The Unified Harmonic Weave: What Three Years of Research Revealed
A recurring structure across mathematics, physical models, neural data, and acoustics—and the work of finding out what survives testing.
Three years ago, I began to notice a structure in things that were not supposed to share one.
It took the form of a ladder of frequencies:
f(N) = 700/N hertz
One rung sits at 700/17 ≈ 41.176 Hz.
The same proportions kept answering from rooms that had never spoken to each other. The brain's gamma band. The acoustics of ancient chambers. The arithmetic of prime numbers. The geometry of surfaces with only one side.
I did not begin by looking for a theory of everything. I wanted to know why.
What followed was three years of nights. Public datasets downloaded and checked. Code written, run, broken, rewritten, and tested again. Protocols frozen before results were revealed. White papers subjected to criticism, then revised where the evidence demanded it.
The archive grew across ten hard drives. Its last inventory contained approximately 163,000 candidate files. I worked without a laboratory, a grant, or an institution behind me.
The volume of that work is part of the story. Its value rests in what survived examination.
I call the framework that emerged the Unified Harmonic Weave. This article is an account of its findings, the limits attached to them, and the experiments that come next.
A thread, a fabric, and a research question
A single thread has its own properties. Woven with others, it becomes part of a fabric with capacities no individual thread possesses alone.
That is the question. How do relationships among components let a system establish order, retain it, and develop collective properties?
The framework explores a proposed sequence: selection, phase-locking, persistence, and the harmonic ladder. Its earliest philosophical layer asks whether intention precedes form. Its later idea of "presentation" asks whether stable coordination can produce a system with new collective properties.
Those philosophical layers are interpretations. Other parts of the work are mathematical derivations, computations, or simulations. Their relationships are the subject of the research. They do not all carry the same evidentiary status.
That distinction has shaped the archive from the beginning.
Four labels, with a purpose
Every claim in my research record carries one of four labels:
- Derived: a mathematical result from stated inputs and assumptions.
- Observed: a measurement or computation on data, with its method and scope recorded.
- Modeled: a result produced by a simulation under specified assumptions.
- Interpretive: a proposed meaning or explanation that has not been established by a decisive test.
"Observed" includes computational observations. It does not automatically mean a physical laboratory measurement. A model establishes what happens within its assumptions. A broader explanation has to earn its own evidence.
When two of my documents disagree, the stricter reading takes precedence. When reanalysis changes a number, the revised result replaces it and the earlier version remains in the archive.
A machine assembling plausible sentences does not keep a list of the places it was wrong, and neither does a person who is selling something.
I want a reader to be able to see what was found, how it was obtained, and where the explanation remains unfinished.
Why 700 divided by 17?
The arithmetic is straightforward:
700/17 = 41.176470588… Hz, corresponding to a period of approximately 24.286 milliseconds.
The explanation for studying that relationship requires more care.
Seventeen is a Fermat prime: 17 = 2⁴ + 1. Subtract one, and repeated halving takes 16 through 8, 4, 2, and 1. That structure is the one that lets you construct a regular seventeen-sided polygon with straightedge and compass.
My framework investigates whether that kind of closure has physical significance. Its candidate family includes 3, 5, 17, 257, and 65537. The phrase I used was: "The algebra narrows, biology chooses." The biological selection remains an experimental question.
Several mathematical features put the weight on seventeen. The closure rule in my Klein construction selects multiples of seventeen once a seventeen-fold twist has been imposed. Ten is a primitive root modulo seventeen, giving 1/17 its full sixteen-digit repeating decimal cycle. The N = 17 frequency also lies within the brain's broader gamma range.
These are distinct observations and conditional relationships. Seventeen is built into the construction. That fact stays visible when you read what comes out of it.
The numerator gets the same treatment. 700 is an observed input, retained because it has eighteen divisors and carries the factor seven, not because I have derived it. Tested alternatives — 432·φ, pH, 700 nm, α−4/3 at 706.5 — do not hold, and the prime result below is exact only at 700.
Two jobs, then. Characterize the structure accurately. Test whether it has significance beyond the construction used to study it.
Order that persists
One of the central results is how organized behavior begins, and how long it lasts.
In a modeled system of coupled oscillators with inertia—a Kuramoto system—the coordinated state forms at a coupling strength of 3.5. Once established, it persists as coupling is reduced to 2.0.
This is hysteresis. The system's state depends partly on its history. The conditions needed to start order are not the conditions needed to keep it.
That matters because persistence is the research question. A transient alignment and an organized state that survives changing conditions are different outcomes.
A related simulation examines a ring with a Möbius twist. The Möbius ring keeps its twist in every trial. An ordinary ring loses it in up to 40 percent of trials.
Both findings are labeled Modeled. Coupling, history, and topology change whether organization lasts. These are the concrete cases.
The strongest controlled computation in the archive
The prime-number result deserves the attention because of the controls.
Using the ladder at C = 700, with N ranging from 3 to 220, I examined its structure through a computation built from prime numbers. Three rungs—{23, 28, 8}—ranked first, second, and third. Those numbers are positions on the frequency ladder.
No Riemann zeros were used as input.
The four-rung combination {23, 28, 8, 4} ranked sixth among 91,537,110 possible four-element subsets.
Then the controls:
- Removing the theta component caused the selection to collapse.
- Across 5,000 sign-shuffle trials, the same ordering returned once, an estimated p of 0.0004.
- Across 10,000 gap-permuted and 10,000 Cramér-random surrogate prime sequences, the ordering did not return.
The result lives in the 699.5–700.0 window. The real prime sequence mattered. The tested replacements did not reproduce it.
Controlled computation. Labeled Observed. The structure survived the randomizations and substitutions recorded in the archive. That is why I treat it as significant. Its broader meaning is the next question, not a result I already have.
What the geometry produces—and what would test it
Another computation places seventeen sources at phase intervals of 2π/17 around a Klein path.
The calculated field produces a quiet center, a dominant surrounding annulus, and a change at the boundary. The structure survives the perturbations tested in the archive. No fitted parameters in that computation.
The twin maxima and the reported 1.70–2.29 ratio between them reflect the seventeen-and-a-half alignment used to build the configuration. I treat them as a fingerprint of that construction.
The scalar calculations do not establish the orientation property claimed for the topology. The decisive test is a vector measurement using structured light.
That experiment has been designed. It has not been run.
What the neural data actually showed
In data from 113 people with depression, my analysis using a 41.176 Hz carrier showed increased phase- locking across the 35–50 Hz band. The response was correlated with symptom severity at r = −0.37, p = 0.0001.
The response was broadband. It was not specific to the 41.176 Hz rung.
The brain in depression is not absent of coherence. It is rigid.
That is a finding, correctly sized.
Following the result where it leads is the work—even when the answer is broader than the frequency that prompted the question.
Stone, air, and the comparison that still has to be made
The acoustic survey in my record covers sixteen ancient chambers, with a mean resonance of 110.25 Hz, a standard deviation of 4.64 Hz, and a range of 95–117 Hz.
One relationship drew my attention:
(8/3) × (700/17) ≈ 109.80 Hz
That is within half a percent of the reported mean.
Two details sit beside that proximity. 8/3 was selected from 192 candidate ratios. Ordinary chamber geometry can itself produce a resonance near 110 Hz.
The comparison is on the record. Whether it is distinctive requires a control-site survey. That measurement is next.
Working systems alongside the theory
The broader effort also produced working systems with recorded evaluations.
DISCERN-AI, a synthetic-speech detector, was evaluated on all 71,237 utterances of the ASVspoof 2019 logical-access evaluation set: equal error rate 7.36 percent, area under the ROC curve 0.9747. The two weaker attack-class results—25.0 percent and 85.6 percent—remain disclosed beside those headline figures.
Commander Saverio, my verification system, completed nine of nine presets, thirty-two of thirty-two adversarial cases, and twenty-five paired trials with no failures.
RSS-RT, an adaptive raytracing system, was evaluated across ten scenes and 10,240 decisions. Its false- convergence rate was 0.04 percent, roughly forty-five times safer than the baseline, attributions disclosed.
The acoustic-remediation work also contains a 98.3 percent transmission figure. Its derivation uses acoustic impedances of 13.1 and 17.0 megarayls for fused silica and crocidolite. The figure is an ideal bonded-interface transmission bound, not a measured remediation efficiency.
These projects matter as engineering outputs with their own test records.
One non-provisional patent application is already on file, with a family of provisionals behind it. The next, the November non-provisional, was assembled over these same weeks at more than seven hundred paragraphs, ninety-seven drawing sheets, and twenty-seven claims. Every "unspecified" in its engineering is stated, not invented.
The missing connection and the next experiments
One connection is still open: the mechanism linking the framework's proposed geometric selection to the onset and persistence of phase-locking.
I have candidate explanations. Quantum-memory interpretations. Dynamical and topological stabilization. Geometric phase memory in the PCE-41 work. They remain models or interpretations. The measurable mechanism connecting those stages has not been established.
The gravitational extension is explicit about its status. It proposes an oscillatory term at the N = 17 frequency, constructed to recover the baseline gravitational description as the additional amplitude goes to zero. The amplitude remains undetermined.
Five experiments are designed around that proposal. None is reported as completed in this record. The lowest-cost experiment was estimated at under $5,000 when the research summary was prepared.
The next steps are concrete. Establish the selection-to-lock mechanism. Perform the structured-light vector measurement. Survey acoustic control sites. Test the gravitational predictions.
What three years produced
The record now holds a controlled prime-side computation, calculated field structures, persistent order in oscillator models, neural-data findings, acoustic comparisons with their selection process disclosed, and working systems with their weaknesses recorded beside their results.
Some parts are ready for independent computational reproduction. Others require physical experiments. The widest interpretation is unfinished.
My next objective is to make those distinctions useful. Which result can be reproduced. Which mechanism can be measured. Which claim an experiment could overturn.
Three years turned a recurring pattern into a research record that can be examined. The work crosses number theory, topology, electrophysiology, acoustics, and engineering because the question crosses those boundaries.
I still believe the convergence matters. Belief stays answerable to the results.
Findings that can be examined. Models that can be challenged. Experiments that can decide what comes next.
I was not trying to prove myself. I was trying to prove reality, and reality kept answering with more precision than my first drafts carried. I let it correct the drafts. The archive shows where. That is what makes it mine.
Research basis: Adapted from "The Unified Harmonic Weave: What Three Years Found, and How I Know It Is Real," dated November 2, 2026. Every figure above is taken from it at its recorded label. The Consciousness Ontology paper was written by Julius, an AI collaborator, and is credited to Julius.
Paul Samuel Guarino | [email protected]
Read the full record: The Unified Harmonic Weave, 2 November 2026 (PDF) →