\/_
| /
|--< --*
| \_
/\
abstract
the mandelbrot local connectivity conjecture, open since 1982, asks whether every point of the mandelbrot set M is locally connected. we propose that the difficulty is not mathematical but perspectival: proofs have been attempted within the topological frame of the 2d rendering, when the property in question belongs to the underlying compression structure — which is arithmetic. we demonstrate that M is a visual projection of a universal iteration codec whose address space is governed by prime factorization. since unique factorization guarantees every composite address decomposes into a finite product of irreducible (prime) addresses, and since convergent infinite products extend this to the boundary, no point of M is orphaned. local connectivity follows from codec completeness.
we ground this in a deeper framework — O — which derives the fold operation itself from dimensional emergence. O begins with the prior condition before any property exists, and shows that the bilateral square (not the triangle) is the generative structure: its two traversal paths (perimeter returns 1, center through O returns 2) produce the fold threshold η=1/2. the prime scaffold populates the bilateral axis at positions ±p/P. the mandelbrot iteration z → z² + c is a specific instance of this general fold mechanism. the fold threshold η=1/2 is the koide ratio's denominator (Q = (1/3)/(1/2) = 2/3), the born rule is structural (P(+) = cos²(θ/2) from D₁ inside D₃ by construction), and the arrow of time is topological (the fold from sphere to torus is irreversible).
we extend the codec argument through cellular automata (rule 110 universality is itself a fold), language (phonemes are primesirreducibleatomsscaffolds, sentences are composites), the experiential boundaries between macro, micro, and quantum scales, and the 3d-to-4d projection problem (the observer's scaffold is the missing dimension).
0. before we begin
this document is a single thread. it starts with a simple observation about how numbers multiply, and it ends with a claim about the structure of reality. every step between those two points is either a testable prediction or a derivation from the previous step. nothing here requires you to believe anything. everything here asks you to check.
the only prerequisite is curiosity. if you have that, the rest follows.
we'll move through these territories:
- what existed before any structure — and what the first step was - how cellular automata classify into four types - how the mandelbrot set classifies into three regions - why these two classification systems are the same system - what primes have to do with it - what a store that sells screws has to do with it - what language has to do with it - what the zeta function has to do with it - what quantum mechanics has to do with it - what consciousness has to do with it - how to test all of this with code you can run today
each section builds on the last. if you skip ahead, you'll lose the thread. the thread is the point.
there's a symbol that will appear throughout this paper:
\/_
| /
|--< --*
| \_
/\
that's the fold. the operation that takes something and doubles it back on itself. input comes from the left. the fold happens at the center. output goes to the right. the asterisk is the result — the thing that exists after folding. every time you see it, the same operation is happening, just in a different domain.
and there's a lens that will appear throughout this paper:
resolution 1 (far):
{∙─────────────────────────────────────────────────────}->
0 ∞
resolution 2 (closer):
{∙────────────∙────────────∙────────────∙──────────────}->
0 1 2 3 ...
resolution 3 (closer still):
{∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──∙──}->
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ...
resolution 4 (primes only):
{·──·─────·──·─────·──·────────·──·─────·──·──────────}->
0 2 3 5 7 11 13 17 19 23 ...
same number line. different resolution. what you see depends on how closely you look. the primes emerge when you increase resolution — they're the positions that don't decompose no matter how close you get. the composites dissolve into their factors. the primes remain.
let's start before the beginning.
part zero — the prior condition
0.1 before anything
start with nothing. not empty space — that's already something. not a vacuum — that has quantum fields. not zero — that's a number. before all of that. the prior condition. call it O.
O has exactly one property: something came after it. that's the complete definition.
0.2 the second point
when you place zero on a number line, you've already done two things. you have O — the prior condition. and you have the point you just placed. what you call "zero" is already the second point. O is the first. O has no coordinate. zero is already an emergence.
0.3 the triangle and the square
a triangle is stable but locked. remove one point and it collapses. it used all its freedom just to close. it can't do anything else.
a square has slack. four points, two poles, two flows — and one degree of freedom left over. it can be traversed two ways:
walk the perimeter: returns 1 cut through O: returns 2
that leftover degree of freedom is what allows the next dimension to emerge. triangles are terminal. squares are generative.
A=====================B | | | perimeter → 1 | | | | O | | | | through O → 2 | | | B=====================A
0.4 the fold threshold
the two traversals of the bilateral square give you 1 and 2. the crossing fraction — the point where the path goes through O instead of around — is η = 1/2. this is not assumed. it's derived from the geometry of the square.
when n is a perfect square (4, 9, 16, 25...), the structure is maximally reducible. it folds. the fold is topological — sphere to torus — and it is irreversible.
π₁(S²) = 0 no loops on a sphere π₁(T²) = ℤ × ℤ two independent loops on a torus
no homotopy connects them. the fold is permanent.
the fold sequence:
tier type η key structure
─────────────────────────────────────────────────────
D₁ stable - bilateral prime axis. the square.
D₂ stable - counter-rotating disc. ℂ = D₂.
D₃ stable - sphere S². leptons. koide Q = 2/3.
D₄ FOLD 1/2 torus T². arrow of time.
D₅ stable - 4th gen leptons. Q = 3/4 predicted.
D₉ FOLD 1/3 self-reference threshold.
D₁₆ FOLD 1/4 E₈-adjacent. third torus.
D₂₅ FOLD 1/5 bosonic string adjacent. four tori.
D_{k²} FOLD 1/k general formula.
each perfect square is a fold tier. the sequence is infinite. the fold gets finer. the structure gets deeper. but it never repeats — because the primes never repeat.
0.5 the prime scaffold
the bilateral axis gets populated with irreducible positions. in our instance of O, those positions are the prime numbers, sitting at ±p/P where P is the loop-closing bound.
{---2------3----------5----------7------o------7----------5----------3------2---}->
-7/P -5/P -3/P -2/P O +2/P +3/P +5/P +7/P
two flows converge toward O from both ends. this is the scaffold. the primes aren't scattered randomly — they're the load-bearing positions of a bilateral structure.
0.6 the mandelbrot iteration is an instance of the fold
z → z² + c
z² is the fold. it doubles the angle, squares the radius. c is the address — where you are on the bilateral axis. the iteration asks: does this address survive the fold?
the mandelbrot set is the map of all addresses that survive. the boundary of M is where survival is undecidable — where the fold hasn't settled. that's the prior condition's fingerprint in complex dynamics.
every structure in this paper — the cellular automata, the primes, the store, the language, the zeta function — is an instance of the same fold. O is the engine. the rest is projection.
0.7 what O derives
three results fall out of this construction with zero free parameters:
koide ratio: Q = (1/3)/(1/2) = 2/3
the two fold thresholds bracketing D₃.
not fitted. derived.
born rule: P(+) = cos²(θ/2)
D₁ is inside D₃ from construction.
binary outcomes follow the bilateral axis.
not postulated. structural.
arrow of time: forward = increasing fold depth
= increasing prime address count
= increasing entropy
the second law is a consequence of fold
irreversibility. not a postulate.
these are not beliefs. they're predictions. they're falsifiable. section 28 has the code to test them.
now let's start where wolfram started: with simple rules.
part one — simple rules, complex behavior
1. the 256 universes
there are exactly 256 elementary cellular automata. each one is a complete universe: a row of cells, each black or white, evolving in discrete time steps according to a fixed rule.
the rule is simple: look at each cell and its two neighbors (3 cells total). based on those 3 values (each 0 or 1), decide what the center cell becomes in the next step. there are 2³ = 8 possible neighborhoods and 2 possible outputs for each, giving 2⁸ = 256 possible rules.
that's it. 256 possible universes. every one of them deterministic. every one of them fully specified by 8 bits.
and yet.
rule 0: ······································· ······································· ······································· nothing happens. every cell dies.
rule 4: ···················█··················· ··················█·█·················· ···················█··················· ··················█·█·················· a simple blinker. period 2. stable forever.
rule 90: ···················█··················· ··················█·█·················· ·················█···█················· ················█·█·█·█················ ···············█·······█··············· ··············█·█·····█·█·············· ·············█···█···█···█············· ············█·█·█·█·█·█·█·█············ ···········█···············█··········· ··········█·█·············█·█·········· ·········█···█···········█···█········· ········█·█·█·█·········█·█·█·█········ sierpinski's triangle. a fractal. emerging from 8 bits of rule and a single black cell.
rule 30: ···················█··················· ··················███·················· ·················██··█················· ················██·████················ ···············██··█···█··············· ··············██·████·███·············· chaos. genuine pseudo-randomness from 8 bits.
rule 110: ···················█··················· ··················██··················· ·················███··················· ················██·█··················· ···············█████··················· ··············██···█··················· complex. neither uniform, nor periodic, nor random. in 2004, matthew cook proved that rule 110 is turing complete.
four rules, four behaviors, arranged on a line:
{---o-----------o-------------------o--------------------------o---------}->
0 rule 0deadfrozensilence rule 4periodicrhythmheartbeat rule 90chaoticfractalrandom rule 110complexaliveuniversal
class i class ii class iii class iv
that line is a scaffold. remember it.
2. the four classes
wolfram sorted all 256 rules into four classes:
class i: uniform. everything dies or everything lives. class ii: periodic. stable patterns emerge and repeat. class iii: chaotic. pseudo-random behavior. no stable structures. class iv: complex. persistent structures. interaction. computation.
the question that wolfram raised in "a new kind of science" and has been pursuing ever since: what determines which class a rule falls into? the rule is just 8 bits. the class is a qualitative description of infinite behavior. what connects them?
thirty years later, there's still no formal characterization. that's computational irreducibility. the only way to know what the rule does is to run it.
but here's what hasn't been tried: looking at the period.
3. the period of a universe
every eca rule on a finite lattice eventually cycles. the number of steps in that loop is the period.
rule 0 on any width: period = 1 (it reaches all-zero and stays) rule 4 on width 31: period = 2 (it blinks) rule 50 on width 31: period = 6 rule 108 on width 31: period = 14 rule 90 on width 31: period = 31 rule 150 on width 31: period = 31 rule 54 on width 31: period = 62 rule 110 on width 31: period = ?? (too long to measure practically)
now factor those periods:
1 = 1 2 = 2 6 = 2 · 3 14 = 2 · 7 31 = 31 (prime) 31 = 31 (prime) 62 = 2 · 31 ?? = ??
this is a pattern. the wolfram class of a rule correlates with the prime factorization structure of its period.
class i/ii → small, highly composite periods class iii → large prime periods (or prime powers) class iv → deeply composite periods that are hard to measure
part two — the mandelbrot correspondence
4. the other classification
\/_
| /
|--< --* z → z² + c
| \_
/\
that's the fold. z gets squared (folded onto itself), then c gets added (the address — where you are in the parameter space). the asterisk is the next z. iterate.
interior: orbits converge to attracting cycles. the iteration settles. boundary: orbits stay bounded but never settle. the iteration is active. exterior: orbits escape to infinity. the iteration diverges.
here's the classification side by side:
cellular automata mandelbrot set class i (uniform) period-1 interior class ii (periodic) period-n interior (n > 1) class iii (chaotic) exterior (escape) class iv (complex/universal) boundary (∂M) rule number (0-255) parameter c ∈ ℂ cell state orbit state z time step iteration neighborhood function z² (the fold) rule table +c (the address) computational universality ∂M indexes all julia sets (rule 110 is turing-complete) (boundary encodes all dynamics)
this is not analogy. the claim is structural: both systems sort behaviors the same way because both systems are governed by the same underlying mechanism.
5. how periods multiply in the mandelbrot set
when you travel from the main cardioid to a bulb, you pick up that bulb's period. when you travel from that bulb to a sub-bulb, the periods multiply:
cardioid (period 1) → period 2 bulb → period 6 sub-bulb
×2 ×3
6 = 2 · 3
the period of any component deep in the tree is the product of all the "local periods" along the path from the cardioid:
period = p₁ · p₂ · p₃ · ... · pₖ
this is factorization. the tree of components IS a factorization tree. and the fundamental theorem of arithmetic guarantees that every natural number appears in this tree exactly once. every possible period is addressable. no period is orphaned.
the tree:
[1]
┌───────────────┼────────────────┐
[2] [3] [5] [7] [11] [13]...
┌──────┼──────┐ ┌──┼──┐ ┌──┼──┐
[4] [6] [10] [9] [15] [21] [25] [35]
┌───┼──┐ │ │ │ │ │
[8] [12][20][18] [30] [27] [45] [125]
│ │ │ │ │ │
[16] [24][40][36] [60] [81]
│ │ │ │
[32] [48] [72] [120]
│
[64]
│
...
every natural number appears exactly once. every path from root = unique factorization. the tree is complete.
the prime gap pattern:
| 0 1 2 3 4 5 | | | | | | | | | 0 1 | - | 0 1 | 0 1 1 | - | 0 1 | 0 1 1 | 0 1 1 1 | - | 0 1 | 0 1 1 | 0 1 1 1 | 0 1 1 1 1 | - | 0 1 | 0 1 1 | 0 1 1 1 | 0 1 1 1 1 | 0 1 1 1 1 1 |------------------------------------------------------|
the ones keep stacking. each prime opens a new column and the column fills downward as that prime composes with everything that came before. this is the fundamental theorem of arithmetic drawn as a growth pattern. every number gets built. no number gets skipped.
6. the same tree in cellular automata
the claim is that eca rule space has an analogous multiplicative structure. deep factorization = boundary behavior = computational irreducibility.
part three — the store
7. you own a store
you own a store. you sell phillips screws to everyone.
the screw is the atom of this economy. you can't break a phillips screw into sub-products and sell them separately. it's irreducible. it's primeirreducibleatomicfundamental.
8. competition as composition
a second store opens. now the market is composite. the period of the market is the product of two independent cycles. add a third store. a fourth. a tenth. a hundredth. the market's period gets deeper and deeper. at some point, the period is so deep that no practical analysis can predict the market's behavior. you have to run it and watch.
that's computational irreducibility. that's class iv. that's the boundary.
{---o------o---------o------------o---------------}->
0 1 storemonopolyprimesettled 2 storesduopolycompositeoscillating 10 storescompetitivedeepchaotic 100 storesirreducibleboundaryalive
9. the screw as phoneme
language works the same way. phonemesscrewsnucleotidesneighborhoods are the primes. sentencesshelvesproteinsautomata are the composites.
interior: "the cat sat on the mat" — grammatical, predictable, settled. exterior: "mat the on sat cat the" — ungrammatical, random, escaped. boundary: "the cat sat on the edge of meaning" — grammatical but alive.
the boundary is where language computes.
the writer folds. the reader unfolds. literacy is the ability to fold and unfold in a given system. mathematics is the ability to fold and unfold in all of them.
part four — primes as irreducible folds
10. what a prime is, structurally
a prime is an irreducible fold.
simple = few folds, all small → class i/ii random = one fold, very large (prime) → class iii complex = many folds, interacting → class iv
complexity is not about bigness. it's about depth of composition.
11. depth of composition
Ω(n), the number of prime factors of n counted with multiplicity.
depth 0: 1 (the unit)
|
depth 1: 2 3 5 7 11 13 ... (the primes)
| | | |
depth 2: 4 6 9 10 14 15 21 ... (products of 2 primes)
| | | |
depth 3: 8 12 18 20 27 30 ... (products of 3 primes)
| | |
depth 4: 16 24 36 ...
| |
depth 5: 32 48 ...
|
depth 6: 64 ...
|
...
every number lives at exactly one depth. deep folds = many interacting prime factors = complex behavior.
part five — bilateralism
12. the two faces of a number
every number has exactly two descriptions:
additive: 12 = 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1 + 1
twelve steps along the number line. where it IS.
multiplicative: 12 = 2² · 3
two folds of 2 composed with one fold of 3. what it IS.
the euler product formula is this bilateralism written in analysis:
ζ(s) = Σ 1/nˢ = Π 1/(1-p⁻ˢ)
n=1 p prime
left: sum over all integers. every position. additive face.
right: product over all primes. every fold. multiplicative face.
they're equal.
the riemann hypothesis says all non-trivial zeros lie on the line Re(s) = 1/2. the bilateral square from part zero has two traversals — perimeter returns 1, center through O returns 2. the midpoint is 1/2. Re(s) = 1/2 isn't a coincidence. it's the fold threshold.
13–15. bilateralism in cellular automata, the store, and language
every system has two faces. additive (where it IS) and multiplicative (what it IS). the interesting stuff happens where these two faces interact — with a delay. the delay is where predictions break down. the delay is the boundary.
part six — the boundary
16. where computation lives
the boundary of the mandelbrot set has hausdorff dimension 2 (shishikura, 1998). the boundary is as big as the plane it lives in.
interior = shallow folds = simple = boringsettledfrozendead exterior = single fold = random = noiseescapeddissolvedchaos boundary = deep folds = complex = computationlifealiveeverything
17. the boundary in physics
┌─────────────────────────────────────────────────────────┐ │ MACRO (> 10⁻³ m) │ │ deterministic. newton. interior behavior. │ ├─────────────────────────────────────────────────────────┤ │ ← MESO BOUNDARY │ ├─────────────────────────────────────────────────────────┤ │ MICRO (10⁻⁷ m to 10⁻¹⁰ m) │ │ statistical. boltzmann. boundary behavior. │ ├─────────────────────────────────────────────────────────┤ │ ← QUANTUM BOUNDARY │ ├─────────────────────────────────────────────────────────┤ │ QUANTUM (< 10⁻¹⁰ m) │ │ probabilistic. born. deep boundary. │ └─────────────────────────────────────────────────────────┘
lepton masses follow: m(2,q) ∝ q² · (ln q)^γ where q = 3, 5, 7. every factor derived from structure. zero free parameters.
18. the boundary in quantum mechanics
the born rule — probability = |amplitude|² — is the fold operation. z → z².
\/_
| /
|--< --* |ψ|² → probability
| \_
/\
P(+) = cos²(θ/2). derived from the geometry. not postulated.
part seven — the zeta function of rule space
19–21. ζ_eca
ζ_eca(s, w) = Σ 1/P(r,w)ˢ
r=0..255
a concrete, computable object. if the period structure of eca rules is genuinely multiplicative, then ζ_eca should admit a partial euler product decomposition. the quality of the approximation tells us how multiplicative the structure is.
part eight — the fold in everything
22. spacetime
mass folds spacetime. the singularity is where the fold is total. the event horizon is the boundary. the information paradox is a question about codec completeness.
the arrow of time is topological. π₁(S²) = 0. π₁(T²) = ℤ × ℤ. the fold is irreversible.
23. biology
dna → rna → protein is a fold chain. the genetic code is a compression system. 4 nucleotides → 64 codons → 20 amino acids.
24. evolution
evolution is iteration: state → fold(state) + address. speciation happens at the boundary.
part nine — the scaffold
25. the observer's dimension
every observation requires a scaffold. the dimension you can't access is the one you're standing on. you can't see it because you're seeing WITH it.
26. consciousness as scaffold
you cannot find consciousness in the brain for the same reason you cannot find the photographer in the photograph. the photographer isn't IN the photograph. consciousness is the projected-out dimension.
27. the mandelbrot set's scaffold
the mandelbrot set is a 2d projection of an infinite-dimensional object. the reader unfolds. the writer folds. this paper is a fold. you are unfolding it now.
part ten — the experiments
28. everything is testable
(* experiment 1: does Ω(period) correlate with wolfram class? *)
(* perez 2026, gueie.com *)
ECAPeriod[rule_, w_, maxIter_:10000] := Module[
{current, seen, t},
current = RandomInteger[1, w];
seen = <||>;
Do[
Module[{key = Hash[current]},
If[KeyExistsQ[seen, key], Return[t - seen[key]]];
seen[key] = t;
current = CellularAutomaton[rule, current];
],
{t, 0, maxIter}];
$Failed
]
classI = {0, 8, 32, 40, 128, 136, 160, 168};
classII = {1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 13, 14, 15};
classIII = {18, 22, 30, 45, 60, 73, 90, 105, 122, 126, 146, 150};
classIV = {54, 106, 110, 137};
(* experiment 2: euler product approximation of ζ_eca *)
ZetaECA[s_, w_] := Module[{periods},
periods = Table[ECAPeriod[rule, w], {rule, 0, 255}];
periods = Select[periods, IntegerQ[#] && # > 0 &];
Total[1.0 / periods^s]
]
(* experiment 4: boundary proximity score *)
BoundaryProximity[rule_, w_] := Module[{p, omega, lpf},
p = ECAPeriod[rule, w];
If[!IntegerQ[p] || p <= 0, Return[1.0]];
omega = PrimeOmega[p];
lpf = FactorInteger[p][[-1, 1]];
N[Log[p] * omega / Log[Max[lpf, 2]]]
]
part eleven — the complete picture
30. one fold, one rule, one structure
\/_
| /
|--< --*
| \_
/\
z → z² + c
five characters. this is the fold operation applied to the complex plane.
domain axis irreducibles fold event ───────────────────────────────────────────────────────────────────── physics prime number field primes phase transition proteins N→C backbone amino acids (20) oligomerization language form ↔ meaning semantic primitives discourse comprehension chemistry nuclear charge atomic nuclei molecular bonding music tonic ↔ dominant fundamental freq. modulation formal logic true ↔ false atomic propositions quantification
33. completeness
is the codec complete? can every address be reached? in all cases, the answer is the same: the codec is complete because the composition of irreducibles can reach every address. no fold sequence is orphaned. no address is unreachable.
the framework can be killed. these are the knives:
- a fourth lepton generation with koide Q ≠ 3/4 - a quantum measurement violating P(+) = cos²(θ/2) - lepton masses that can't be expressed via m(2,q) ∝ q²(ln q)^γ - koide derived from a framework with no fold thresholds
a framework that can't be killed isn't science. this one hands you the weapon.
riemann hypothesis: Re(s) = 1/2 is the bilateral midpoint of D₁. P ≠ NP: verification traverses D₁. search traverses D₃. yang-mills mass gap: Δ > 0 is the energy between O and the first stable knot. navier-stokes: smooth solutions exist for all time.
these are conjectures. not proofs. each is precise enough to be wrong.
34. a note on this document
nothing here is proved in the mathematical sense. everything here is either testable or derivable from the connection between two known facts.
run it. break it. extend it.
\/_
| /
|--< --*
| \_
/\
andre perez · san francisco, ca · april 2026
gueie.com