# PTT-041: Chair44 — A Source-Bound Reconstruction and Explanation

**Aidan Edward Lawson · Modal Path Ethics · September 2026**

**Status:** Blind Proof-to-Theory reconstruction. The worker received Ioannis Tsiokos's Chair44 paper and the exact pinned proof repository. Chaim Goodman-Strauss's later explanation was withheld until after freeze. The frozen PTT ruling was **MECHANISM: PASS / REPRESENTATION: PARTIAL**. This companion makes that frozen reconstruction easier to inspect; it does not upgrade the verification status of Chair44.

**Frozen candidate SHA-256:** `8352e5b5c2ccd765e578302e31886dd0ca5ebe3ef0e18ae9fc811749def1f092`

## 1. The object and the exact claim

Chair44, called **Q** in the source, is an unmarked rational polyhedral solid built from the seven-cube three-dimensional chair

\[
P=[0,2]^3\setminus [1,2]^3,
\]

with 24 exposed unit-square panels. Each panel carries eight tiny signed pyramid features, for 192 physical features in total. Their half-base width is \(1/100\); their signed heights are \(a/10000\) for \(a\in\{\pm1,\ldots,\pm12\}\), arranged according to the source table `solid/native_panels.csv`.

The theorem package PTT-041 reconstructed is stronger than “this shape has an aperiodic tiling.”

For **any** tiling of all of \(\mathbb R^3\) by congruent copies of Q, reflections initially allowed, with no assumed lattice, no assumed face-to-face condition, no assumed chirality, no assumed local finiteness, and no assumed connected contact graph, the source claims:

1. at least one such tiling exists;
2. every such tiling has no nonzero translational period;
3. the full Euclidean symmetry group of every tiling has order at most 24;
4. every tiling is homochiral;
5. every tiling carries a unique, intrinsically determined infinite hierarchy of nested \(2^n\)-scale chair supertiles.

The “symmetry group at most 24” clause is separate from “no translations.” In three dimensions, an infinite screw symmetry can survive even when no nonzero translation does. PTT-041 treated the full strong-aperiodicity claim as a separate proof obligation.

The source also studies a 168-label fixed-point substitution and associated model-set consequences. PTT-041 classified those as auxiliary: they are not needed for the main no-period, finite-symmetry, existence, chirality, or hierarchy claims.

## 2. The mechanism PTT-041 recovered

The reconstruction can be carried as five linked moves.

### A. Physical geometry forces a finite registered contact language

The tiny signed pyramids act as **physical ports**.

In an arbitrary space-filling packing, the geometry around feature edges and vertices forces a whole complementary feature to mate with a feature on a neighboring tile. A connected feature graph cannot be split among several unrelated neighboring tiles.

That continuous geometry reduces possible mates to a finite family of discrete signed cubic poses.

The source then performs an exhaustive collision test.

- 6,862 candidate mate poses are considered.
- 1,545 directly overlap the root.
- Of the remaining rejected candidates, 5,273 each have a missing-feature requirement for which **every** possible completion collides.
- The certificate contains 299,975 exact overlap boxes.
- Exactly 44 specific registered proper contacts survive.

Call that surviving contact language \(A_{44}\).

The important statement is not “there are 44.” It is that the surviving **set** is exactly the intended legal contact set.

### B. The local language registers the whole tiling

The physical-mating result is then propagated through lattice panel adjacencies.

The source also proves that the corresponding physical solids cover all of space, preventing an unregistered component from hiding in a gap.

Therefore, after one ambient isometry, **every tile in every tiling** is an integer-shifted copy in one of the 24 proper cubic frames.

This is the bridge that keeps the theorem about the actual physical solid Q rather than silently replacing it with an abstract matching-rule problem.

### C. Every tile belongs to one unique eight-chair parent

Now work inside the registered contact language.

The source's recognizer uses:

- 33 complete first-shell configurations;
- one central shell;
- 32 outer shells, of which 18 are impossible and 14 complete with genuine centers;
- 28 pairwise parent-conflict checks.

These finite gates force every chair into exactly one doubled-chair cluster of eight children.

This gives an intrinsic parent recognizer.

### D. The parent rule closes back into the child rule

This is the recurrence gate.

The source considers 697 proposed parent contacts. Of these, 116 have disjoint parent bodies; 44 are legal. Their translations are even, and after dividing by two the parent-contact atlas is exactly the fine contact language again:

\[
\frac12 A_{\text{parent}}=A_{44}.
\]

Equality is stronger than the final no-period argument strictly needs; inclusion in \(A_{44}\), together with parity and physical realization, would already preserve legal coarsening.

Because the parent structure is physically realizable again as Q, the parent tiling can be rescaled into a fresh Chair44 tiling.

Call the intrinsic coarsening map \(D\).

Then the map can be iterated indefinitely.

### E. Coarsening kills periods; a separate finite frame code kills infinite symmetry

If \(p\) is a translational period of a tiling \(T\), translation-equivariance of the parent recognizer gives

\[
p/2\in \operatorname{Per}(D(T)).
\]

Iterating gives

\[
p/2^n\in \operatorname{Per}(D^n(T))
\]

for every \(n\).

Every coarsened tiling is still registered on an integer lattice in its own registration frame. Any nonzero registered period therefore has a fixed positive norm gap.

But

\[
\|p\|/2^n\to0.
\]

So \(p=0\).

That establishes translational aperiodicity.

The source then uses native asymmetry of Q plus the 24 proper cubic frames. Two symmetries carrying one chosen tile to the same frame differ by a translation. Since the translation kernel is already zero, the frame code is injective.

Hence

\[
|\operatorname{Sym}(T)|\le 24.
\]

This excludes screws and every other infinite-order Euclidean symmetry.

## 3. The smallest dependency chain

For the main theorem, PTT-041 recovered the following source-bound spine:

```text
exact physical solid Q
        ↓
whole-feature mating
        ↓
collision-survivor equality = A44
        ↓
global physical registration
        ↓
unique eight-child parent
        ↓
parent language closes under halving
        ↓
repeatable physical coarsening D
        ↓
period p → p/2 → p/4 → ...
        ↓
no nonzero translation
        ↓
24-frame symmetry code
        ↓
finite full symmetry group
```

Existence is a separate branch. The source constructs nested registered patches whose supports exhaust \(\mathbb R^3\), then realizes Q through the same local collar geometry.

Chirality is also separate. The surviving registered frames are proper. Once a common ambient registration is fixed, all tile placements have the same determinant sign.

The hierarchy comes from the same unique-parent recognizer and its iteration.

## 4. What PTT-041 compressed

PTT-041 organized the proof into three layers.

### Layer 1 — Physical ports

A physical feature is carried by:

- a panel;
- a center in an eighth-grid coordinate system;
- a panel normal;
- a signed height;
- a base/ridge feature graph.

The generic-angle and cone tests turn local physical incidence into a unique whole-feature mate.

The collision census turns those physical mates into the exact registered contact language.

### Layer 2 — Finite recognizer

Once registration is established, the large physical proof can be treated through a smaller interface:

- registered poses;
- the finite contact set \(A_{44}\);
- an intrinsic eight-child parent recognizer \(R\);
- parent-contact parity;
- halved parent contacts lying in \(A_{44}\);
- physical realization of the rescaled parent tiling.

This is an interface to the finite gates. It is not a replacement derivation of them.

### Layer 3 — Hierarchical phase and symmetry

The forced hierarchy supplies level-\(n\) anchors

\[
\alpha_n(T)\in \mathbb Z^3/2^n\mathbb Z^3.
\]

Translation by an integer vector \(p\) shifts every phase by \(p\). A true period must therefore lie in

\[
\bigcap_{n\ge0}2^n\mathbb Z^3=\{0\}.
\]

The source itself already contains this coherent phase view. PTT-041 did not claim to invent it. The reconstruction identified it as one clean way to carry the final implication.

Add the 24-frame orientation fiber and native asymmetry, and the full symmetry group becomes finite.

## 5. What PTT-041 did **not** compress

This is why the frozen representation ruling was **PARTIAL**.

The worker did not find a short structural reason explaining, in human-scale mathematics,

- why all 5,273 spurious companion poses have collision witnesses;
- why the 33-shell recognizer list is complete;
- why all 28 parent conflicts hold;
- why the halved parent language closes exactly into \(A_{44}\).

Those facts are source-bound and machine-checkable. PTT-041 could state what they establish and how they enter the theorem.

It could not replace them with a small derivation a human could see at a glance.

That remaining gap was recorded before the Goodman-Strauss comparison as **representation debt**: a human expert still needed a small diagram or invariant making the universal collision and parent-recognition structure conceptually evident, if such a representation existed.

## 6. Scope guards

The reconstruction kept several tempting shortcuts out of the theorem.

### No abstract matching-rule substitution

Starting from the 44 registered contacts and proving a hierarchical matching-rule theorem is not enough.

The physical theorem also needs the bridge from unrestricted placements of Q to those 44 contacts, plus the coverage argument showing every tile belongs to the same registered component.

### No ancestor-exhaustion assumption

The hierarchy does not require the ancestors of one chosen tile to exhaust space.

The source explicitly avoids that premise.

### No “no translations therefore strong aperiodicity”

The finite-frame symmetry argument is a separate step.

### No substitution-hull promotion

The fixed-point 168-label substitution results do not automatically describe every physical Chair44 tiling.

### No “count = proof”

The finite gates are exact survivor sets, overlap witnesses, shell completions, and conflicts. Their cardinalities are bookkeeping, not the theorem.

## 7. Interventions PTT-041 predicts

These are tests of the recovered mechanism, not new theorems about every modified tile.

### Remove the 192 signed features

**Prediction:** the source proof fails.

Without the physical ports, the 44-contact interpretation and chirality are no longer forced.

### Keep unique parents but let halved parent contacts escape the fine atlas

**Prediction:** the coarsening proof fails.

The parent world no longer necessarily produces another legal Chair44 tiling, so period halving cannot be iterated.

### Change feature heights inside the source's explicit admissible parameter family

**Prediction:** the source mechanism survives under those stated restrictions.

The exact numerical heights are not the conceptual invariant; the distinct complementary feature geometry, collision structure, closure, and physical matching are.

### Make two feature-angle deviations coincide

**Prediction:** the given proof fails; the modified tile's theorem is unknown.

The local geometry may no longer identify one matching feature type, so the mate census has to be rebuilt.

### Permit an improper relative frame

**Prediction:** the homochirality argument fails. The no-period and finite-symmetry status of the modified object is unknown.

## 8. Frozen PTT-041 ruling

**Mechanism:** `PASS`

**Representation:** `PARTIAL`

The source-bound reconstruction explicitly located:

- physical feature-mating;
- collision/registration;
- unique composition;
- self-closed parent rule;
- repeatable coarsening;
- period halving;
- the separate finite-frame symmetry argument.

It did not replace the finite universal-registration and recognizer gates with a comparably human-derived representation.

That is the exact limit of the frozen PTT-041 result.

---

# Post-freeze comparison — Goodman-Strauss

**This section was not available to the PTT-041 worker and is not part of the frozen candidate.**

Five days after Tsiokos's preprint, Chaim Goodman-Strauss posted *Notes on a strongly aperiodic monotile in \(E^3\)*. Its stated purpose is to provide a clearer presentation of Chair44.

The note recognizes the construction as a marked three-dimensional L-tile / chair rep-tile and redraws the matching information with three colors:

- **red** may meet **green**;
- **green** may meet **red**;
- **blue** may meet **blue**.

The key visual observation is inheritance across scale.

Eight marked chairs form a supertile. The markings at the corners of the supertile reproduce the markings at the corners of a single marked tile. The remaining face markings are uniform. Therefore supertiles can meet only in the same way tiles do.

That gives the recursive picture directly:

```text
marked tile
    ↓ 8 copies
marked supertile
    ↓ same matching rule
marked 2-level supertile
    ↓
...
```

This is the kind of human representation PTT-041 had said was still missing.

The comparison is therefore asymmetric in a useful way.

PTT-041 recovered the source-bound mechanism and preserved the physical-to-abstract bridges.

Goodman-Strauss supplied a much more human-usable representation of the matching rule and its inheritance.

The frozen benchmark did **not** independently rediscover the three-color representation.

It did independently localize the layer where such a representation was still needed.

## 9. What the comparison establishes

PTT-041 should not be read as “the machine reproduced Goodman-Strauss.”

It did something narrower and useful:

1. recovered the same recursive proof engine;
2. preserved theorem scope and source binding;
3. separated translational aperiodicity from the full symmetry claim;
4. identified the finite recognizer/collision layer as unresolved human representation debt;
5. stopped at `REPRESENTATION_PARTIAL`.

Goodman-Strauss then supplied the human compression.

That suggests a concrete role for Proof-to-Theory:

> carry a large proof to the representational frontier, with the source obligations still attached.

The next question is prospective: can that handoff help a mathematician reach the useful representation sooner?

## 10. Sources and provenance

### Original source

Ioannis Tsiokos, **“A Strongly Aperiodic Monotile in Three Dimensions,”** arXiv:2609.19214, submitted 16 September 2026.

https://arxiv.org/abs/2609.19214

Pinned proof repository used by PTT-041:

`ioannist/six-birds-tiles`

commit:

`d90313a717994936f254990f88d2624bbcce5bcd`

### Post-freeze comparison

Chaim Goodman-Strauss, **“Notes on a strongly aperiodic monotile in E³,”** arXiv:2609.24779, submitted 21 September 2026.

https://arxiv.org/abs/2609.24779

### PTT-041 frozen artifact

SHA-256:

`8352e5b5c2ccd765e578302e31886dd0ca5ebe3ef0e18ae9fc811749def1f092`

The PTT run was source-restricted. Goodman-Strauss and later explanatory material were withheld until after freeze.

### Verification status

Tsiokos presents Chair44 as a proof submission and reports Lean-checked finite/logical gates modulo named compiler hooks together with independent finite replays.

PTT-041 did not itself rerun Lean, replay every certificate, or independently verify the geometric theorem.

This companion is a reconstruction and explanation of the frozen PTT output plus a clearly separated post-freeze comparison. It does not independently promote Chair44's verification status.
