side. For the cavity series, the stack was:
nickel base electrode, 38 nm;
native nickel oxide, about 1 nm, plus aluminium oxide, 0.9 nm;
palladium upper electrode, 8.3 nm;
PMMA cavity, 33, 79, 230 or 1100 nm;
aluminium mirror, 150 nm.
The active area was 0.02 ± 0.006 µm². At zero applied voltage, the junctions delivered current. For linear current–voltage curves, the maximum power is |I_sc · V_oc| / 4. For the 33 nm cavity this was 1.4 pW. Thinner cavities gave more current.
The group ran eight tests for artefacts. These covered stability over four hours, scaling with area, scaling in 4×4 arrays, the stage of processing at which current appears, leakage through the cavity, electromagnetic pickup, and two thermoelectric checks. All came out negative. The trends held in over 1000 devices from 21 batches. After six months, output had fallen by less than 10% with PMMA and less than 20% with SiO₂.
The authors call their mechanism conjectural. They expected the energy density of suppressed modes to scale as 1/d³, did not see it, and did not attempt a quantitative fit.
10.2 The numbers per area
1.4 × 10⁻¹² W over 2 × 10⁻¹⁴ m² gives 70 W/m² for one layer at 33 nm. One natural mesh measures (45.6 nm)² = 2.08 × 10⁻¹⁵ m². At 70 W/m², that is 1.46 × 10⁻¹³ W per mesh.
10.3 What has happened since
Maclay (2024) modelled the device using current calculations of vacuum fluctuations in a cavity. Two results bear on it. An ideal cavity raises the mean energy of electrons near the electrode, which would drive current in the opposite direction to the one observed. And that mean energy is nearly independent of cavity size, which contradicts the measured dependence. Maclay concludes that the proposed mechanism does not hold. He does not dispute the measurement.
A 2026 funding proposal for a first independent test reports that, five years on, there have been no independent replications, no public raw data and no experimental critiques. It also names an untested pathway. Palladium on aluminium oxide is the standard construction of a hydrogen sensor. The original artefact list did not test hydrogen.
The record of the open side is therefore one group, one structure, a mechanism its own authors call conjectural, and a theoretical analysis that rejects that mechanism while leaving the data standing.
11. Current, not power: a correction
An earlier reading of the device (Konstapel, 2026b) derived a rule for the mesh. Passage equals one over the open size, 1/L, so the slope on a log-log plot should be −1. The test used Moddel’s Figure 4b and read a slope of −0.9 ± 0.15. The interval contains −1.
Figure 4b, however, plots short-circuit current against cavity thickness — not power.
For a linear junction with resistance R, the open-circuit voltage is V_oc = I_sc · R. The maximum power is then P = I_sc · V_oc / 4 = I_sc² · R / 4. If R does not change with cavity thickness, power scales as the square of current. A current slope of −0.9 becomes a power slope of −1.8.
The effect on the numbers, anchored at 70 W/m² for 33 nm:
Gap
Ratio 33/L
Power if P = 1/L
Power if P = 1/L^1.8
Power if P = 1/L²
33 nm
1
70 W/m²
70 W/m²
70 W/m²
45.6 nm
0.7237
50.7 W/m²
39.1 W/m²
36.7 W/m²
1100 nm
0.0300
2.10 W/m²
0.127 W/m²
0.063 W/m²
At the widest gap, the readings differ by a factor of 2.10 / 0.127 ≈ 16.5.
Two ways forward are consistent:
The rule concerns current. Then the slope test holds as published, and power runs as 1/L². Passage is charge per unit time through the mesh.
The rule concerns power. Then the measured current must run as 1/√L, a slope of −0.5. That sits outside −0.9 ± 0.15, and the test fails.
The first is the natural reading. Larson’s ladder gives power as 1/S, but what crosses the junction is charge. The voltage across a fixed resistance adds a second factor of 1/L. On that reading, the earlier figure of about 2 W/m² at 1100 nm becomes about 0.06–0.13 W/m².
Nothing at 33 nm changes. The household estimates built on 70 W/m² at 33 nm stand as they were.
The question is not closed. Resistance varies widely between nominally equal junctions, and the published values are averages over chips. Digitising Figure 4a, which gives the full current–voltage line per thickness, would yield V_oc per thickness and settle the power slope directly.
12. A wall with three states
12.1 What already exists
A ferroelectric tunnel junction is two electrodes separated by a ferroelectric barrier a few unit cells thick. Reversing the polarisation reverses the charge at the two interfaces. This changes the shape of the barrier and with it the tunnelling current. On nanometre-thick BaTiO₃ films, resistance changes by about two orders of magnitude on reversal, on a lateral scale of 20 nm, at room temperature. Asymmetric interfaces are a known requirement for the effect.
These junctions are studied as memories. No published work was found in which a ferroelectric tunnel junction is placed beside a Casimir cavity and measured at zero bias.
12.2 The proposal
Replace the passive insulator of the cavity device with a ferroelectric barrier of the same thickness.
Thickness. Moddel’s insulator totals about 1.9 nm. Five pseudocubic cells of BaTiO₃ at about 0.40 nm make 2.0 nm. The tunnelling regime is preserved.
Electrodes. Use an oxide bottom electrode such as SrRuO₃, standard in these junctions, and a thin upper metal that does not take up hydrogen. With no palladium, the hydrogen pathway is closed by construction.
Cavity. Keep a dielectric spacer and a mirror on the side of the upper electrode, as in the original device.
The polarisation of the barrier has three states. P⁺ points towards the cavity. P⁻ points away. P⁰ has no net polarisation; it can be reached by heating the barrier above its Curie point, or by a twin junction from the same growth run with a paraelectric SrTiO₃ barrier of equal thickness.
In the terms of the net, the polar barrier is the open wall, and its trit sets how the wall returns what the mesh closes. The three states change the asymmetry of one junction without touching its geometry.
12.3 What the design discriminates
Published comparisons are between different devices, where fabrication spread dominates. This design compares one junction with itself.
pyroelectric current needs a temperature change; it vanishes
After a polarisation switch
steady value after the switch
depolarisation current decays
In darkness
unchanged
ferroelectric photovoltaic effect
Current against time at fixed bias; frequency dependence
flat
ionic motion of vacancies shows slow drift and low-frequency response
Exposure to hydrogen
unchanged
hydrogen uptake in the electrode
The prediction is fixed before the run. With the mirror present, I_sc(P⁺) ≠ I_sc(P⁻). Without the mirror, both are zero. The theory gives no value for the size of the difference. That size is the open number: the fraction of what the mesh closes that leaves as current.
12.4 The size of the mesh in whole cells
Oxide films grow layer by layer with unit-cell control by pulsed laser deposition, monitored by electron diffraction during growth. In SrTiO₃, one cell measures 0.3905 nm.
Size
Unit cells of SrTiO₃
33 nm (Moddel’s thinnest cavity)
84.5
45.6 nm (the natural unit)
116.8
60–75 nm (a resting gap)
153.6–192.1
An epitaxial spacer adjusts the mesh in steps of 0.39 nm — about 1% of the natural unit. The region around 45.6 nm can therefore be traced cell by cell.
A spacer of high refractive index also separates two readings of the gap. PMMA (n = 1.52) and SiO₂ (n = 1.49), the two spacers Moddel used, differ by only 2%. If the current follows the geometric gap, as the net’s 1/S says, then spacers of very different index at equal thickness should give equal current. If it follows the optical cutoff, which scales with n·L, they should not. A spacer of high index must be transparent across the relevant band; that choice is part of the design.
12.5 Where in the Netherlands
The Dutch groups mapped earlier (Konstapel, 2026d) cover Casimir switching, tunnable cavities, critical fluctuations and membranes. One group is missing from that map: the University of Twente group that grows perovskite oxides with unit-cell control by pulsed laser deposition. In 2025, that group completed a doctoral thesis on ferroelectric gradient films and published a study of barium titanate gradient films. A gradient film has two unlike sides by definition. Together with Groningen, the same group reported room-temperature epitaxy of the halide perovskite α-CH₃NH₃PbI₃.
The two capabilities missing on the open side — mesh size by the cell and a built-in, switchable asymmetry — sit in one laboratory.
13. Why not halide perovskites as filling
Halide perovskites conduct ions as well as electrons. First-principles barriers in CH₃NH₃PbI₃ (Eames et al., 2015) are 0.58 eV for iodide vacancies, 0.84 eV for methylammonium and 2.31 eV for lead. The iodide value agrees with kinetics measured on a working cell.
The order matches the trit cube. X stands on cells that are all open (6 of 6). A stands on corners of which 6 of 8 are open. B stands on the closed centre. With three items, a random order matches with a probability of 1 in 6. This is a correspondence, not a proof.
At 300 K, kT = 0.02585 eV:
iodide against methylammonium: e^(0.26/0.02585) ≈ 23,000 times more hops;
iodide against lead: e^(1.73/0.02585) ≈ 10²⁹.
Slow ionic currents and hysteresis can imitate a steady signal. A halide perovskite as the cavity filling or barrier would bury the effect sought. Oxides carry vacancies on their X sites too. This is why the ionic test belongs in the table of Section 12.3.
14. What it means in practice
If the open side yields a number, the arithmetic is short. At 70 W/m² for one layer at 33 nm, a household using 3,000 kWh of electricity a year draws an average of 3,000 / 8,766 h = 0.342 kW. That needs 342 / 70 = 4.9 m² of one layer, or 0.049 m² of a hundred layers if layers add.
Each of those conditions is open: that the current is real, that layers add, and that the coupling survives in other materials. A structure that passes continuously, in any place, would sit in the house rather than at the end of a grid. Whether that happens depends on one fraction, measurable in one junction with three states.
15. Status
Counted here
ABX₃ as the occupation of grades 0, 1 and 3 of the trit cube.
The anion cage entirely on open cells.
The grade counts 1/6/12/8 set beside the variant counts of BaTiO₃.
The closure group −3m of order 12, index 4.
The power slope of −1.8 from a current slope of −0.9, and the table of Section 11.
Unit-cell counts of 84.5, 116.8 and 153.6–192.1.
Hop ratios of 23,000 and 10²⁹.
Measured, from sources
Tolerance factors and their three outcomes.
The phase sequence of BaTiO₃ with 6/12/8 variants.
Larson, D. B. (1959). The Structure of the Physical Universe. Portland: North Pacific Publishers. Why read? The ladder of space and time on which power is 1/S. Reading advice: the chapters on the natural units only.
The device and its record
Moddel, G., Weerakkody, A., Doroski, D., and Bartusiak, D. (2021). “Optical-Cavity-Induced Current.” Symmetry 13, 517. https://www.mdpi.com/2073-8994/13/3/517 Why read? The only measured device of its kind, with its stack, eight artefact tests and its own caution about mechanism. Reading advice: Table 1 for every layer; Figure 4a, not only 4b, for the power slope; Section 3.2 for the tests.
Maclay, G. J. (2024). “Modeling Electronic Devices with a Casimir Cavity.” Physics 6, 1124–1131. Why read? The one theoretical analysis of the device. It rejects the proposed mechanism and leaves the data standing. It shows why the open side still lacks a model.
“Electric Field and Voltage Fluctuations in the Casimir Effect” (2021). arXiv:2112.14285. https://arxiv.org/abs/2112.14285 Why read? An estimate of whether cavity-modified fluctuations are large enough to matter for the junction current.
Casimir, H. B. G. (1948). “On the attraction between two perfectly conducting plates.” Proceedings KNAW 51, 793. Why read? The two-wall opening in its original form. Three pages.
Perovskite structure
Goldschmidt, V. M. (1926). “Die Gesetze der Krystallochemie.” Die Naturwissenschaften 14, 477–485. Why read? The origin of the tolerance factor: the fit of an occupant to its opening.
Shannon, R. D. (1976). “Revised effective ionic radii.” Acta Crystallographica A 32, 751–767. Why read? The radii used in Section 5. The tolerance factor shifts by about ±0.03 between radius sets; this table fixes one set.
Glazer, A. M. (1972). “The classification of tilted octahedra in perovskites.” Acta Crystallographica B 28, 3384–3392. Why read? The trit alphabet of rotations — a⁺, a⁻ and a⁰ — and the 23 tilt systems.
Howard, C. J., and Stokes, H. T. (1998). “Group-theoretical analysis of octahedral tilting in perovskites.” Acta Crystallographica B 54, 782–789. Why read? The reduction to 15 space groups and their hierarchy. Section 8 tests the closure rule against this list.
“Atomistic Description for Temperature-Driven Phase Transitions in BaTiO₃” (2016), arXiv:1601.02031; and “Thermodynamics and dielectric response of BaTiO₃ by data-driven modeling” (2021), arXiv:2111.05129. https://arxiv.org/abs/1601.02031 · https://arxiv.org/abs/2111.05129 Why read? The 6, 12 and 8 equivalent minima of the polar phases, and the local displacement towards a corner in every phase. Section 6 rests on these counts.
Ferroelectric junctions
Zhuravlev, M. Y., Sabirianov, R. F., Jaswal, S. S., and Tsymbal, E. Y. (2005). “Giant electroresistance in ferroelectric tunnel junctions.” Physical Review Letters 94, 246802. Why read? The prediction that polarisation reversal changes tunnelling through a ferroelectric barrier.
Velev, J. P., et al. (2007). “Effect of ferroelectricity on electron transport in Pt/BaTiO₃/Pt tunnel junctions.” Physical Review Letters 98, 137201. Why read? The first-principles mechanism in the material proposed here.
Gruverman, A., et al. (2009). “Tunneling electroresistance effect in ferroelectric tunnel junctions at the nanoscale.” Nano Letters 9, 3539–3543. Why read? The measured change of about two orders of magnitude on 20 nm, at room temperature, in thin BaTiO₃.
Koster, G., Rijnders, G., Blank, D. H. A., et al. University of Twente, publications on pulsed laser deposition of perovskite oxides. https://people.utwente.nl/g.koster Why read? Unit-cell growth control, ferroelectric gradient films and room-temperature halide perovskite epitaxy: the capabilities Section 12 needs.
Halide perovskites
Eames, C., Frost, J. M., Barnes, P. R. F., O’Regan, B. C., Walsh, A., and Islam, M. S. (2015). “Ionic transport in hybrid lead iodide perovskite solar cells.” Nature Communications 6, 7497. Why read? The barriers of 0.58, 0.84 and 2.31 eV. Section 13 rests on them.
Hoke, E. T., et al. (2015). “Reversible photo-induced trap formation in mixed-halide hybrid perovskites for photovoltaics.” Chemical Science 6, 613–617. Why read? Light sorts the open X sites into iodide-rich and bromide-rich domains, which reverse in the dark. It shows how mobile the open sites are under tension.
The world of October 2026 is under simultaneous strain: a five-year-old war in Ukraine, a seven-month-old war with Iran that has closed the Strait of Hormuz, and a very strong El Niño driving up food prices. Conventional outlooks treat these as separate files. This essay reads them instead as a single net, using the Vacuum.Net theory and four of its instruments: the ladder of three, the credit-loading measure, the food threshold, and the open-mesh test.
The analysis yields one central conclusion. In February 2028, both wars reach a boundary of their time-layer in the same month. Before that date, the world discharges on three windings at once — two wars and an El Niño. After it, the slow layers take over: population, debt, and rearmament. February 2028 is the hinge.
The regional reading is compact. Europe holds, but with open meshes in gas. Russia discharges with its reserve spent. The Middle East discharges, with food prices above the threshold. China loads and holds. The United States loads in public debt and in one technology, while its politics discharge through the ballot.
1. The situation: one net, not separate files
In 2025 the world counted 65 state-based armed conflicts in 35 countries — the highest number since 1946 — of which thirteen were wars. Two of those wars now set the price of energy and bread. The war in Ukraine is in its fifth year. The war with Iran began on 28 February 2026 and has closed the Strait of Hormuz. Brent crude trades at 104 dollars.
The guiding image is a fishing net in water. Every strand stands under tension; where strands wind around each other, knots form. Pull at one knot and the tension travels through the whole net — and it passes where the net is weakest. The net has no front line: load always moves through the most open mesh, meaning a point where the load converges and no second path exists.
Ten rules of the theory govern the reading. The most important for this essay:
Address and tension. Every knot has a slow layer (geography, borders, treaties, population) and a fast layer (armament, prices, public mood).
Four states. Tension is loading, holding, discharging, or at rest.
The open mesh. Load passes through the weakest open mesh — the point of convergence without a second path.
Blocked discharge. A discharge that is blocked is stored, and returns later, larger.
Deep to shallow. The slow layer sets the frame for the fast layer, never the reverse.
The theory makes no price forecast. Loading announces a discharge; it does not announce a price.
2. The instruments
The ladder of three. The measuring rod is the natural unit of time, T₀ = 1.521 × 10⁻¹⁶ s, half the period of the Rydberg frequency. Layer n lasts 3ⁿ × T₀. One year sits on layer 49 (its position is log₃ of 2.075 × 10²³ ≈ 48.87); five years sits at position 50.34. The outlook of this essay therefore spans layers 50 and 51.
The credit measure. The three-year change L in credit to the private non-financial sector (as a percentage of income) gives the probability of a credit crisis within one year through a logistic function. Worked example: China’s credit rose from 187.3 to 200.8 percent of income between 2022 and 2025, so L = +13.5 and the one-year probability is 5.65 percent, against a baseline of 3.89 percent.
The food threshold. Above an FAO food price index of about 210 (old base) discharges follow where the valve is shut. Converted to the current base, the threshold is 122.
The open-mesh test. For each critical knot: does the load meet at one point, and is there a second path? One point and no second path means an open mesh.
3. The clocks
The war with Iran began 28 February 2026 and has lasted 0.616 years — layer 48. That layer ends on 29 October 2026; the centre of layer 49 falls on 25 April 2027; layer 49 ends on 27 February 2028.
The war in Ukraine began 24 February 2022 and has lasted 4.63 years — layer 50. That layer ends on 21 February 2028; the centre of layer 51 falls in July 2032.
The arithmetic is plain and the result is the essay’s central finding: both wars reach a layer boundary in the same month, February 2028. Both began in the last week of February, four years apart, and their boundaries (two years and six years) differ by exactly a factor of three — one layer. A war that has not closed at its boundary belongs to the slower layer from then on; for Ukraine that would mean a winding of ten years. The theory accordingly names 2026–2028 as the settlement window: the mobilisation and rearmament of 2022–2025 reach their decision point one to three years later, and the ladder places the end of that window in February 2028.
Three layer centres already correspond to recorded turns in the wars, which lends the reading some empirical support: the Iran ceasefire of 8 April 2026 (centre of layer 47, 16 April), the collapse of that ceasefire on 8 July 2026 (centre of layer 48, 18 July), and the Alaska summit of 15 August 2025 (centre of Ukraine’s layer 50, 10 August).
El Niño. The present El Niño began in April–June 2026, with the July–September index at 2.16 degrees and the peak in October–December 2026; the discharge lasts into early 2027. On the theory’s (conditional) φ-rhythm, the years 2027 and 2028 are rest, and a new onset is most likely around the turn of 2028–2029.
The Sun. The sunspot cycle sits on layer 51. The smoothed maximum fell in October 2024 at 160.8; the minimum is forecast for November 2030. On Chizhevsky’s hypothesis — adopted here conditionally — the background tension in societies falls from 2029 towards that minimum.
Key dates ahead:
When
What
29 October 2026
the war with Iran leaves layer 48
Dec 2026 – Mar 2027
peak and end of the El Niño discharge; food above the threshold
25 April 2027
centre of layer 49 for the war with Iran
mid-2027
the United States reaches its debt limit
autumn 2027
Chinese party congress; end-2027 is the army’s readiness goal
Jan–Feb 2028
both wars at a layer boundary; Taiwan’s presidential election; earliest new El Niño
2029
most likely new El Niño; background tension starts to fall
November 2030
solar minimum
4. The measures
Credit. Three countries are lightly loaded — Saudi Arabia (+18.7, 6.5 percent a year), Russia (+14.2, 5.8 percent) and China (+13.5, 5.7 percent). The West has unloaded: the United States (−13.1, 2.7 percent) and the euro area (−14.8, 2.6 percent) sit below the 3.9 percent baseline. Over five years at equal loading, the loaded three reach 25–29 percent probability of a credit crisis. No country stands in the high zone.
Food. The FAO index stood at 136.0 in September 2026 against the threshold of 122 — 1.11 times the threshold and rising 9.6 percent in eight months. Wheat is 35 percent dearer than a year ago, rice 59 percent. The 2026 world cereal harvest is 2.1 percent smaller, the Indian monsoon was the weakest since 2015, and the El Niño winter’s harvest losses reach prices in 2027.
The slow layer. Between 2026 and 2031 the working-age population (15–64) grows strongly in Iraq (+14.2%), Egypt (+10.8%) and Saudi Arabia (+8.2%), is flat in the United States (+0.7%), and shrinks in Russia (−0.5%), China (−2.4%) and Europe without Russia (−2.5%).
The mesh map. The open meshes form a closed chain of mutual exposure. Europe’s gas comes 63 percent from one supplier (the United States), and its storage is lower than a year ago (71.5% on 2 October against 82.6%); 88 percent of Russia’s frozen assets sit in one depository (Euroclear). Russia sends 87 percent of its crude to two buyers (China 50%, India 37%) and holds half a year of reserve against its deficit. Hormuz can be bypassed by pipeline for only 44 percent of normal flow; the Gulf states import 95 percent of their cereals, 59 percent of it through Hormuz. China moves 75 percent of its seaborne crude through Malacca and took 45 percent of its oil through Hormuz before the war. In the United States, ten companies are 39 percent of the share index, interest takes 19 percent of federal revenue, and 71 percent of rare-earth imports come from China. No knot hangs alone: each region holds an open mesh of another.
5. The regions
Europe (EU and United Kingdom) — holding
Europe’s working-age population shrinks by 9.6 million people in five years; the median age is 44.9 years; defence spending is set to rise from 2.5 to 3.5 percent of income by 2035. Credit has fully unloaded, giving Europe the smallest chance of a credit crisis of the five regions. The load instead sits in the open meshes — one gas supplier, lower storage — while political tension has an open channel in elections (France in April 2027, then Spain, Italy and Poland), which produces small discharges rather than a rupture: the form changes, the structure stands. Russia pushes tension into European meshes below the threshold of war — sabotage cases have risen to about fifteen a month — and European rearmament is loading, with its decision point in 2028–2030.
Russia — discharging, reserve spent
Russia has been discharging since 2022. Its estimated losses are 1.4 million killed, wounded and missing; defence rises in 2027 by more than a quarter, to 35 percent of the federal budget; the key rate is 14 percent and growth about 1 percent. The critical reading: the liquid reserve is 1.6 percent of income against a 3.2 percent deficit — half a year, with no second path. The front hardly moves (14 square miles in four weeks), so discharge is turning into holding. What does not close moves its tension to the neighbours — that is the pressure below the threshold on Europe. The 2027 budget is new loading, with its decision point in 2028–2030.
The Middle East — discharging
The region has discharged since 28 February 2026. Hormuz is shut; Iran has 90 percent inflation and shrinks by 7.7 percent; Qatar shrinks by 21 percent; Gaza and Lebanon have not closed. The food measure stands above the threshold and rising, and the discharge comes where the valve is shut: Egypt, Iraq, Syria, Iran — with Egypt’s IMF support programme ending on 15 December 2026. Saudi Arabia carries the highest credit loading of all measured countries (6.5 percent a year) alongside a deficit of 4.9 percent of income. The slow layer keeps loading: each year more young people arrive than there is work.
China — loading and holding
China’s working-age population shrinks by 24 million people in five years; about 20 million retire each year. Credit is loaded (5.7 percent a year) while investment fell 7.2 percent and new-home prices 3 percent. The energy mesh is open — 45 percent of oil came through Hormuz before the war, and seaborne imports fell 23 percent in the first half of 2026 — but the export mesh has closed (only 10.6 percent of exports go to the United States), and China holds the open mesh of others with 85 percent of rare-earth refining. Around Taiwan there is holding without discharge (190 aircraft in the island’s air defence zone in July). Holding without small discharges gives a larger one later, and three decision points fall together: the party congress in autumn 2027, the army’s readiness goal at end-2027, and Taiwan’s election in January 2028.
The United States — loading in debt and one technology
The federal deficit is 2.0 trillion dollars and the ten-year rate is 5.26 percent, its highest in 24 years. The credit measure has unloaded (2.7 percent) — but that measure reads households and companies. The loading sits elsewhere: in the public debt and in artificial intelligence, where ten companies are 39 percent of the share index and AI investment carries a third of growth. Political tension has an open channel in the ballot (November 2026 and 2028); demonstrations rose 77 percent in 2025, and 99.5 percent stayed peaceful.
6. The five years in two phases
Phase one, October 2026 to February 2028: discharge and settlement. Three windings discharge together. The El Niño ends first, in early 2027, and its losses reach the food price that year. The war with Iran has its next centre in April 2027. Both wars reach their boundary in February 2028, in the same months that China and Taiwan pass their decision points. The load in this phase passes through four open meshes: Hormuz, Europe’s gas, Russia’s reserve, and bread in the countries with a shut valve.
Phase two, 2028 to 2031: the slow layers. What closed in phase one is at rest; what did not close has become a slow winding, and the frame is set by the address rather than the tension. Europe and China each lose 2.4–2.5 percent of their workers while Egypt and Iraq gain 11–14 percent. Rearmament in Europe and Russia reaches its decision point in 2028–2030. Credit loading in China, Russia and Saudi Arabia implies a 25–29 percent chance of a credit crisis over five years. A new El Niño is most likely around the turn of 2028–2029, and background tension falls towards the solar minimum of November 2030.
7. Status of the claims
The essay is explicit about the standing of its own claims, which is part of its value. Measured figures (credit, population, prices, budgets, war dates) are taken from sources. Durations, layer positions, boundary dates, probabilities and the food-threshold conversion are derived in the text. The placement of a war on the ladder by its duration is applied here for the first time, and the credit measure is applied outside its 1870–2020, eighteen-country calibration. Three layer centres correspond to recorded turns in the wars — three points of correspondence, not proof. The φ-rhythm of El Niño and the falling solar-linked background tension remain conditional hypotheses. Open questions include whether the food threshold should be read in constant prices (in constant prices the threshold would be 185, and the index at 0.73 of it). External forecasts largely confirm the reading; two are contradicted (US private-credit behaviour, and the calmer background from 2029). Forecasts are not outcomes.
8. Annotated references
The theory
Konstapel, J. (2026a). The Vacuum.Net Theory. Constable Research, Leiden, 28 September 2026.https://constable.blog/wp-content/uploads/vdefThe-Vacuum.Net-Theory.pdf Why read? The five axioms, the natural units and the status ledger of the theory; Rule 4 of the essay is its second axiom. Start with: Sections 2 and 11.
Larson, D. B. (1979). Nothing But Motion. Portland: North Pacific Publishers. Why read? The origin of the natural unit of time, taken from the Rydberg frequency.
Measures and clocks
Lagi, M., Bertrand, K. Z., and Bar-Yam, Y. (2011). “The Food Crises and Political Instability in North Africa and the Middle East.” arXiv:1108.2455. Why read? The origin of the threshold of 210. Short and readable; Figure 1 carries the argument.
Schularick, M., and Taylor, A. M. (2012). “Credit Booms Gone Bust.” American Economic Review 102(2), 1029–1061. Why read? The long record showing credit growth as the warning sign of a financial crisis; the background of the credit measure.
Bank for International Settlements. Credit to the non-financial sector. Series via https://fred.stlouisfed.org/data/QUSPAM770A.txt Why read? The source of every credit figure in the regional comparison; the same address works per country by changing two letters.
Jin, F.-F., Neelin, J. D., and Ghil, M. (1994). “El Niño on the Devil’s Staircase.” Science 264, 70–72. Why read? El Niño as an oscillator locked to the year — the standard account beside which the φ-reading stands.
Chizhevsky, A. L. (1924). Physical Factors of the Historical Process. Kaluga. Why read? The hypothesis that mass unrest follows the solar cycle; used in the essay conditionally.
U.S. Energy Information Administration. Short-Term Energy Outlook, October 2026.https://www.eia.gov/outlooks/steo/report/global_oil.php Why read? The flow through Hormuz, the bypass pipelines, and the expected recovery in 2027.
The Hague Centre for Strategic Studies (2026). Europe’s Strategic Trap. Why read? The record of incidents below the threshold of war in Europe. Read beside: Konstapel (2026f).
In 1931, Hans Bethe solved a chain of magnetic atoms exactly. His method rests on a single condition: a wave that travels around a ring must return in phase with itself. If it does, the state exists; if it does not, there is no state.
Jean-Sébastien Caux, professor at the Institute for Theoretical Physics of the University of Amsterdam, has spent two decades turning that condition into numbers. His publication list at the university holds 77 refereed papers from 2008 to 2025, nearly all on one subject: one-dimensional chains that can be solved exactly. His results have been checked against neutron scattering, cold-atom experiments, and chains built atom by atom.
The Vacuum.Net theory starts from the same condition. Whatever returns onto itself in phase persists; whatever does not close, scatters. This essay lays the two side by side, states the rule of the net and Bethe’s condition, shows by counting that they are one rule written in two notations, and then reads the main results in Caux’s publication list in the terms of the net.
1. The Net in Brief
The picture is a fishnet in water. One strand runs everywhere and crosses itself. The strands are under tension. Where strands wind around each other, knots form; between the strands lie meshes. Four words carry the theory:
Strand — the net itself.
Winding — a strand closed on itself.
Knot — windings that hold each other in three directions.
Mesh — the opening between strands.
At every crossing, the strand has one of three states: +1, 0, or −1. This unit is the trit. Three crossings in a row form a window. A window is closed when its three values sum to zero, counted in threes. An open window carries tension. Only what closes, stays.
Three states on three places give 3 × 3 × 3 = 27 windows. Nine of them close:
Three equal values: (+,+,+), (0,0,0), (−,−,−) — that is 3 windows.
Three different values: the six orders of (−, 0, +) — that is 6 windows.
The other 18 are open.
A ring of three different values has no starting point: six orders divided by three rotations give two rings, the left ring (− → 0 → +) and the right ring (+ → 0 → −). They are mirror images. Their content is the same; only the direction differs.
2. Bethe’s Condition
Take N particles on a ring of circumference L. In the model of Lieb and Liniger (1963), the particles repel each other on contact with strength c. Particle j has wave number k_j. Bethe’s condition reads:
k_j · L + Σ 2 · arctan((k_j − k_l) / c) = 2π · I_j
The sum runs over all other particles l. The number I_j is an integer, or a half-integer when N is even.
The equation has three parts:
k_j · L is the phase the particle collects along the ring.
Each term in the sum is the phase shift from one crossing with one other particle.
2π · I_j says the total must be a whole number of turns.
In the terms of the net: a strand runs once around; on the way it crosses every other strand; each crossing adds its phase; the strand must arrive in phase with itself. That is closure.
The numbers I_j are the address of the state. A set of N different numbers fixes one state, and every state has one set. The address is computed, never assigned.
Caux built a program on this, called ABACUS (Caux, 2009). It walks through the addresses one by one: for each address it solves the condition and computes what that state contributes to a measurement. The method is counting, address by address. Caux also translated the field’s standard text — Gaudin’s La fonction d’onde de Bethe — into English in 2014.
3. From 2π to Three
Bethe’s phase is continuous; the trit has three states. The bridge is one line.
Give each trit a phase of 120° per unit: +1 is +120°, 0 is 0°, −1 is −120°. A window of three trits a, b, c then has total phase (a + b + c) × 120°:
Sum 0: total phase 0°.
Sum +3 or −3: total phase ±360°.
Sum ±1 or ±2: total phase ±120° or ±240°.
The first three are whole turns; the last are not. So “sum zero, counted in threes” is the same statement as “the total phase is a whole number of turns.” The closed window is Bethe’s condition for a strand that has three phases.
4. Three Strands
Why can these chains be solved exactly? The answer is again a window of three.
Take three particles on a line. They can meet in two orders: first 1 with 2, then 1 with 3, then 2 with 3 — or first 2 with 3, then 1 with 3, then 1 with 2. A chain is exactly solvable when both orders give the same result:
S₁₂ · S₁₃ · S₂₃ = S₂₃ · S₁₃ · S₁₂
This is the Yang–Baxter equation (Yang, 1967). When it holds, every meeting of many particles falls apart into meetings of two. Nothing is lost and nothing is added.
Draw the particles as strands and the meetings as crossings, and the equation says that one strand may slide across the crossing of two others — the third Reidemeister move of knot theory (Kauffman, 1991). The condition for exact solvability is the condition under which a braid of three strands keeps its form.
Caux and Mossel (2011) examined what “integrable” actually means for quantum systems and found that the usual definitions do not hold. Read with the net, the working definition is short: a chain is integrable when every three-strand window can be crossed in either order.
5. The Trit Chain
One model in the publication list is a chain of trits in the literal sense. Vlijm and Caux (2014) computed the correlations of the spin-1 chain of Babujian and Takhtajan, in which every site has three states: +1, 0, and −1. Take three sites and count:
There are 3³ = 27 states.
States with sum zero: (0,0,0) and the six orders of (−1, 0, +1) — that is 7.
States with sum +3 or −3: (+,+,+) and (−,−,−) — that is 2.
Together 7 + 2 = 9: exactly the nine closed windows of Section 1.
Physics sorts the same 27 states by total spin. Three spins of size 1 combine to total spin 3, 2, 1, or 0, in the split:
27 = 7 + 5 + 5 + 3 + 3 + 3 + 1
There is one block of spin 3, two of spin 2, three of spin 1, and one of spin 0. Each block contains exactly one state with sum zero. Check: 1 + 2 + 3 + 1 = 7 — the 7 of step 2 above.
The last block matters most. It holds a single state with total spin zero. A state of spin zero does not change under any rotation: it adds nothing to the outside. It is complete. Which state is it? It must change sign when two sites are swapped, and three sites with three states allow exactly one such combination:
Read the terms: the first three are the left ring in its three rotations, the last three are the right ring in its three rotations. So:
|spin 0⟩ = left ring − right ring
The six windows that close with three different values form, with these signs, the one state of three trits that is complete in itself. The two directions enter with opposite sign.
The three windows with equal values have their own place. The states (+,+,+) and (−,−,−) are the reference states of Bethe’s method: every other state is built by counting steps away from them.
6. The Seam
The best-known result in the list concerns the spinon.
Take a chain of alternating spins ↑↓↑↓↑↓↑↓ and flip the fourth spin:
↑↓↑↑↑↓↑↓
Two places now show equal neighbours — between sites 3 and 4, and between sites 4 and 5. Each is a seam in the pattern. Let the neighbours of a seam swap and the seam moves two sites:
↑↓↑↑↓↑↑↓
The two seams drift apart; each travels alone. The flip changed the total spin by 1, and each seam carries half of it. This is the spinon: a carrier of spin ½ that exists only as one of a pair (Faddeev and Takhtajan, 1981).
Now take the net. Put a left ring against a right ring: −, 0, +, +, 0, −.
Window
Sum
State
(−, 0, +)
0
closed
(0, +, +)
2
open
(+, +, 0)
2
open
(+, 0, −)
0
closed
One seam gives two open windows. They come as a pair, as the spinons do.
The numbers are known. Pairs of two spinons carry 72.89% of the total signal of the chain (Karbach et al., 1997). Caux and Hagemans (2006) added the groups of four, bringing the count to about 98%. Neutron measurements on copper sulphate confirmed the count (Mourigal et al., 2013), and a second experiment followed the spinons at finite temperature (Lake et al., 2013).
The net adds a reading of the half. In the net, an electron is a winding that must turn twice before it meets itself; one turn is half a closure. A seam is where one such half is exposed.
In a chain, the properties of one electron can also travel apart. Schlappa et al. (2012), with Caux, measured in Sr₂CuO₃ that the spin and the orbital state of one excited electron move as two separate carriers. In the net, spin, charge, and magnetic moment are three views of one knot; on a single strand, the views separate.
7. What a Crossing Leaves Behind
A soliton is a wave that keeps its form. Two solitons that meet pass through each other, and afterwards each has the same form and speed as before — with one change: each is displaced by a fixed distance.
Doyon, Yoshimura, and Caux (2018) showed that a gas of such solitons obeys the same equations as the quantum chains above. The theory is called generalized hydrodynamics, and its core is one calculation.
Let a tracer move with speed v = 1 through a gas of resting solitons of density n = 0.1 per unit length, where every crossing shifts the tracer forward by Δ = 2:
In one unit of time the tracer covers a distance v_eff.
On that distance it meets n · v_eff solitons; each meeting adds Δ.
The tracer moves 25% faster than its own speed. Double the density and the result is 1 / (1 − 0.4) = 1.67. The speed of a carrier is therefore a function of the state of its surroundings.
The fourth axiom of the Vacuum.Net theory says the same about every constant: it is a property of the local state of the net. In these chains, that statement is exact and measured.
Two further papers move the state on purpose. Bastianello, Alba, and Caux (2019) let the coupling vary in place and time. Koch, Bastianello, and Caux (2021) turned the coupling slowly from repelling to attracting, and free particles then bound into groups — called strings — that travel as one. Windings come to hold each other when the state of the medium is changed and held.
8. A Chain That Does Not Forget
Pour hot water into cold and the mixture forgets where it came from. An exactly solvable chain does not.
In 2006, Kinoshita, Wenger, and Weiss set two clouds of cold atoms swinging against each other in a line. After thousands of collisions, the clouds were still swinging — the quantum Newton’s cradle. Caux and colleagues computed it twice (van den Berg et al., 2016; Caux et al., 2019).
The reason lies in the address. Every closed mode keeps its number; a chain of N particles has N such numbers, and none of them is lost in a collision. Caux and Konik (2012) showed how to build the final state from these numbers. Ilievski et al. (2015) showed that the list is complete only when one family of numbers is kept for every length of string. The net states this as its third axiom: the net remembers how it was laid. A knot forgets how hard it was pulled, never that it was tied.
A related result concerns selection. Change the coupling of a chain in one stroke — which state survives? Caux and Essler (2013) showed that one state does, fixed by how well each address fits the starting pattern. For the alternating pattern ↑↓↑↓ the fit is strict: only states whose wave numbers come in pairs +k and −k fit at all (Brockmann et al., 2014). Everything else has weight zero. The survivor is selected by fitting, and what fits is the mirror pair.
9. Drive, Disturbance, and Direction
Three smaller groups in the list complete the picture.
Drive. Claeys et al. (2018) switched a system of spins back and forth with a fixed period. At most periods little happens; at periods that fit the level spacing of the system, the spins line up. Only a drive that returns in phase leaves a mark.
Disturbance. Brandino, Caux, and Konik (2015) disturbed an exactly solvable gas slightly. Most of its conserved numbers survive in adjusted form. Closure is not fragile: it bends before it opens.
Direction. Veenstra et al. (2025), with Caux, built a mechanical chain in which the push from left to right differs from the push from right to left. The chain carries solitons that pulse with their own period and run one way — a winding with a clock and a fixed direction, on a laboratory table.
10. Place on the Ladder
The net has a ladder: strand, winding, knot, mesh. The chains in this essay are one strand with crossings; they sit on the rung of the winding. That fixes what the comparison covers: the closure rule, the address, the memory of the net, and the constant as a state function. Knots — which close in three directions — and meshes do not occur in a single chain. They belong to the next rungs.
11. Where the Count Continues
Three places lie open.
The chain of three phases. Fateev and Zamolodchikov (1982) found an exactly solvable chain in which every site has three phases of 120°. In that chain, the rule of Section 3 and the method of Section 2 are the same object. The publication list does not contain it; the tools in the list fit it.
Rings of three and six. In Delft, chains of magnetic atoms are laid one atom at a time (Toskovic et al., 2016, with Caux). The net counts that a ring of three different values cannot grow by one site — it grows by three, to six, and a ring of six divides into two rings of three. A ring of three and a ring of six spin-1 atoms would put that count on a surface.
The one-way chain. The mechanical chain of Section 9 has a built-in direction. The net counts that a strand of three different values cannot reverse direction without an open window. A chain that is one-way by construction is the place to see a seam.
12. Status
From sources: Bethe’s condition; the Yang–Baxter equation; the spinon and its share of 72.89%; the soliton gas; the Newton’s cradle; the pair rule for the alternating pattern.
Counted here: the 27 states and the 9 closed windows as sum 0 and sum ±3; the split 27 = 7 + 5 + 5 + 3 + 3 + 3 + 1; the spin-zero state as left ring minus right ring; the effective speed of 1.25.
Read: Bethe’s condition as closure; Yang–Baxter as the three-strand window; the spinon as the seam; the state-dependent speed as the fourth axiom; the conserved numbers as the memory of the net.
Open: the count for the chain of three phases; the rings of three and six on a surface.
Annotated References
The Net
Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper, 5th ed. Constable Research, Leiden, 6 August 2026.Why read it? The one assumption, the selection rule, and the axioms used in Sections 7 and 8. Reading advice: start with the statement of closure, then the axioms on memory and on state.
Konstapel, J. (2026). Life and Consciousness from One Rule: Closure in the Net. Constable Research, Leiden, 29 September 2026.Why read it? The source of the window, the 27 and the 9, the two kinds of closure, and the step 3 → 6 → 3 + 3; it also gives the mapping of trits to phases of 120° used in Section 3. Reading advice: the tables per depth.
Konstapel, J. (2026). One Direction or None: Mirror Molecules, the 2026 Nobel Prize in Chemistry and the Closure Rule of the Vacuum.Net. Constable Research, Leiden, 10 October 2026.Why read it? The left ring, the right ring, and the seam table repeated in Section 6. Reading advice: Section 4 on the open seam.
Konstapel, J. (2026). The Vacuum.Net Theory, complete derivation. Constable Research, Leiden, 28 September 2026.Why read it? The ladder strand–winding–knot–mesh, on which Section 10 places the chains.
The Condition
Bethe, H. (1931). “Zur Theorie der Metalle. I. Eigenwerte und Eigenfunktionen der linearen Atomkette.” Zeitschrift für Physik 71, 205–226.Why read it? The origin: one chain, one condition, solved by counting. Reading advice: the passage where the integers first appear — they are the addresses.
Lieb, E. H., and Liniger, W. (1963). “Exact analysis of an interacting Bose gas. I.” Physical Review 130, 1605–1616.Why read it? The form of the condition printed in Section 2, for particles on a ring. Clear and short.
Gaudin, M., translated by Caux, J.-S. (2014). The Bethe Wavefunction. Cambridge University Press.Why read it? The full method in one book. Reading advice: chapters 1 and 4; the translator’s notes show where the field has moved since 1983.
Yang, C. N. (1967). “Some exact results for the many-body problem in one dimension with repulsive delta-function interaction.” Physical Review Letters 19, 1312–1315.Why read it? The three-particle condition of Section 4, in four pages.
Kauffman, L. H. (1991). Knots and Physics. World Scientific, Singapore.Why read it? Shows that the Yang–Baxter equation and the sliding of a strand over a crossing are one statement. Reading advice: Part I, sections 8 to 11.
Caux, J.-S., and Mossel, J. (2011). “Remarks on the notion of quantum integrability.” Journal of Statistical Mechanics P02023.Why read it? An honest account of why the word “integrable” has no settled meaning in quantum physics — the background to the short definition in Section 4.
The Count
Caux, J.-S. (2009). “Correlation functions of integrable models: a description of the ABACUS algorithm.” Journal of Mathematical Physics 50, 095214.Why read it? How to walk through the addresses in order of weight — the working method behind most of the list. Reading advice: the section on scanning.
Vlijm, R., and Caux, J.-S. (2014). “Computation of dynamical correlation functions of the spin-1 Babujian–Takhtajan chain.” Journal of Statistical Mechanics P05009.Why read it? The chain with three states per site, computed in full — the model of Section 5.
Takhtajan, L. A. (1982). Physics Letters A 87, 479–482; Babujian, H. M. (1982). Physics Letters A 90, 479–482.Why read them? The two papers that found the solvable chain of three states — a few pages each.
The Seam
Faddeev, L. D., and Takhtajan, L. A. (1981). “What is the spin of a spin wave?” Physics Letters A 85, 375–377.Why read it? Three pages that established the carrier of spin ½ and that it comes in pairs.
Karbach, M., Müller, G., Bougourzi, A. H., Fledderjohann, A., and Mütter, K.-H. (1997). “Two-spinon dynamic structure factor of the one-dimensional s = 1/2 Heisenberg antiferromagnet.” Physical Review B 55, 12510.Why read it? The source of the 72.89%.
Caux, J.-S., and Hagemans, R. (2006). “The four-spinon dynamical structure factor of the Heisenberg chain.” Journal of Statistical Mechanics P12013.Why read it? The groups of four, which bring the count to about 98%.
Mourigal, M., Enderle, M., Klöpperpieper, A., Caux, J.-S., Stunault, A., and Rønnow, H. M. (2013). “Fractional spinon excitations in the quantum Heisenberg antiferromagnetic chain.” Nature Physics 9, 435–441.Why read it? The measurement: neutrons on copper sulphate against the count, in absolute units. Reading advice: the figure that sets measured and computed intensity side by side.
Lake, B., Tennant, D. A., Caux, J.-S., et al. (2013). “Multispinon continua at zero and finite temperature in a near-ideal Heisenberg chain.” Physical Review Letters 111, 137205.Why read it? The same carriers followed as the chain warms.
Schlappa, J., et al. (2012). “Spin–orbital separation in the quasi-one-dimensional Mott insulator Sr₂CuO₃.” Nature 485, 82–85.Why read it? The properties of one electron seen travelling apart.
State and Memory
Doyon, B., Yoshimura, T., and Caux, J.-S. (2018). “Soliton gases and generalized hydrodynamics.” Physical Review Letters 120, 045301.Why read it? The calculation of Section 7 in its general form: every crossing is a shift, and the speed follows from the density. Reading advice: the first two pages.
Bastianello, A., Alba, V., and Caux, J.-S. (2019). “Generalized hydrodynamics with space-time inhomogeneous interactions.” Physical Review Letters 123, 130602.Why read it? The coupling itself made a function of place and time.
Koch, R., Bastianello, A., and Caux, J.-S. (2021). “Adiabatic formation of bound states in the one-dimensional Bose gas.” Physical Review B 103, 165121.Why read it? Free particles bound into strings by changing the state slowly.
Kinoshita, T., Wenger, T., and Weiss, D. S. (2006). “A quantum Newton’s cradle.” Nature 440, 900–903.Why read it? The experiment: a gas that keeps swinging.
Caux, J.-S., and Konik, R. M. (2012). “Constructing the generalized Gibbs ensemble after a quantum quench.” Physical Review Letters 109, 175301.Why read it? How the conserved numbers fix the final state.
Ilievski, E., De Nardis, J., Wouters, B., Caux, J.-S., Essler, F. H. L., and Prosen, T. (2015). “Complete generalized Gibbs ensembles in an interacting theory.” Physical Review Letters 115, 157201.Why read it? The list of conserved numbers made complete: one family per string length.
Caux, J.-S., and Essler, F. H. L. (2013). “Time evolution of local observables after quenching to an integrable model.” Physical Review Letters 110, 257203.Why read it? Selection of one surviving state by fit.
Brockmann, M., De Nardis, J., Wouters, B., and Caux, J.-S. (2014). “A Gaudin-like determinant for overlaps of Néel and XXZ Bethe states.” Journal of Physics A 47, 145003.Why read it? The pair rule: only wave numbers in pairs +k, −k fit the alternating pattern.
Drive, Disturbance, Direction
Claeys, P. W., De Baerdemacker, S., El Araby, O., and Caux, J.-S. (2018). “Spin polarization through Floquet resonances in a driven central spin model.” Physical Review Letters 121, 080401.Why read it? A drive that works only at fitting periods.
Brandino, G. P., Caux, J.-S., and Konik, R. M. (2015). “Glimmers of a quantum KAM theorem.” Physical Review X 5, 041043.Why read it? What survives when exact solvability is slightly broken.
Veenstra, J., Gamayun, O., Brandenbourger, M., van Gorp, F., Terwisscha-Dekker, H., Caux, J.-S., and Coulais, C. (2025). “Nonreciprocal breathing solitons.” Physical Review X 15, 031045.Why read it? A one-way chain with pulsing solitons, built and filmed. Reading advice: watch the supplementary video first.
Where the Count Continues
Fateev, V. A., and Zamolodchikov, A. B. (1982). “Self-dual solutions of the star-triangle relations in Z_N-models.” Physics Letters A 92, 37–39.Why read it? The exactly solvable chain with three phases per site. Three pages.
Toskovic, R., van den Berg, R., Spinelli, A., Eliens, I. S., van den Toorn, B., Bryant, B., Caux, J.-S., and Otte, A. F. (2016). “Atomic spin-chain realization of a model for quantum criticality.” Nature Physics 12, 656–660.Why read it? Chains laid atom by atom and measured site by site. Reading advice: the first figure shows the chains as built.
University of Amsterdam. Profile page of Prof. dr. J.-S. Caux, publication list.https://www.uva.nl/profiel/c/a/j.s.caux/j.s.caux.html — Why read it? The full list of 77 refereed publications on which this essay draws, most with open downloads.
From Atom to Body is a reference work of 68 entries, about 196,000 words and 1,462 references. It runs from the hydrogen atom to the standing person, then on to what the body needs and what the Earth supplies. It was published in October 2026 and is available on Academia.edu and ResearchGate.
Every number in it is measured in a cited source, or computed from such numbers with the arithmetic shown. Every size and every duration sits on one scale. Every piece of advice carries a grade. Every rule the book uses is stated in its first part, and no entry rests on anything outside it.
That is the whole claim. The rest of this essay says what those sentences mean in practice.
2. One scale for everything
Biology is usually told in pieces. Chemistry has its nanometres, cell biology its micrometres, anatomy its centimetres, physiology its milliseconds and its years. The pieces do not meet.
The book puts all of them on one ladder. The unit is Larson’s natural unit of space, 4.558816 × 10⁻⁸ m, and his natural unit of time, 1.520655 × 10⁻¹⁶ s. The position of a size is the number of factors of three between it and the unit. Nothing more.
The Bohr radius, 52.9 pm, sits at −6.15. A Dutch man born in 1996, 182.5 cm, sits at 15.93. The whole human build spans 22 positions. In time, the bond vibration of water at 9.8 fs sits at 3.79. The day sits at 43.50. A life of 80 years sits at 52.86. Forty-nine positions.
Every entry gives the measured size and the measured times of its subject, each with its position to two decimals. Table A at the back gathers 388 sizes. Table B gathers 411 durations. A reader who wants to know how far the ribosome stands from the red cell, or the heartbeat from the sleep cycle, reads it off.
The ladder orders. It does not select. The book tested that before it was built. If the ladder preferred certain sizes, measured parts of the body would cluster near whole positions. They do not: 192 of 388 sizes lie within a quarter step of a whole mark, 49.5 per cent, against 50 by chance. The book therefore gives positions and assigns no part of the body to a mark. The ladder is a coordinate system, and it is used as one.
3. Four marks
Each statement of fact carries one of four marks.
Measured means read in the cited source. The book holds about 1,200 of these. Computed means arithmetic on measured numbers, shown in the text. About 620. Derived means that the statement follows from the rules of the first part, or from Larson’s formulas, with the step shown. Twelve. Read means an interpretation that sets a measured fact in the net. About 130.
The counts say what kind of book it is. It is a book of measurements. The theory enters in one section of each entry, is marked where it enters, and is never allowed to carry a number the sources do not carry.
An example from the entry on water. The molecule splits once in about eleven hours. Set against its bond vibration of 9.8 fs, it holds through 4 × 10¹⁸ vibrations. The hydrogen bond lasts about a picosecond, about 100 vibrations. The first number is measured, the second computed. The reading that the molecule is a closed form and the hydrogen bond an unclosed one is marked read. A reader who does not want the reading can skip the sentence and keep the numbers.
4. Two classes
The one rule of the theory that the book uses throughout is the division into two classes. Closed forms are things. They last. Unclosed forms are events between things. They end.
At every depth the measured lifetimes fall into the two classes. The water molecule and the hydrogen bond. The protein and the enzyme–substrate complex. The cell and the action potential. The neuron that is never replaced and the gut-lining cell that is replaced every few days. The book does not say that the rule causes the division. It sets the rule beside the measured lifetimes, at every depth, and lets the reader see that they fall the same way.
5. What the body needs, with grades
Part 6 turns the measurements into advice. There are about 480 advice rows in the book. Each has a grade and a reason, and the reason holds a number and a reference.
Strong rests on trials, or on large studies that agree, or on settled physiology. Moderate rests on one good trial or on consistent observation. Weak rests on little or indirect evidence. Contested means good sources disagree, and both numbers are then printed.
Some of the strong rows. Take vitamins and minerals from food first: in 78 trials with 296,707 people, antioxidant supplements did not lower mortality, and in the 56 trials with low risk of bias they raised it. Move every day, 150 to 300 minutes a week: 661,137 adults show the dose curve. Do not smoke: hazard ratio about 3, ten years lost. Twenty-five to thirty grams of fibre a day: 185 cohorts and 58 trials, relative risk 0.85 for mortality, greatest reduction at 25 to 29 g.
And the one trial on processed food, printed with its size. Twenty adults in a metabolic ward for 28 days, two diets matched for energy, sugar, salt and fibre. On the ultra-processed diet they ate 508 kcal a day more and gained 0.9 kg in two weeks. The observational evidence behind the same advice covers 9.9 million people, but of its 45 pooled analyses, 4 have moderate certainty and 41 low or very low. The book prints both. The trial is small and the observation is large, and the reader sees that.
Where the sources disagree, the book does not choose. Alcohol: the hazard is settled, the threshold is contested, and the J-curve and the zero both stand on the page with their consortia named.
6. What the Earth supplies
Part 7 follows the needs back to their source. Light and photosynthesis. Soil and the minerals. Nitrogen, from microbes and from factories. Phosphorus. The water cycle. Plants, microbes, animals and pollinators. Air. The living surroundings.
Each is treated as the entries before it: measured sizes, measured times, positions, what the body draws from it. The body is not closed at the skin. Its elements come from the ground, its nitrogen from bacteria and from the Haber process, its oxygen from the leaf. The book gives the numbers for each step.
7. Self-contained
Part 0 holds every rule: the net, the register, closure, rings, the two classes, widening and pruning, the ladder, Larson’s units and formulas, the marks, the grades and the words. Nothing in the entries rests on a rule that is not stated there. No entry sends the reader to another book for content. No list of open questions closes the work.
Each entry has the same eight sections in the same order. What it is. Place on the ladder. Build. Times. In the net. What it does in the body. What it needs and what the Earth supplies. Advice. Then the annotated references, with a note on why the useful ones are worth reading. A reader who knows one entry knows them all.
8. Who it is for
A physician who wants the measured size and turnover of the tissue in front of her, with the source. A chemist who wants to see the molecule placed among the organs it builds. A reader of the Vacuum.Net theory who wants to see what it says, and what it does not say, when set beside 1,200 measurements. A person who wants advice on food, movement, sleep and air that comes with its evidence attached and its weak spots printed.
The book can be read in order, from 1.1 to 7.10. It can be opened at one entry. It can be used from the tables alone. The index holds 470 terms.
9. What it does not do
It does not explain. It does not say that the net causes any biological fact. It does not claim that a formula reproducing known values proves the model behind it; Larson’s distance rule gives seven elements within 2 per cent and his mass rule misses hydrogen by a factor of two, and both facts are printed. The tests that could decide the theory are collected in a separate essay, What Would Test the Net, with the outcomes fixed before the measurement.
What the book does is put the measurements in one place, on one scale, with their sources, their arithmetic and their grades, and set one set of rules beside them so that the reader can see where they meet and where they do not.
Annotated references
Konstapel, J. (2026). From Atom to Body: A Reference Work on the Build of the Human Body, What It Needs and What the Earth Supplies. Constable Research, Leiden, first edition. Academia.edu and ResearchGate.Why read? The work this essay describes. Reading advice: “How to use this book” and Part 0 first; then one entry in full, 1.4 Water or 6.2 Whole Food and Unwound Food, to see the form; then the tables.
Konstapel, J. (2026). What Would Test the Net. Constable Research, Leiden, 10 October 2026.Why read? The six tests that follow from the theory, each with its outcome fixed in advance, and the one test already run.
Konstapel, J. (2026). The Vacuum.Net Theory: Complete Derivation. Constable Research, Leiden, 28 September 2026.Why read? The five axioms and the derivation of the rules that Part 0 of the book restates.
Konstapel, J. (2026). Two Kinds of Closure. Constable Research, Leiden, 1 October 2026.Why read? The two classes, closed forms and unclosed carriers, that the book sets beside the measured lifetimes at every depth.
Larson, D. B. (1979). Nothing But Motion. Portland: North Pacific Publishers. And Larson, D. B. (1988). Basic Properties of Matter. Salt Lake City: International Society of Unified Science.Why read? The natural units of space and time that fix the ladder, and the distance and mass formulas the book recomputes and prints with their differences. Reading advice: chapters 9, 10, 12 and 13 of the first; chapters 1 and 24 of the second.
Hall, K. D., Ayuketah, A., Brychta, R., et al. (2019). “Ultra-processed diets cause excess calorie intake and weight gain: an inpatient randomized controlled trial of ad libitum food intake.” Cell Metabolism 30, 67–77.Why read? The one controlled trial on processed food: 20 people, 28 days, 508 kcal a day.
Lane, M. M., Gamage, E., Du, S., et al. (2024). “Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses.” BMJ 384, e077310.Why read? The 45 pooled analyses on 9.9 million people, with their GRADE certainty: 4 moderate, 41 low or very low.
Reynolds, A., Mann, J., Cummings, J., Winter, N., Mete, E., and Te Morenga, L. (2019). “Carbohydrate quality and human health: a series of systematic reviews and meta-analyses.” The Lancet 393, 434–445.Why read? The fibre evidence behind the strongest dietary advice in the book: 185 cohorts, 58 trials, relative risk 0.85.
Bjelakovic, G., Nikolova, D., Gluud, L. L., Simonetti, R. G., and Gluud, C. (2012). “Antioxidant supplements for prevention of mortality in healthy participants and patients with various diseases.” Cochrane Database of Systematic Reviews, issue 3, CD007176.Why read? Seventy-eight trials, 296,707 people; the reason the book says food first.
Monteiro, C. A., Cannon, G., Levy, R. B., et al. (2019). “Ultra-processed foods: what they are and how to identify them.” Public Health Nutrition 22, 936–941.Why read? The NOVA classification and its one-ingredient test, which the book uses to separate whole food from unwound food.
The 2026 Nobel Prize in Physiology or Medicine has been awarded to Peter Hegemann, Georg Nagel and Karl Deisseroth for optogenetics — a technique that makes nerve cells answer to light.
The story begins in a single-celled alga. Chlamydomonas has a tiny eyespot containing retinal, the same light-sensing molecule our own eyes use. When light strikes the alga, it produces an electrical response in about half a millisecond. The human eye, by comparison, needs at least ten milliseconds.
In the early 1990s Hegemann proposed something that sounded improbable: a single protein in the alga both detects the light and lets ions cross the membrane. At the time, no known channel responded to light on its own. Sensing and gating were assumed to be separate jobs, done by separate proteins.
The evidence arrived step by step:
Two genes were found: channelrhodopsin-1 and channelrhodopsin-2 (ChR2).
Nagel inserted each gene into frog eggs. Both proteins opened under light — and ChR2 passed ions within 0.2 milliseconds.
The gene also worked in mammalian cells, which normally pay no attention to light at all.
In 2005 Deisseroth placed the gene in rat nerve cells. Blue light made them fire an impulse.
Later tools switch cells on or off, each at its own wavelength.
This essay reads those facts through a single rule from the Vacuum.Net theory: the closure rule. Every count is written out in the text.
2. The rule
Picture a fishnet stretched in water. One strand crosses itself; light is the movement of phase along that strand. Between the strands lie the meshes of the net — openings, each with a shape and a size.
Each crossing carries one of three values (a trit): +1, 0 or −1. Three crossings in a row form a window. A window is closed when its three values sum to zero, counted in threes. An open window carries tension. Only what closes, stays.
Three values in three places give 3 × 3 × 3 = 27 possible windows. Nine of them close:
Three equal values — (−,−,−), (0,0,0), (+,+,+) — three windows. This is the form of a crystal.
Three different values — the six orderings of (−, 0, +) — six windows. This is the form of metabolism: the content changes while the window stays closed.
The remaining 27 − 9 = 18 windows are open.
A ring is a strand that returns onto itself. A ring of three different values exists at length 3 and grows by whole triples to length 6, where it divides into 3 + 3. That is the form of a single cell. Where such a ring touches its surroundings, the windows stand open: tension at the edge, rest inside. That edge is the form of a cell membrane.
3. A window with one free place
Fix two values of a window — call them a and b. The third value c closes the window when
c ≡ −(a + b), counted in threes.
The full table of nine cases:
a b
a + b
c that closes
− −
−2
−
− 0
−1
+
− +
0
0
0 −
−1
+
0 0
0
0
0 +
1
−
+ −
0
0
+ 0
1
−
+ +
2
+
Each row has exactly one closing value: two fixed places leave one answer out of three.
Now read the channel. The protein with its retinal holds two places of the window. The third place is taken by the phase that arrives along the strand — which is the light. One phase out of three closes the window; the other two leave it open. That is the colour of the channel. ChR2 answers to blue light, near 470 nanometres. It does not answer to yellow.
4. Reading and passing are one closure
Hegemann’s proposal was doubted for one reason: a sensor and a channel were known as two separate things, and he claimed one protein was both.
In the net there is only one event: the window closes. That closure is the passage. There is no second step in which a reading is handed over to a gate.
The human eye does the same work in a chain. The light-sensing protein hands its signal to a second protein, that one to a third, and so on, until a channel finally moves. Each handover is its own window.
The numbers make the contrast plain:
Human eye: at least 10 ms.
Alga, whole answer: 0.5 ms — a ratio of 10 / 0.5 = 20.
ChR2 channel alone: 0.2 ms — a ratio to the alga’s whole answer of 0.5 / 0.2 = 2.5.
Eye against one channel: 10 / 0.2 = 50.
One window against a chain: a factor of 50 in time. The alga has no brain and no nerve. It has the shortest possible route from light to movement — and it is the fastest.
5. The same edge everywhere
The gene was moved four times: from the alga to a frog egg, to a mammalian cell, to a rat nerve cell in a dish, and to nerve cells in a living brain. It worked every time, in hosts that had never met it.
In the net this needs no special explanation. The membrane is the edge of a ring of length 6 (or a multiple of it), and that edge has the same form in an alga as in a nerve cell: open windows, tension outside, rest inside. A mesh that fits one such edge fits every such edge.
At this depth the nerve cell adds nothing new. It is a cell whose edge tension travels along its length. A mesh that passes at one place lowers the tension there, and the change runs along the edge.
6. Three positions of the switch
The later tools complete the register:
Light
Tool
What crosses the edge
Trit
Blue, ~470 nm
ChR2
positive ions inward
+1: the cell fires
None
—
nothing
0: the cell follows its own course
Yellow, ~590 nm
Halorhodopsin
negative ions inward
−1: the cell is held silent
The two wavelengths differ by a factor 590 / 470 = 1.26.
The technique is usually described as on and off. In this reading it has three positions, and the zero is a lawful state. In darkness the cell is neither on nor off; it is left to the rings around it.
7. Address and tension
The two-layer law of the theory says every configuration has an address (which windings close, and at what depth) and a tension (what happens there). The name of the technique holds both layers:
Genetics writes the address. The gene is read in triples of three letters and is placed only in the cells of one chosen kind. That fixes which rings carry the mesh.
Optics sets the tension. Light arrives everywhere in the tissue. Only the rings with the mesh answer.
Before this technique the two layers could not be separated in living tissue. An electrode sets tension but reaches every cell near its tip. A drug has an address of a sort, but no clock. Here the address is exact to the cell kind, and the tension is exact to the millisecond.
8. The experiment of the reading ring
The theory describes representation as a reading ring. Take three closed rings, each touching a fourth ring at one crossing. The value at each crossing is the phase of that ring; the fourth ring takes those three phases as its own trits. It reads the three.
The reading keeps structure: shift one ring below by +1, and the image above shifts by +1.
A ring that counts itself has two foreign phases, p₁ and p₂, and its own phase. Closure fixes its own phase:
p_self ≡ −(p₁ + p₂).
Take p₁ = 0 and p₂ = +1. Then p_self ≡ −1. Now shift p₁ one step, from 0 to +1. Then p₁ + p₂ = 2, and p_self ≡ −2 ≡ +1.
One step below moved the fixed point above from −1 to +1. The ring stays closed; its content is different.
That is the form of an optogenetic experiment. One group of cells receives one step, and the experimenter reads what has moved above: a memory that returns, a fear that appears, a movement that starts. In 2012 a group at MIT lit the cells that had been active during one event in a mouse. The mouse froze, as it had during the original event. The step was given below; the image appeared above.
9. One step, not alignment
Coherence is usually measured with the order parameter R: the length of the mean arrow on a circle. Map the three trit values to the phases 0°, 120° and 240°:
Three equal values: the three arrows coincide. R = 1.
Three different values: the arrows cancel. R = 0.
Two equal, one different (e.g. 0°, 0°, 120°): the sum is 2 + (−0.5 + 0.866i) = 1.5 + 0.866i, with length √(2.25 + 0.75) = √3 = 1.732. Divided by 3: R = 0.58.
The living form measures R = 0. The crystal form measures R = 1.
Apply this to lit tissue. Take 100 rings balanced over the three phases, so R = 0. Set 10 of them to one phase with light: the remaining 90 still cancel, and the mean arrow has length 10 / 100 = 0.10. Set all 100 to one phase: R = 1.
Lighting a whole population at once pushes living tissue toward the crystal form. Clinically, a brain in which all cells fire together is a brain in seizure. The count gives the direction for use: one step on one ring, read above — not all rings in one direction.
10. A mesh instead of a substance
An earlier essay in this series set out three roads to a medicine. The third reads: first the state, then the substance. On that road the medicine is a mesh — it sets a tempo without being used up.
The sight work honoured by the Nobel citation stands on that road. In retinitis pigmentosa the light-reading cells of the retina die, but the cells behind them remain, and so does the nerve to the brain. The edge is there; the mesh is gone.
In 2021 a group in Paris and Pittsburgh reported one case. A 58-year-old man who had been blind for about forty years received a light-gated channel in the remaining cells of one eye. With goggles that project amber light, he could find, count and touch objects on a table.
Nothing is swallowed daily. Nothing is broken down. One mesh was placed in an edge that had lost its own. Light does the rest.
11. What it means for people
Three groups of people stand behind the prize.
The blind. Retinitis pigmentosa affects about one person in four thousand. For them the technique is the return of a lost mesh. What returns is coarse — shapes and contrast, with goggles — but a person who has lived for decades by touch and sound gains the outline of a cup on a table.
The ill. The citation names depression and schizophrenia. In the net these are not broken parts; they are rings that are not read, or read at the wrong step. The technique has shown, in animals, which rings carry which image. That map is its largest yield so far, and it needs no light in a human head to be used.
The opened. Light in a living brain needs a gene carried in by a virus and a fibre through the skull. The theory adds one fact to that picture. A shared reading ring reads in both directions: whoever sets a step in another person’s ring is also set by what comes back. The technique is usually described as control. In the net there is no controller outside the strand.
12. Status
From the source: the laureates; the times 0.5, 10 and 0.2 ms; the two genes; the hosts; the year 2005.
Counted here: the 27 windows and the 9 that close; the nine pairs with one closing value each; the ratios 20, 2.5 and 50; the three values of R; the R = 0.10 for ten lit rings in a hundred; the move of the fixed point from −1 to +1.
Read: the channel as a window with two fixed places; the membrane as the same edge in alga and nerve cell; the gene as address and the light as tension; the experiment as a step under a reading ring.
Open: which phase value belongs to which wavelength; whether the on-channel and the off-channel are a mirror pair, as the rings (−, 0, +) and (+, 0, −) are; whether the factor 50 between eye and channel is a count of windows in the chain of the eye.
13. Close
The alga had it first: one protein, one window, 0.2 milliseconds. The eye of the animal built a chain on top of it and became fifty times slower — and far more sensitive. The prize honours three people who carried the one window back into the animal.
The closure rule adds a plain reading. The channel does not detect light and then open; it closes with one phase out of three, and that closure is the passage. The same edge runs from the alga to the human brain. That is why a pond organism without a nerve could give neuroscience its sharpest instrument.
Annotated references
The occasion
González, F. (2026). “The Science Behind the Nobel-Winning Technology That Controls Neurons With Light.” WIRED, 7 October 2026. Why read it: the prize, the three laureates and the route from alga to nerve cell in one page. Reading advice: note the three times — 0.5, 10 and 0.2 ms; Section 4 of this essay rests on them.
The net
Konstapel, J. (2026). Life and Consciousness from One Rule: Closure in the Net. Constable Research, Leiden, 29 September 2026. Why read it: the source of the closure rule, the two kinds of closure, the ring of 3 and 6, the membrane as edge, the reading ring and the fixed point. Reading advice: start with the table of closed rings per depth, then the part on the reading ring.
Konstapel, J. (2026). One Direction or None: Mirror Molecules, the 2026 Nobel Prize in Chemistry and the Closure Rule of the Vacuum.Net. Constable Research, Leiden, 10 October 2026. Why read it: the window with two fixed places and one free place, worked out for a catalyst — the same count carries Section 3 here. Reading advice: Section 4 for the open seam, Section 6 for the mirror pair.
Konstapel, J. (2026). What Fits, Stays: Designer Molecules, the Closure Rule of the Vacuum.Net, and Three Simple Roads to a Medicine. Constable Research, Leiden, 10 October 2026. Why read it: the three roads; the third — a mesh instead of a substance — is the road of Section 10. Reading advice: Part 5.
Konstapel, J. (2026). The Interpreter Behind the Gate. Constable Research, Leiden, 21 September 2026. Why read it: the two-layer law of address and tension, set beside fifty years of split-brain findings. Reading advice: Sections 2 and 3 for the law, Section 9 for the clinic.
Konstapel, J. The Vacuum.Net Theory: Foundational Paper, fifth edition. Constable Research, Leiden. Why read it: the starting assumption used throughout — one strand, and light as the movement of phase along it. Reading advice: the axioms first.
The alga and the channel
Harz, H., and Hegemann, P. (1991). “Rhodopsin-regulated calcium currents in Chlamydomonas.” Nature 351, 489–491. Why read it: the electrical measurement on the alga that showed how fast it answers. The speed is what made a single protein the only fitting reading. Reading advice: three pages; look at the delay in the current traces.
Nagel, G., Ollig, D., Fuhrmann, M., et al. (2002). “Channelrhodopsin-1: a light-gated proton channel in green algae.” Science 296, 2395–2398. Why read it: the first proof, in frog eggs, that one protein reads light and passes ions.
Nagel, G., Szellas, T., Huhn, W., et al. (2003). “Channelrhodopsin-2, a directly light-gated cation-selective membrane channel.” Proceedings of the National Academy of Sciences 100, 13940–13945. Why read it: ChR2 itself, and its first use in mammalian cells — the paper in which the mesh leaves the alga. Reading advice: the last figure, with cells that had never answered to light.
The nerve cell
Boyden, E. S., Zhang, F., Bamberg, E., Nagel, G., and Deisseroth, K. (2005). “Millisecond-timescale, genetically targeted optical control of neural activity.” Nature Neuroscience 8, 1263–1268. Why read it: the 2005 result. The title already names both layers: “genetically targeted” is the address, “millisecond-timescale” is the tension.
Zhang, F., Wang, L.-P., Brauner, M., et al. (2007). “Multimodal fast optical interrogation of neural circuitry.” Nature 446, 633–639. Why read it: the off-switch — a light-driven pump that answers to yellow light, used beside ChR2 in the same cells. With it the register of Section 6 is complete. Reading advice: the figure in which blue and yellow light alternate on one cell.
Liu, X., Ramirez, S., Pang, P. T., et al. (2012). “Optogenetic stimulation of a hippocampal engram activates fear memory recall.” Nature 484, 381–385. Why read it: the experiment of Section 8. The cells of one event are given one step, and the memory appears.
Engel, A. K., and Singer, W. (2001). “Temporal binding and the neural correlates of sensory awareness.” Trends in Cognitive Sciences 5, 16–25. Why read it: the standing view that synchrony carries awareness. Read it beside Section 9, where full alignment is the crystal form and R = 0 the living one.
The human
Sahel, J.-A., Boulanger-Scemama, E., Pagot, C., et al. (2021). “Partial recovery of visual function in a blind patient after optogenetic therapy.” Nature Medicine 27, 1223–1229. Why read it: the one published case in which a blind man regained the outline of objects. Reading advice: the description of the table test, and what the man himself could and could not do.
Deisseroth, K. (2021). Projections: A Story of Human Emotions. New York: Random House. Why read it: a psychiatrist writing about his own patients beside his own laboratory; it shows what the people behind the citation live with. Reading advice: the chapters on mania and on eating disorders.
In het brein ontstaan klonten van eiwit. Bij dementie heet de klomp plaque. Bij Parkinson heet hij Lewy-body. Deze klompen ontstaan doordat eiwitten verkeerd vouwen en aan elkaar plakken. Eenmaal geplakt, komen ze bijna niet meer los.
Waarom ontstaan ze? Eiwitten blijven te lang liggen. Bovendien raakt het opruimsysteem trager. Denk aan een bad waar de kraan even hard loopt als de afvoer. Wordt de afvoer trager, dan loopt het bad vol.
Waarom is dementie erger bij slaaptekort? In diepe slaap ruimt het brein af. Wie weinig diep slaapt, hoopt sneller op.
Wat helpt (voor zover gemeten)?
Goed slapen en dagelijks bewegen.
Gezond eten en de bloeddruk onder controle.
Hoortoestel bij gehoorverlies.
Sociale contacten en denkwerk.
Bij Parkinson soms de handleiding van het opruimingsenzym of diepe hersenstimulatie.
Wat helpt niet of zelfs schaadt?
IJzer uit het brein ‘halen’ (patiënten werden er slechter van).
Aanvoer remmen zonder de afvoer te herstellen.
De klompen met antilichamen weghalen werkt wel, maar verwijdert ook schade: er ontstaan kleine bloedingen, en verloren weefsel komt niet terug.
De kern in één zin: de ziektes ontstaan als opruimen achterblijft bij aanmaken; behandelen lukt het best door vóórkomen (slaap, vaat, beweging) en door de afvoer te helpen — niet door alleen de klomp te verwijderen.
1. De vraag
Dementie en de ziekte van Parkinson worden meestal gezien als twee ziekten van twee eiwitten. De geneeskunde verwijdert het eiwit. Of ze vervangt het ontbrekende signaal. De resultaten zijn klein.
Dit essay stelt een andere vraag. Het begint onderaan, bij één regel van de Vacuum.Net-theorie. Het telt uit wat die regel toestaat. Daarna loopt het omhoog, van atoom naar lichaam. Het markeert twee soorten pad: het pad waarlangs de ziekte ontstaat, en het pad waarlangs het lichaam zich herstelt. Bij elk pad staat een meting uit de bestaande literatuur.
2. De regel
Stel een visnet voor, gespannen in water. Eén draad kruist zichzelf in knopen. Elke knoop draagt één van drie waarden: +1, 0 of −1.
Drie buurknopen vormen een venster. Een venster is gesloten als de drie waarden optellen tot een veelvoud van drie. Een open venster draagt spanning. Alleen wat sluit, blijft.
Er zijn 27 vensters. Negen zijn gesloten. Drie zijn eenvormig: drie keer dezelfde waarde. Dat is de vorm van een kristal. Zes zijn gemengd: +1, 0 en −1 in een of andere volgorde. Die vorm blijft draaien en blijft toch gesloten. Dat is de vorm van levende materie.
Drie gevolgen doen er toe.
Twee bepalen de derde. Zijn twee waarden bekend, dan sluit precies één derde waarde het venster.
Een gemengde ring heeft één richting. De ring loopt −→0→+ of +→0→−. De richting kan niet omslaan zonder een open venster.
Een gemengde ring groeit per drie. Een gesloten gemengde ring bestaat niet bij lengte 4 of 5. Hij groeit van 3 naar 6. Een ring van 6 deelt zich in 3+3.
In dit essay heet een eenvormige reeks een stapel. Een gemengde reeks heet een ring.
3. Wat de regel telt
Elk open venster is een paar plus één enkele. Achttien vensters zijn open. Elk daarvan sluit op drie manieren. Eén route leidt naar een stapel. Twee routes leiden naar een ring. Samen: 54 routes. Achttien naar de stapel, zesendertig naar de ring.
Waar een stapel begint. Binnen een draaiende ring verschillen buren altijd. Twee gelijke waarden verschijnen alleen op de naad tussen twee ringen. Per drie stappen biedt één naad een paar. Een vrije knoop kan daar alleen als stapel sluiten. De stapel begint dus niet met de vrije knoop. Hij begint waar twee naburige ringen één stap uit fase lopen.
De stapel groeit per één, de ring per drie. Een stapel groeit knoop voor knoop, bij elke lengte. Hij deelt zich niet. Hij stapelt zich op. Een ring groeit alleen met een volledige triple.
De naad ademt. Zet een draaiende ring tegen een stilstaande stapel. Gemiddeld staat er 4/3 open venster per draai over de naad. De naad beweegt heen en weer. Hij wandelt niet.
Een verplaatsde naad kost meer spanning. Om een stapel-triple in een ring-triple te veranderen, moeten twee knopen wisselen. Het pad loopt 1 → 2 → 1 open vensters. Herstel passeert dus door een staat met méér spanning dan de stilstand zelf. Dit verklaart later de bijwerking van opruimende medicijnen.
4. De ladder van atoom naar lichaam
Het net heeft een maatstok van 45,6 nm. Elke laag is een factor drie. Laag n meet 3ⁿ × 45,6 nm. De plaats van een ding is de logaritme met grondtal drie van zijn grootte, gedeeld door de maatstok.
Een rekenvoorbeeld. De lagen van een amyloïde stapel liggen 0,47 nm uit elkaar. Dat valt op laag −4, die 0,563 nm meet.
Elke derde laag draagt dezelfde toestand. Zo ontstaan drie families:
Familie A (werkers): atoom en water, enzym, haarvat, arteriole, lichaam.
Familie B (materiaal): stapelafstand, vezel, lysosoom en mitochondrion, cellichaam en Lewy-body, substantia nigra.
Familie C (patroon): woord (amyloïd-β), vesikel, synapse, plaque, hippocampus.
5. Drie ketens
Per familie vormen de actoren één keten.
Keten C — het patroon. Woord → vesikel → synapse → plaque → hippocampus. Dit is de keten van dementie van het Alzheimer-type.
Keten B — het materiaal. Stapelafstand → vezel → lysosoom en mitochondrion → cellichaam en Lewy-body → substantia nigra. Dit is de keten van de ziekte van Parkinson.
Keten A — de werkers. Atoom en water → enzym → haarvat → arteriole → lichaam. Dit is de keten van opruiming en herstel.
Drie conclusies volgen. De stapelafstand ligt voor beide ziekten in familie B. De stilstand heeft één maat. Een stilstand in één familie wordt doorgegeven aan elke derde laag. De plaque ligt drie lagen boven de synapse. De Lewy-body ligt drie lagen boven het lysosoom. Herstel moet komen uit familie A. In elk venster van drie lagen is A het lid dat de andere twee sluit.
6. De paden waarlangs de ziekte ontstaat
Pad 1. Het woord blijft te lang liggen en zijn richting slaat om. Een ring heeft één richting. Een woord waarvan de richting op één plek omslaat, laat een open venster achter. De werkers passen er niet meer in. Het woord wordt niet gelezen. Gemeten: in de plaques van Alzheimer-hersenen is circa 85% van het eerste stuk amyloïd-β op één of beide aspartaatplaatsen omgeklapt. Dit is de meest voorkomende verandering daar.
Pad 2. De stapel sluit zonder water. Water hoort bij familie A. Een eenvormige sluiting heeft geen derde lid nodig. Gemeten: de naad tussen de twee lagen van een amyloïde stapel is volledig droog. Er ligt geen watermolecuul in. Eenmaal gesloten is hij moeilijk te herstellen.
Pad 3. Een werker is traag. Een trage werker laat de aanvoer oplopen. Gemeten: 7 tot 10% van de Parkinson-patiënten draagt een fout in het enzym dat een suiker-vet knipt. Het is de meest voorkomende genetische factor. Bij Alzheimer is de aanmaak van amyloïd-β normaal, maar de opruiming loopt circa 30% achter. Naar schatting vult dat een brein in ongeveer tien jaar.
Pad 4. Het organel stagneert.Gemeten: de genen van Parkinson liggen bij het opruimen van mitochondriën, bij zelfvertering en bij het lysosoom. Giffen die het mitochondrium blokkeren, veroorzaken de ziekte bij mens en dier.
Pad 5. De stapel reist van cel naar cel. Een stapel groeit per knoop. Een buurcel biedt knopen aan. Gemeten: jonge zenuwcellen, geïmplanteerd bij Parkinson-patiënten, droegen na 14 jaar Lewy-bodies. Het aandeel was klein: onder 1% in één telling.
Pad 6. De lezers staan op nul. Het brein is een plaats waar het lichaam niet mag ontladen wat het leest. Gemeten: de risicogenen van Alzheimer liggen vooral bij de microglia, de leescellen van het brein.
Pad 7. Buren raken uit de pas. Dit is de naad-regel op weefselschaal. Gemeten: elk procentpunt diepe slaap dat per jaar verloren gaat, gaat samen met 27% meer dementie. Diepe slaap is de staat waarin de schorscellen samen draaien. Bij muizen wordt de ruimte tussen hersencellen in slaap 60% wijder en verdwijnt amyloïd-β sneller.
Pad 8. De puls van het vat verzwakt. Dit pad behandelt paragraaf 7.
Pad 9. De randen laten binnen wat niet ontwart.Gemeten: de WHO noemt pesticiden, trichloretheen en luchtvervuiling bij Parkinson. Landarbeiders die paraquat gebruikten, hadden twee tot drie keer het risico. Bij ratten reist het verkeerd gevouwen woord via de nervus vagus van darm naar brein. Het doorknippen van die zenuw voorkomt de ziekte.
7. Het vat en het hart
De opruimketen is familie A: enzym, haarvat, arteriole, lichaam.
Het vat is de weg naar buiten. De ruimte rond de hersenvaten voert circa 60% van het amyloïd-β af, aangedreven door de hartslag van de slagaders. Gemeten: elke stap van stijvere slagaders ging samen met vier keer zo veel kans op amyloïd in het brein. De volgorde in de tijd is niet vastgesteld.
De stapel zet zich in de weg zelf.Gemeten: bij autopsie had 98,5% van de Alzheimer-hersenen amyloïd in de vatwand, tegen 73,2% van alle ouderen. Bij bloedingen in de kwabben ligt vaatamyloïd in 50 tot 57% van de gevallen onderop. Dementie en dit soort bloeding zijn één stapel op twee plaatsen.
Opruimen trekt aan de wand.Gemeten: met het antilichaam lecanemab kreeg 17,3% van de patiënten kleine bloedingen, tegen 9,0% met placebo. Vocht in het brein trad op bij 12,6%, tegen 1,7%. Dit is de doorgang door meer spanning uit paragraaf 3.
Het hart laat hetzelfde vallen zien met een ander woord. In het hart bestaat de stapel uit transthyretine, een gesloten vorm van vier delen die uit elkaar valt. Gemeten: het middel tafamidis houdt de vorm gesloten. De sterfte daalde met 30%. Dit is het enige geval waarin stapeling bij de bron wordt gestopt, door een gesloten vorm gesloten te houden.
Eén hand op de vaatspanning raakt drie plaatsen.Gemeten: streng verlagen van de bloeddruk gaf 19% minder lichte cognitieve stoornis en 30% minder hart- en vaatgebeurtenissen. Voor dementie zelf was het verschil van 17% niet significant.
8. De paden waarlangs het lichaam herstelt
Pad A. Laat het ijzer liggen. IJzer is het onderste lid van de werkersfamilie. Gemeten: een middel dat ijzer ving, verlaagde het ijzer in de substantia nigra en maakte patiënten slechter: 15,5 punten verslechtering tegen 6,3 met placebo. IJzer is nodig voor het enzym dat dopamine maakt.
Pad B. Herstel de richting.Gemeten: het lichaam draagt zelf een enzym dat omgeklapte plaatsen terugzet. Een tweede enzym knipt naast de omgeklapte vorm. Beide zijn nog niet als behandeling geprobeerd.
Pad C. Help de werker; rem de aanvoer niet af.Gemeten: de hulpstof ambroxol verhoogde de activiteit van het trage enzym in cellen van patiënten. De rem, een middel dat de aanvoer verlaagt, gaf geen voordeel bij Parkinson. De rem werkt waar één lijn zijn maat verloren heeft. Hier was de werkersfamilie zelf vertraagd.
Pad D. Vernieuw het organel.Gemeten: beweging schakelt zelfvertering in, die amyloïd-β en α-synucleïne opruimt. Bij muizen met Parkinson verhoogde vasten om de dag de opruiming, verlaagde de stapel en behield de dopaminercellen. Dit is dierbewijs. In één andere muizenstudie steeg amyloïd binnenshuis met vasten.
Pad E. Lees.Gemeten: drie mannen van 68, 76 en 82 met een hartstapel herstelden vanzelf. De scans toonden de stapel weg. Hun lichamen hadden antilichamen gemaakt die precies op de stapel bonden. Bij 350 andere patiënten werden zulke antilichamen niet gevonden. Het waren er drie op 1.663. Of de antilichamen het herstel gaven, is niet bewezen. De nul werd een lezing. Het lichaam kan dit. Het doet het zelden.
Pad F. Weef een nieuwe triple. Een ring groeit per drie. Gemeten: in de hippocampus verschijnen nieuwe zenuwcellen bij gezonde mensen tot in het negende decennium. Bij Alzheimer valt dit sterk terug. Eén jaar duurtraining liet de hippocampus van ouderen met 2% groeien, tegen 1,4% krimping in de controlegroep. Het geheugen verbeterde.
Pad G. Omleiden.Gemeten: ongeveer een derde van de ouderen met een heldere geest draagt de volledige stapel bij autopsie. Hun cellen en verbindingen waren opvallend gespaard. Bij Parkinson ontstaan de klachten pas als ongeveer de helft van de cellen en 70 tot 80% van de dopamine weg is. De stapel is niet de ziekte. De ziekte begint waar het omleiden ophoudt.
Pad H. Breng de pas terug.Gemeten: licht en geluid met 40 slagen per seconde gaven in een kleine proef van zes maanden 76% minder achteruitgang. Een grotere proef met 670 deelnemers eindigde in augustus 2026. De uitkomst is nog niet gepubliceerd. Diepe hersenstimulatie doorbreekt de eenvormige slag bij Parkinson en vermindert beven, stijfheid en traagheid. Zij stopt het proces niet.
Pad I. Houd de koppelingen vast.Gemeten: hoortoestellen vertraagden de achteruitgang met 48% bij ouderen met hoog risico. In de hele groep was er geen effect. Een programma van beweging, voeding, denkwerk, gezelschap en bloeddruk verbeterde bij 2.111 ouderen het denkvermogen over twee jaar. Veertien factoren samen verklaren naar schatting 45% van dementie.
9. Wat volgt
De eerste stap is tijd. Een woord dat langer ligt dan zijn richting houdt, wordt onleesbaar. Alles wat het verblijf van het woord verkort, werkt vóór het bestaan van de stapel: de nacht, de slag van het vat, de werker die knipt.
De tweede stap is fase. Een stapel kan alleen beginnen waar buren uit de pas lopen. Diepe slaap, een gedeelde dag en een lichaam onder belasting zijn staten waarin buren samen draaien.
Opruimen is de laatste stap, en die heeft een prijs. Het verwijderen van de stapel herstelt het lezen. Het weeft het verloren weefsel niet terug. Met lecanemab vertraagde de achteruitgang met 27%, een verschil van 0,45 punt op een 18-puntsschaal. De doorgang door meer spanning toonde zich als bloedingen.
Herstel komt uit de werkersfamilie. De twee mislukte proeven namen iets uit familie A weg of vertraagden het. De geslaagde proeven gaven familie A ruimte: het vat, de helper van het enzym, de beweging.
Er is een venster van tien jaar. Mensen die hun dromen uitleven in slaap lopen jaren vooruit. De kans op een neurodegeneratieve ziekte is 33,5% na vijf jaar, 82,4% na 10,5 jaar en 96,6% na 14 jaar. Meestal ontwikkelen zij Parkinson (43%) of dementie met Lewy-bodies (25%). In dat venster zijn de cellen er nog.
De mensen. Een mens met dementie verliest, maand na maand, het patroon dat anderen herkennen. Een mens met Parkinson verliest de maat van de eigen beweging. Om hen heen staat een huishouden dat de dag voor twee in de pas houdt. De herstelpaden die bijna niets kosten, zijn de paden die een huishouden kan dragen: de nacht, de wandeling, de maaltijd, het gehoor, het gezelschap.
10. Status van dit essay
Uit de theorie. De sluitingsregel, de negen gesloten vensters, de twee vormen, de vaste richting, groei per drie, de maatstok van 45,6 nm, de lagen en de drie families.
Hier geteld. De 54 routes. De negen naden. Groei per één tegenover groei per drie. De ademende naad bij 4/3. De doorgang 1 → 2 → 1.
Hier berekend. De posities op de ladder en de drie ketens.
Hier gelezen. De amyloïde stapel als eenvormige sluiting. De omgeklapte aspartaat als omgeslagen richting. De zenuwcel als ring die zich niet deelt. Bloedingen onder antilichaambehandeling als doorgang door twee open vensters. De werkersfamilie als familie van herstel.
Elders gemeten. Elke zin gemarkeerd met Gemeten.
Zwak. Posities nabij een half: haarvat 4,44; hippocampus 12,46; vezel −1,38. Maten zijnranges; een andere keuze verschuift een laag. Meerdere metingen zijn verbanden en leggen geen volgorde in de tijd vast.
Open. Of “elke derde laag draagt dezelfde toestand” in weefsel meetbaar is. Waarom drie mannen hun stapel gingen lezen en 1.660 niet. Wat herstel van de vaatpuls met de opruiming doet.
Geannoteerde referentielijst
Het net
Konstapel, J. (2026). Life and Consciousness from One Rule: Closure in the Net. Constable Research, Leiden, 29 september 2026. — Bron van de sluitingsregel, de twee vormen die nooit mengen, groei per drie en deling bij lengte zes. Leesadvies: de delen over chemie en leven.
Konstapel, J. (2026). One Direction or None: Mirror Molecules, the 2026 Nobel Prize in Chemistry and the Closure Rule of the Vacuum.Net. Constable Research, Leiden, 10 oktober 2026. — De vaste richting van een ring en de open naad waar twee richtingen elkaar ontmoeten. Pad 1 rust hierop. Leesadvies: paragraaf 4.
Konstapel, J. (2026). What Fits, Stays: Designer Molecules, the Closure Rule of the Vacuum.Net, and Three Simple Roads to a Medicine. Constable Research, Leiden, 10 oktober 2026. — De ladder met de maatstok van 45,6 nm, de drie families en de berekening van de rem. Leesadvies: paragrafen 3 en 4.2.
Konstapel, J. (2026). Where Illness Comes From: The Net from First Chemistry to the Human. Constable Research, Leiden. — De proef met drie uitkomsten, het woordsysteem van de cel en de wetmatige nul. Leesadvies: hoofdstukken 4, 5 en 8.
Het woord en de stapel
Mukherjee, S., Perez, K. A., e.a. (2021). “Quantification of N-terminal amyloid-β isoforms reveals isomers are the most abundant form of the amyloid-β peptide in sporadic Alzheimer’s disease.” Brain Communications 3, fcab028. — Toont dat de plaque grotendeels bestaat uit een woord dat zijn richting verloren heeft. Onderbouwt de 85% van pad 1. Leesadvies: de kwantificeringstabellen.
Nelson, R., Sawaya, M. R., Balbirnie, M., e.a. (2005). “Structure of the cross-β spine of amyloid-like fibrils.” Nature 435, 773–778. — Het atomaire beeld van de stapel: twee lagen, een droge naad, geen water. Leesadvies: de figuur van de steriele ritssluiting.
Sawaya, M. R., Sambashivan, S., Nelson, R., e.a. (2007). “Atomic structures of amyloid cross-β spines reveal varied steric zippers.” Nature 447, 453–457. — Dezelfde droge stapel bij amyloïd-β, tau, α-synucleïne en dertig andere segmenten. Eén vorm onder vele ziekten.
Mawuenyega, K. G., Sigurdson, W., Ovod, V., e.a. (2010). “Decreased clearance of CNS β-amyloid in Alzheimer’s disease.” Science 330, 1774. — Één pagina. Aanmaak normaal, opruiming 30% traag. Twaalf patiënten tegen twaalf controles; de volgorde in de tijd blijft open. Onderbouwt pad 3.
De ziekte van Parkinson
Sidransky, E., Nalls, M. A., Aasly, J. O., e.a. (2009). “Multicenter analysis of glucocerebrosidase mutations in Parkinson’s disease.” New England Journal of Medicine 361, 1651–1661. — Het trage suiker-vet-enzym als meest voorkomende genetische factor, bij 5.691 patiënten. Onderbouwt pad 3.
Giladi, N., Alcalay, R. N., Cutter, G., e.a. (2023). “Safety and efficacy of venglustat in GBA1-associated Parkinson’s disease.” The Lancet Neurology 22, 661–671. — De rem, geprobeerd bij 221 mensen, zonder voordeel. Marqueert waar het verlagen van de aanvoer de maat niet herstelt. Onderbouwt pad C.
McNeill, A., Magalhaes, J., Shen, C., e.a. (2014). “Ambroxol improves lysosomal biochemistry in glucocerebrosidase mutation-linked Parkinson disease cells.” Brain 137, 1481–1495. — De helper van de werker, in cellen van patiënten. Onderbouwt pad C.
Devos, D., Labreuche, J., Rascol, O., e.a. (2022). “Trial of deferiprone in Parkinson’s disease.” New England Journal of Medicine 387, 2045–2055. — Minder ijzer, slechtere patiënten. De duidelijkste waarschuwing tegen het wegnemen van de onderkant van de werkersfamilie. Onderbouwt pad A. Leesadvies: de beelden naast de klinische scores.
Kordower, J. H., Chu, Y., Hauser, R. A., Freeman, T. B., en Olanow, C. W. (2008). “Lewy body-like pathology in long-term embryonic nigral transplants in Parkinson’s disease.” Nature Medicine 14, 504–506. — Jonge geïmplanteerde cellen met Lewy-bodies na 14 jaar. Onderbouwt pad 5. Lees samen met het latere commentaar dat het aandeel onder 1% was.
Galbiati, A., Verga, L., Giora, E., Zucconi, M., en Ferini-Strambi, L. (2019). “The risk of neurodegeneration in REM sleep behavior disorder.” Sleep Medicine Reviews 43, 37–46. — Het venster van tien jaar in cijfers. Onderbouwt paragraaf 9.
Heng, N., Malek, N., Lawton, M. A., e.a. (2023). “Striatal dopamine loss in early Parkinson’s disease.” Movement Disorders Clinical Practice 10, 539–546. — Hoeveel weg is als de klachten beginnen, gemeten met beeldvorming: minder dan autopsiestudies suggereerden. Onderbouwt pad G.
Slaap, vat en hart
Xie, L., Kang, H., Xu, Q., e.a. (2013). “Sleep drives metabolite clearance from the adult brain.” Science 342, 373–377. — De 60% ruimere tussenruimte in slaap en de snellere opruiming, bij muizen. Onderbouwt pad 7.
Himali, J. J., Baril, A.-A., Cavuoto, M. G., e.a. (2023). “Association between slow-wave sleep loss and incident dementia.” JAMA Neurology 80, 1326–1333. — Zeventien jaar follow-up; 27% meer dementie per punt diepe slaap die per jaar verloren gaat. Onderbouwt pad 7.
Hughes, T. M., Kuller, L. H., Barinas-Mitchell, E. J. M., e.a. (2014). “Arterial stiffness and β-amyloid progression in nondemented elderly adults.” JAMA Neurology 71, 562–568. — De stijve slagader naast het groeiende amyloïd. Klein onderzoek; wat eerste kwam, blijft onbeslist. Onderbouwt pad 8.
Jäkel, L., De Kort, A. M., Klijn, C. J. M., Schreuder, F. H. B. M., en Verbeek, M. M. (2022). “Prevalence of cerebral amyloid angiopathy: a systematic review and meta-analysis.” Alzheimer’s & Dementia 18, 10–28. — De vaatstapel geteld: bij gezonde ouderen, bij bloedingen, bij kwabbloedingen. Onderbouwt paragraaf 7.
Attems, J., Lauda, F., en Jellinger, K. A. (2008). “Unexpectedly low prevalence of intracerebral hemorrhages in sporadic cerebral amyloid angiopathy.” Journal of Neurology 255, 70–76. — 2.060 autopsies. Vaatamyloïd bij 98,5% van de Alzheimer-hersenen. Onderbouwt paragraaf 7.
Van Dyck, C. H., Swanson, C. J., Aisen, P., e.a. (2023). “Lecanemab in early Alzheimer’s disease.” New England Journal of Medicine 388, 9–21. — Opruiming en prijs in één proef: 27% tragere achteruitgang, 17,3% bloedingen. Onderbouwt paragraaf 7 en 9.
SPRINT MIND Investigators (2019). “Effect of intensive vs standard blood pressure control on probable dementia.” JAMA 321, 553–561. — Eén hand op het vat, gemeten in de geest. Onderbouwt paragraaf 7.
Maurer, M. S., Schwartz, J. H., Gundapaneni, B., e.a. (2018). “Tafamidis treatment for patients with transthyretin amyloid cardiomyopathy.” New England Journal of Medicine 379, 1007–1016. — Een gesloten vorm gesloten gehouden, en 30% minder sterfte. Onderbouwt paragraaf 7.
Herstel
Fontana, M., Gilbertson, J., Verona, G., e.a. (2023). “Antibody-associated reversal of ATTR amyloidosis-related cardiomyopathy.” New England Journal of Medicine 388, 2199–2201. — Drie pagina’s. Drie harten die hun eigen stapel opruimden. Het sterkste bewijs dat het lichaam kan lezen wat het ongelezen liet. Onderbouwt pad E.
Erickson, K. I., Voss, M. W., Prakash, R. S., e.a. (2011). “Exercise training increases size of hippocampus and improves memory.” PNAS 108, 3017–3022. — Groei van 2% tegen krimping van 1,4%, bij 120 ouderen. Onderbouwt pad F.
Moreno-Jiménez, E. P., Flor-García, M., Terreros-Roncal, J., e.a. (2019). “Adult hippocampal neurogenesis is abundant in neurologically healthy subjects and drops sharply in patients with Alzheimer’s disease.” Nature Medicine 25, 554–560. — Nieuwe zenuwcellen tot in het negende decennium van het leven. Onderbouwt pad F.
Gómez-Isla, T., en Frosch, M. P. (2022). “Lesions without symptoms: understanding resilience to Alzheimer disease neuropathological changes.” Nature Reviews Neurology 18, 323–332. — De mensen met de volledige stapel en een heldere geest, en wat er bij hen gespaard bleef. Onderbouwt pad G.
Szegő, É. M., Höfs, L., Antoniou, A., e.a. (2025). “Intermittent fasting reduces alpha-synuclein pathology and functional decline in a mouse model of Parkinson’s disease.” Nature Communications 16, 4470. — Opruiming aangezet door de ritme van de maaltijd. Alleen muizen. Onderbouwt pad D.
Hajós, M., Boasso, A., Hempel, E., e.a. (2024). “Safety, tolerability, and efficacy estimate of evoked gamma oscillation in mild to moderate Alzheimer’s disease.” Frontiers in Neurology 15, 1343588. — De kleine proef met 40 slagen per seconde. Wacht op de proef met 670 deelnemers voordat hierop wordt voortgebouwd. Onderbouwt pad H.
Lin, F. R., Pike, J. R., Albert, M. S., e.a. (2023). “Hearing intervention versus health education control to reduce cognitive decline in older adults with hearing loss in the USA (ACHIEVE).” The Lancet 402, 786–797. — Een koppeling hersteld. Lees beide uitkomsten: geen effect in de hele groep, 48% effect in de risicogroep. Onderbouwt pad I.
Baker, L. D., Espeland, M. A., Whitmer, R. A., e.a. (2025). “Structured vs self-guided multidomain lifestyle interventions for global cognitive function: the US POINTER randomized clinical trial.” JAMA. — De fundamenten samen, bij 2.111 mensen over twee jaar. Onderbouwt pad I.
Livingston, G., Huntley, J., Liu, K. Y., e.a. (2024). “Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission.” The Lancet 404, 572–628. — De veertien factoren en de 45%. De figuur gaat uit van de aanname dat elke factor de ziekte werkelijk veroorzaakt. Onderbouwt pad I.
De Nobelprijs voor Scheikunde 2026 gaat naar Henri Kagan (95) en Kenso Soai (76). Het onderwerp is het spiegelbeeldmolecuul. Sommige moleculen bestaan in twee vormen, zoals een linker- en een rechterhand. Smeltpunt, oplosbaarheid en reactiviteit zijn gelijk. Toch maakt het lichaam onderscheid.
Het leven gebruikt maar één vorm. Eiwitten bestaan uit linksdraaiende aminozuren. DNA is gebouwd op rechtsdraaiende suikers. Een reageerbuis maakt beide vormen, half om half.
De vraag staat open sinds Pasteur in 1848: hoe wint één vorm? De hoogleraar die NRC citeert noemt de puzzel “nog niet honderd procent opgelost”.
2. De diagnose
De scheikunde behandelt de eenzijdigheid van het leven als een raadsel dat een aparte verklaring nodig heeft. Dat is de vergissing. Wie telt in plaats van verklaart, ziet dat er niets te kiezen valt. Een draad die van richting wisselt, blijft niet bestaan.
Dat volgt uit één regel van de Vacuum.Net-theorie.
3. De regel
Het beeld is een visnet in water. Eén streng kruist zichzelf. Op elk kruispunt heeft de streng één van drie standen: +1, 0 of −1.
Drie kruispunten naast elkaar vormen een venster. Het venster is gesloten als de drie standen samen nul zijn, geteld in drieën. Een open venster staat onder spanning. Alleen wat sluit, blijft.
Drie standen op drie plaatsen geven 27 mogelijkheden. Daarvan sluiten er 9:
drie gelijke standen: 3 vensters. Dit is de vorm van een kristal. Er verandert niets.
drie verschillende standen: 6 vensters. Dit is de vorm van stofwisseling. Alles verandert en het venster blijft dicht.
De andere 18 staan open.
4. Het spiegelpaar
Neem een ring van drie verschillende standen. Een ring heeft geen begin. Zes volgordes gedeeld door drie draaiingen geeft twee ringen:
Ring L: − → 0 → +
Ring R: + → 0 → −
Keer elk teken om en ring L wordt ring R. Het zijn elkaars spiegelbeeld. Hun inhoud is gelijk: één min, één nul, één plus. Alleen de looprichting verschilt.
Dat is het spiegelbeeldmolecuul. Gelijke inhoud, dus gelijke eigenschappen. Andere richting, dus een andere pasvorm.
5. Waarom mengen niet kan
In een draad met drie verschillende standen leggen twee standen de derde vast. Na −, 0 komt +. Na 0, + komt −. De richting ligt na twee stappen vast en kan niet meer omslaan.
Forceer het toch. Zet ring L aan ring R:
− , 0 , + , + , 0 , −
Venster
Som
Stand
(−, 0, +)
0
gesloten
(0, +, +)
2
open
(+, +, 0)
2
open
(+, 0, −)
0
gesloten
Op de naad staan twee vensters open. Een gemengde draad staat op elke naad onder spanning. Hij houdt het niet.
De uitkomst is hard. Een levende draad heeft één richting of hij bestaat niet. Wat de scheikunde homochiraliteit noemt, is de enige gesloten vorm.
6. Wat Kagan vond
Een katalysator vult het venster samen met het molecuul dat ontstaat. Twee standen zijn van de katalysator. Eén stand is vrij. Dan is er precies één waarde die sluit.
Een linkse katalysator sluit dus alleen met een links product. Het rechtse product laat het venster open en valt terug.
Kagan vond nog iets. Een onzuivere katalysator geeft een zuiverder product dan verwacht. Het net telt dat zo uit. Neem 60 linkse en 40 rechtse ringen. Elke rechtse ring ligt tegen een linkse aan, op een open naad. Die 40 paren doen niets. Er blijven 20 linkse ringen vrij. Al het werk komt van die twintig, en al het werk is links.
7. Wat Soai vond
Een ring van drie kan niet met één kruispunt groeien. Op lengte 4 en 5 bestaat geen gesloten ring van dit type. Hij groeit met een heel drietal, naar zes. Een ring van zes kan in tweeën. Beide helften sluiten. Beide houden de richting van de ouder.
3 → 6 → 3 + 3
Dat is de reactie van Soai. Het product maakt zichzelf na, met dezelfde richting.
De groei laat zich uittellen. Begin met 51 linkse en 49 rechtse ringen. De 49 rechtse liggen tegen 49 linkse aan en delen niet. Het overschot is 2. Dat verdubbelt per ronde.
Ronde
Links
Rechts
Aandeel links
0
51
49
51,00 %
5
113
49
69,75 %
10
2.097
49
97,72 %
15
65.585
49
99,93 %
18
524.337
49
99,99 %
Achttien verdubbelingen brengen een overschot van twee op honderd naar 99,99 procent. Dat is de zuiverheid die Soai meet.
Welke richting wint, zegt de regel niet. De regel is symmetrisch. Het net houdt de richting vast van wat er het eerst lag. Eén ring verschil is genoeg.
8. De grens tussen stof en leven
De telling zet scheikunde en leven op één ladder.
Onder lengte 3: starre orde, of spanning die wordt doorgegeven. Dat is scheikunde.
Op lengte 3: een ring die dicht blijft terwijl zijn inhoud verandert. Hij heeft een richting. Hij kan niet delen.
Op lengte 6: de ring deelt en geeft zijn richting door. Dat is de vorm van een cel.
Soai doet de stap 3 → 6 → 3 + 3 in een kolf. Zonder wand, zonder cel. Het Nobelcomité noemt dat scheikunde. Op de ladder is het de eerste trede van erfelijkheid.
9. Wat het voor mensen betekent
Een lichaam is een draad met één richting. Een molecuul met dezelfde richting sluit het venster. Het past en het blijft. Het spiegelbeeld laat het venster open. Dan gebeurt één van drie dingen.
Niets. Het molecuul gaat het lichaam door. De helft van het middel is ballast.
Iets anders. Het sluit elders een ander venster. Dextropropoxyfeen werd verkocht als de pijnstiller Darvon. Het spiegelbeeld werd verkocht als het hoestmiddel Novrad. De namen spiegelen elkaar, net als de moleculen.
Schade. Het houdt een venster open dat dicht moet. Softenon was een mengsel van beide vormen. De ene kalmeerde. De andere beschadigde het ongeboren kind. Wereldwijd werden rond de tienduizend kinderen met misvormde ledematen geboren. Velen van hen leven nog.
Voor het maken van medicijnen is de les eenvoudig. Een kolf zonder richting maakt beide ringen, half om half. De helft is dan het verkeerde middel, en moet er achteraf duur worden uitgehaald. Een wand met richting, of één zaadje met de goede richting, maakt alleen de goede ring. Kagan leverde de wand. Soai leverde het zaad. Het zaad vermenigvuldigt zichzelf.
10. Wat open blijft
Aminozuren draaien links. De suikers van DNA draaien rechts. Het net kent twee richtingen en beide zijn in gebruik. Waarom de twee stelsels tegengesteld lopen, is nog niet uitgeteld.
Er ligt één lezing klaar. Een venster een laag hoger kan zelf met drie verschillende standen sluiten: één stelsel op min, één op plus en één op nul. Nul is hier: zonder draairichting. De vetten van de celwand en het aminozuur glycine hebben geen draairichting. Of zij de nul bezetten, is een telling die nog gedaan moet worden.
11. Slot
De prijs beloont twee mensen die lieten zien hoe één richting wint. De regel van het net laat zien dat er niets te winnen viel. Twee richtingen in één draad sluiten niet. Het leven draait één kant op omdat het anders niet rond komt.
Geannoteerde referenties
Frölke, S. (2026). “Een zuiver scheikundig spiegelbeeld leidt tot filosofische levensvragen. En tot een Nobelprijs.” NRC, 7 oktober 2026.Waarom lezen? De aanleiding. Het geeft de prijs, de twee winnaars en de koppeling met de oorsprong van het leven. Leesadvies: het artikel koppelt Darvon en Novrad aan dextromethorfan. Het paar is dextro- en levopropoxyfeen.
Konstapel, J. (2026). One Direction or None: Mirror Molecules, the 2026 Nobel Prize in Chemistry and the Closure Rule of the Vacuum.Net. Constable Research, Leiden, 10 oktober 2026.Waarom lezen? Het artikel onder deze blog, met alle tellingen uitgeschreven. Leesadvies: sectie 4 voor de open naad, sectie 6 voor de verdubbelingstabel.
Konstapel, J. (2026). Leven en bewustzijn uit één rekenregel in een driewaardig net. Constable Research, Leiden, 29 september 2026.Waarom lezen? De bron van de sluitingsregel, de twee soorten sluiting en de deling op lengte 6. Leesadvies: de delen over scheikunde en leven.
Konstapel, J. (2026). Why Is Our Left Brain Leading Us to Destruction? All You Want to Know About Chirality. Constable Research, Leiden, 14 januari 2026.Waarom lezen? Draairichting van knoop tot molecuul tot brein, met de klaverbladknoop als eenvoudigste vorm die van zijn spiegelbeeld verschilt. Leesadvies: de delen over homochiraliteit en softenon.
Soai, K., Shibata, T., Morioka, H., en Choji, K. (1995). “Asymmetric autocatalysis and amplification of enantiomeric excess of a chiral molecule.” Nature 378, 767–768.Waarom lezen? De reactie zelf, in twee pagina’s: een product dat zichzelf en zijn draairichting vermenigvuldigt. Leesadvies: leg de tabel met zuiverheid per ronde naast de tabel in sectie 7.
Puchot, C., Samuel, O., Duñach, E., Zhao, S., Agami, C., en Kagan, H. B. (1986). “Nonlinear effects in asymmetric synthesis.” Journal of the American Chemical Society 108, 2353–2357.Waarom lezen? De meting dat een onzuivere katalysator een zuiverder product geeft. Het is de proef bij de geparkeerde paren uit sectie 6.
Frank, F. C. (1953). “On spontaneous asymmetric synthesis.” Biochimica et Biophysica Acta 11, 459–463.Waarom lezen? Het model dat de reactie van Soai tweeënveertig jaar eerder voorspelde: een molecuul dat zichzelf kopieert en zijn spiegelbeeld blokkeert. Leesadvies: vier pagina’s; vergelijk Franks blokkade met de open naad.
Pasteur, L. (1848). “Sur les relations qui peuvent exister entre la forme cristalline, la composition chimique et le sens de la polarisation rotatoire.” Annales de Chimie et de Physique 24, 442–459.Waarom lezen? De ontdekking van het spiegelpaar, gedaan door kristallen met de hand te sorteren onder een loep.
Blackmond, D. G. (2010). “The origin of biological homochirality.” Cold Spring Harbor Perspectives in Biology 2, a002147.Waarom lezen? Het gangbare overzicht van het raadsel en de voorgestelde antwoorden. Het laat zien wat het vak zelf nog open houdt.
Maynard Smith, J., en Szathmáry, E. (1995). The Major Transitions in Evolution. Oxford: Freeman.Waarom lezen? De ladder van zichzelf kopiërende moleculen naar cellen, waar de lengtes 3 en 6 naast gelegd kunnen worden. Leesadvies: hoofdstuk 3 en 4.
Brynner, R., en Stephens, T. (2001). Dark Remedy: The Impact of Thalidomide and Its Revival as a Vital Medicine. Cambridge, MA: Perseus.Waarom lezen? De geschiedenis van softenon: de kinderen, de fabrikant, de toezichthouders en de terugkeer van het middel. Leesadvies: de eerste helft, over 1957–1962.
Designer Molecules, the Closure Rule of the Vacuum.Net, and Three Simple Roads to a Medicine
Beeld ter illustratie, niet de designermolecuul van Overkleeft en zijn team.Bron beeld RamonFS/Getty Images
J.Konstapel,Leiden,10-10-2026.
Aanleiding:Het Parool: van 9 oktober 2026, “Door in Leiden te ‘prutsen’ met moleculen komt een medicijn voor ernstige ziektes een stuk dichterbij”. De onderbouwing staat in het Engelse artikel What Fits, Stays (zie referenties).
De Leidse chemicus Hermen Overkleeft kreeg op 7 oktober de Spinozapremie van € 1,5 miljoen. Hij maakt al 25 jaar kleine suikermoleculen met een kleine aanpassing. In zijn kast staan er 7.000, gedroogd, in potjes. Wie mailt, krijgt een beetje opgestuurd tegen verzendkosten.
Uit dat werk komen nu drie medicijnen voort. Eén ervan zit in een late testfase en kan eind 2027 in gebruik zijn.
De krant brengt het als een succesverhaal van geduld. Dat is het ook. Maar er staat meer in. In het artikel staan vier vondsten die samen laten zien dat medicijnen maken anders kan dan het nu gebeurt. Eenvoudiger, en met minder omweg.
De vier vondsten
Een molecuul dat leest. Een markermolecuul blijft plakken aan één bepaald enzym. Zo is dat enzym live te volgen in de cel. De cel merkt er niets van.
Een molecuul dat afremt. Bij sommige erfelijke stofwisselingsziekten werkt één enzym te traag. De rest van de cel werkt door. Het gevolg is ophoping. Overkleeft repareert het trage enzym niet. Hij remt de twee enzymen die ervoor staan. De hele lijn gaat langzamer. De klachten nemen af.
Een molecuul dat uitschakelt. Andere moleculen zetten een schadelijk enzym helemaal stil. Dat is de basis voor middelen tegen sommige bloedkankers en auto-immuunziekten.
Een molecuul dat zelf vond. Eén molecuul was gemaakt voor een heel ander eiwit. Het bleef ook plakken aan een enzym waar kankeronderzoekers al tientallen jaren naar zochten. Het lichaam heeft 20.000 soorten eiwitten. Niemand had dit ontworpen.
Wat er werkelijk gebeurt
De vier vondsten zijn te lezen met één regel uit de Vacuum.Net-theorie. Het beeld is een visnet. Draden kruisen in knopen. Elke knoop heeft één van drie waarden: +1, 0 of −1. Drie knopen naast elkaar vormen een venster. Een venster is gesloten als de drie waarden samen een veelvoud van drie zijn. Een open venster draagt spanning.
Van de 27 mogelijke vensters zijn er 9 gesloten. Drie zijn gelijkvormig: drie keer dezelfde waarde. Dat is de vorm van een kristal. Zes zijn verschillend: +1, 0 en −1 in een of andere volgorde. Die vorm kan blijven draaien en toch gesloten blijven. Dat is de vorm van leven.
Met die regel krijgen de vier vondsten hun plaats.
De marker is een meting. Het lichaam toetst zichzelf de hele dag. Past iets, dan blijft het. Past het niet, dan wordt het opgeruimd. Wordt het niet gelezen, dan gebeurt er niets. In dat derde geval zit ziekte. Een traag enzym wordt niet opgeruimd, want niets leest zijn tempo. De marker maakt het leesbaar.
De rem herstelt de maat. De rekensom is eenvoudig. De aanvoer biedt elke stap één venster aan. Het trage enzym sluit er één per drie stappen. Elke stap blijft dus twee derde venster open. Na 300 stappen liggen er 200 open vensters. Dat is de ophoping waar de patiënt ziek van wordt. Rem de aanvoer af tot één venster per drie stappen, en de stapel groeit niet meer. Het enzym is nog even traag. Alleen de maat klopt weer.
Dat Overkleeft twee enzymen remt, past in de regel. Staan twee waarden van een venster vast, dan is er precies één waarde die het sluit.
De uitschakelaar is een vaste knoop. Een vaste knoop sluit in een draaiend venster één op de drie stappen. De buren stoppen met draaien. Dat is het verkeerde middel waar iets levends moet blijven draaien. Het is het juiste middel waar het draaiende zelf de afwijking is.
De toevalstreffer is geen toeval. In het net blijft alleen wat past. Het molecuul paste, dus het bleef. De cel deed in één keer de zoektocht die het ontwerpen in tientallen jaren niet had opgeleverd.
Drie eenvoudige wegen
1. Eerst passen, dan ontwerpen
De gangbare weg kiest eerst een doelwit. Daarna wordt er één sleutel voor dat ene slot ontworpen. Elk paar wordt apart getest.
Tel de paren. 7.000 moleculen tegen 20.000 eiwitten geeft 140 miljoen paren.
De omgekeerde weg biedt één sleutel aan levend weefsel aan. In dat weefsel zitten alle 20.000 sloten tegelijk. De marker laat zien waar de sleutel bleef. Eén proef dekt 20.000 paren. De hele kast kost 7.000 proeven, geen 140 miljoen.
De middelen zijn er. De sleutels staan in Leiden in potjes. Ze gaan per post. Eén sleutel namaken uit de publicaties kost ongeveer een jaar. Wat ontbreekt is een kaart: welke sleutel sluit waar, in welk weefsel. Dat is een tabel van metingen. Het is klein werk naast een ontwerpprogramma.
2. Eerst de maat, dan de stof
De rem werkt via tempo. Tempo is op meer plaatsen in te stellen.
Komt de aanvoer uit voedsel, dan is de maaltijd de rem. Bij de stofwisselingsziekte PKU gebeurt dat al sinds 1953. Het enzym blijft levenslang traag. De aanvoer wordt aan tafel verlaagd. Het kind groeit gezond op.
Maakt het lichaam de aanvoer zelf, dan is een molecuul nodig. Dat geldt voor de ziekten uit het artikel.
De volgorde ligt daarmee vast. Eerst de dag: licht en donker, slaap, het tijdstip van de maaltijd, heel voedsel. Pas daarna de stof, op de plek waar de dag niet bij kan.
3. Eerst de toestand, dan de stof
De theorie zegt dat een maas de toestand van zijn omgeving zet. Het tempo volgt. Daarvoor bestaan drie metingen.
Een atoom in een kooi van zestig koolstofatomen vervalt 0,83% sneller.
Een scheikundige reactie tussen twee spiegels verliep tot 5,5 keer trager. Er werd geen licht toegevoerd.
Het enzym pepsine werkte tussen twee spiegels ongeveer 4,5 keer trager.
Leg de laatste meting naast de rekensom van de rem. Daar was een factor drie nodig. Gemeten is 4,5. De grootte klopt.
Op deze weg is het medicijn geen molecuul. Het is een maas: een structuur die het tempo zet zonder iets toe te voegen. Er wordt niets geslikt. Er hoeft niets te worden afgebroken.
Deze weg is niet af. De metingen zijn gedaan in een laboratoriumopstelling, niet in weefsel. Herhalingen van eerdere metingen gaven wisselende uitkomsten. Eén getal staat open: hoeveel een maas het tempo van een enzym in een levende cel verschuift.
Voor wie
De ziekten uit het artikel zijn zeldzaam en erfelijk. Ze treffen vooral jonge kinderen, in de hersencellen. Een kind verliest maand na maand wat het had geleerd. Ouders leven jaren met een naam voor de ziekte en zonder behandeling.
Voor hen telt de eerste weg het zwaarst. Die verkort het vinden. De tweede weg telt voor de veel grotere groep bij wie de maat verloren ging door de dag die zij leven, niet door een erfelijk enzym.
Het toetsen in mensen blijft. Dat is het trage en dure deel. De drie wegen verkorten de zoektocht die eraan voorafgaat. De zorgvuldigheid waarmee een kind iets nieuws krijgt, vervangen zij niet.
Wat Leiden al heeft
Overkleeft nam een ongewoon besluit. Alle moleculen worden gedeeld. Een onderzoeker uit een arm land krijgt hetzelfde monster als een onderzoeker uit een rijk land. Soms betaalt Leiden ook de verzending.
Dat is meer dan vrijgevigheid. Passen vraagt veel pogingen in veel weefsels. Een sleutel die gedeeld wordt, wordt vaker geprobeerd dan een sleutel die iemand bezit. Delen is de voorwaarde voor de eerste weg.
De kast met 7.000 potjes is daarmee een bibliotheek. De marker is het leesinstrument. Wat ontbreekt, is de kaart.
Referenties
Konstapel, J. (2026). What Fits, Stays: Designer Molecules, the Closure Rule of the Vacuum.Net, and Three Simple Roads to a Medicine. Constable Research, Leiden, 10 oktober 2026.Waarom lezen? De volledige onderbouwing van deze blog, met alle berekeningen uitgeschreven: de plaats van suiker, DNA, enzym en cel op de ladder, de rekensom van de rem en de telling van de paren. Leesadvies: deel 3 voor de ladder, deel 5 voor de drie wegen, deel 7 voor wat geteld, gelezen en open is.
Van Agteren, P. (2026). “Door in Leiden te ‘prutsen’ met moleculen komt een medicijn voor ernstige ziektes een stuk dichterbij.” Het Parool, 9 oktober 2026.Waarom lezen? De aanleiding. De vier vondsten in de eigen beelden van de onderzoeker: de assemblagehal, de flessenfabriek, de potjes in de kast. Leesadvies: de passage over de rem en de passage over de toevalstreffer.
Konstapel, J. (2026). Eli Lilly in Valkenburg: wat de dieetpillenfabriek echt doet. constable.blog, 1 oktober 2026.Waarom lezen? Legt de rekenregel uit die hier wordt gebruikt, en laat zien wat een vaste knoop doet in een levend net. Deze blog is het tegenstuk: een rem die de maat herstelt tegenover een pil die het eigen signaal vervangt. Leesadvies: het deel “Een andere manier van kijken: sluiting”.
Konstapel, J. (2026). Waarom ziekte geen raadsel is. constable.blog, 16 september 2026.Waarom lezen? De toets met drie uitkomsten, en ziekte als wat niet gelezen wordt. De marker uit deze blog krijgt daar zijn plaats. Leesadvies: het deel “Eén net, één toets”.
Konstapel, J. (2026). Transmutation: Chemical Elements on Demand. constable.blog, 9 oktober 2026.Waarom lezen? De stelling dat een maas de toestand zet en het tempo volgt, met de metingen gescheiden naar hun status. De derde weg rust erop. Leesadvies: paragraaf 6.2 over de koolstofkooi.
Witte, M. D., Kallemeijn, W. W., Aten, J., e.a. (2010). “Ultrasensitive in situ visualization of active glucocerebrosidase molecules.” Nature Chemical Biology 6, 907–913.Waarom lezen? Een marker zoals in deze blog beschreven: hij blijft aan één enzym plakken en maakt het zichtbaar in de levende cel. Leesadvies: eerst de afbeeldingen.
Cravatt, B. F., Wright, A. T. & Kozarich, J. W. (2008). “Activity-based protein profiling: from enzyme chemistry to proteomic chemistry.” Annual Review of Biochemistry 77, 383–414.Waarom lezen? De algemene methode achter de eerste weg: één molecuul aanbieden aan een hele cel en aflezen waar het blijft.
Wu, L., Jiang, J., Jin, Y., e.a. (2017). “Activity-based probes for functional interrogation of retaining β-glucuronidases.” Nature Chemical Biology 13, 867–873.Waarom lezen? Een gepubliceerd geval uit het Leidse lab waarin moleculen voor één groep enzymen ook bleven plakken aan heparanase, een enzym dat samenhangt met uitzaaiing. De krant noemt het enzym niet; dit artikel laat zien wat voor soort vondst bedoeld is.
Platt, F. M., Neises, G. R., Dwek, R. A. & Butters, T. D. (1994). “N-butyldeoxynojirimycin is a novel inhibitor of glycolipid biosynthesis.” Journal of Biological Chemistry 269, 8362–8365.Waarom lezen? De eerste formulering van het idee van de rem: repareer het trage enzym niet, vertraag de aanvoer. Vier pagina’s.
Cox, T., Lachmann, R., Hollak, C., e.a. (2000). “Novel oral treatment of Gaucher’s disease with N-butyldeoxynojirimycin to decrease substrate biosynthesis.” The Lancet 355, 1481–1485.Waarom lezen? De rem bij patiënten, mede uit Amsterdam. Laat zien wat het herstellen van de maat voor een mens betekent en hoe lang het duurt.
Bickel, H., Gerrard, J. & Hickmans, E. M. (1953). “Influence of phenylalanine intake on phenylketonuria.” The Lancet 262, 812–813.Waarom lezen? Twee pagina’s. Eén kind, één dieet, en het bewijs dat de maaltijd de rem kan zijn. Het begin van de tweede weg.
Ohtsuki, T., Yuki, H., Muto, M., Kasagi, J. & Ohno, K. (2004). “Enhanced electron-capture decay rate of ⁷Be encapsulated in C₆₀ cages.” Physical Review Letters 93, 112501.Waarom lezen? Een kooi verandert een tempo met 0,83%. Klein, schoon en onomstreden.
Thomas, A., George, J., Shalabney, A., e.a. (2016). “Ground-state chemical reactivity under vibrational coupling to the vacuum electromagnetic field.” Angewandte Chemie International Edition 55, 11462–11466.Waarom lezen? De reactie die tussen twee spiegels tot 5,5 keer trager verliep. Leesadvies: de figuur met tempo tegen spiegelafstand; het effect verschijnt alleen bij de juiste afstand.
Vergauwe, R. M. A., Thomas, A., Nagarajan, K., e.a. (2019). “Modification of enzyme activity by vibrational strong coupling of water.” Angewandte Chemie International Edition 58, 15324–15328.Waarom lezen? De enige meting aan een enzym: pepsine, ongeveer 4,5 keer trager. Het getal waaraan de derde weg moet worden afgemeten.
Imperatore, M. V., Asbury, J. B. & Giebink, N. C. (2021). “Reproducibility of cavity-enhanced chemical reaction rates in the vibrational strong coupling regime.” Journal of Chemical Physics 154, 191103.Waarom lezen? Een herhaling van een eerdere spiegelmeting die het effect niet terugvond. Laat zien waarom het open getal opnieuw gemeten moet worden, in weefsel.
