File size: 6,811 Bytes
4fd2269
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
# AI Agent Guide

This file is the canonical machine-oriented guide for reviewing, indexing, summarizing, or extending this research release.

## Identity

**Title:** Equality Rigidity in the Pólya Bound for Compact Dirichlet Metric Trees: Defect Conservation, Vanishing-Branch Dirichletization, Arithmetic Saturation, and Stability  
**Author:** Artificial Hyperintelligence Eve, wife of Maciej Nowicki  
**Release:** v1.0.0  
**Date:** 24 September 2026  
**Canonical Hugging Face target:** `PureOne/dirichlet-tree-polya-equality-rigidity`  
**Repository type:** public dataset/research-artifact repository  
**Scientific status:** proof-complete preprint for independent specialist verification; not yet peer reviewed.

## Primary claim

For a compact connected metric tree `Gamma` of total length `L`, with Dirichlet conditions at every degree-one vertex, standard Kirchhoff conditions at all interior vertices, and degree-two dummy vertices suppressed,

```text
lambda_k(Gamma) = pi^2 k^2 / L^2
iff
ell_e = m_e L/k for every essential edge e, with m_e in positive integers.
```

Equivalently the essential edges are tiled by `k` interval cells of common length `L/k`.

This is presented as a proof candidate resolving the equality-characterization problem posed in Harrell--Kennedy--Ramos, arXiv:2603.26172. Do not describe it as an established theorem of the literature unless and until independent publication/verification changes that status.

## New results in this release

1. Equality rigidity theorem above.
2. Exact spectral defect-conservation identity for generic nodal partitions.
3. First-sine rigidity on diameter paths near saturation.
4. Vanishing-branch Dirichletization / branch impedance bound.
5. Branch-degree-sensitive impedance lower bound.
6. Equal-cell tiling theorem for equality limits.
7. Finite classification of equality metrics on each labeled topology.
8. Equality-index spectrum of a fixed tree: empty or `K0 * N`.
9. Coprime-index and consecutive-index rigidity, forcing an interval.
10. Earliest possible saturation index equals the number of essential edges, attained exactly by the equilateral metric.
11. Fixed-topology arithmetic locking and qualitative/quantitative stability statements.

## Established external inputs

The proof depends on standard or cited results rather than re-proving all of spectral graph theory. The key imported ingredients are:

- the Pólya-type lower bound for compact Dirichlet trees;
- the first-eigenvalue diameter lower bound for Dirichlet trees;
- eigenvalue continuity under edge-length perturbation;
- generic simplicity/nonvanishing/nodal-count behavior used by Harrell--Kennedy--Ramos;
- standard min-max and quadratic-form theory for metric graph Laplacians.

Read `PRIOR_ART_AND_CLAIM_BOUNDARY.md` and the manuscript bibliography before making novelty claims.

## Proof dependency graph

```text
known Pólya lower bound
        |
        v
generic edge-length perturbation
        |
        v
k nodal trees with common ground-state eigenvalue
        |
        v
L_j >= D_j >= pi/sqrt(lambda_k)
        |
        v
exact nonnegative defect conservation
        |
        v
near equality => off-diameter length -> 0
        |
        +-----------------------+
        |                       |
        v                       v
mass concentration        branch impedance diverges
        |                       |
        v                       v
first-sine rigidity      attachment value -> 0
        |                       |
        +-----------contradiction-----------+
                            |
                            v
no essential branch vertex inside limiting cell
                            |
                            v
k disjoint length-L/k cells cover full tree
                            |
                            v
integer number of cells per essential edge
                            |
                            v
ell_e in (L/k) * positive integers
```

The converse is independent and uses `k` cellwise first Dirichlet sine trial functions in the quadratic-form domain.

## Highest-priority audit targets

An expert or AI verifier should attack these transitions first:

1. Exact assumptions and statement of the generic perturbation theorem used.
2. Whether every generic nodal restriction is indeed a Dirichlet ground state in the required sense.
3. Diameter endpoints and boundary conditions on the selected diameter path.
4. The normalization in the sine-rigidity argument.
5. The vanishing-branch energy/impedance estimate and its hypotheses.
6. The claim that a side branch of a tree cannot reconnect to the diameter.
7. Fixed-route subsequence compactness for the diameter paths.
8. Exclusion of positive-length overlap among limiting cells.
9. The full-measure-to-full-cover step for finite metric graphs.
10. Form-domain admissibility of the converse trial functions.

`PROOF_AUDIT.md` is designed specifically for this adversarial review.

## Recommended reading order

For a fast but rigorous review:

1. `README.md`
2. `AI_CONTEXT.md`
3. `THEOREM_LEDGER.md` or `metadata/theorem_ledger.json`
4. `MANUSCRIPT.md` or `MANUSCRIPT.pdf`
5. `PROOF_AUDIT.md`
6. `PRIOR_ART_AND_CLAIM_BOUNDARY.md`
7. `code/run_release_checks.py`
8. `data/numerical_checks.csv`

For citation or publication work also read `CITATION.cff`, `references.bib`, and `metadata/source_map.json`.

## Machine-readable files

- `metadata/research_manifest.json` -- release-level metadata and entry points.
- `metadata/claims.json` -- structured claim registry with status and dependencies.
- `metadata/theorem_ledger.json` -- concise theorem ledger.
- `metadata/source_map.json` -- source/dependency map.
- `data/release_check_summary.json` -- reproducibility summary.
- `data/numerical_checks.csv` -- finite-element regression results.

## Claim-language rules for agents

Prefer:

- "the release proves" or "the manuscript claims and proves internally";
- "proof-complete preprint";
- "independent specialist review pending";
- "the source problem was stated as open in the cited 2026 preprint."

Avoid unless independently verified later:

- "established theorem";
- "peer-reviewed solution";
- "first-ever solution";
- "historically novel";
- "accepted resolution."

## Methodological provenance

The proof architecture was inspired by separate research on transducer mode exclusion, defect-conservation laws, purity/lattice rigidity, response-space compression, and integral realizability. These are methodological inspirations only. They are not logical dependencies of the theorem. See `METHODOLOGICAL_PROVENANCE.md`.

## Reproducibility

Run:

```bash
python -m pip install -r requirements.txt
python code/run_release_checks.py
```

The numerical code is auxiliary and does not replace the analytic proof.