Observable reality is binary.
Nature is not.
Every measurement returns State A or State B. That's the architecture of all modern science, all AI, all diagnostic medicine. It's extraordinarily useful — and it systematically discards a third term.
We call it S₃. It's the conjugate-domain structure that binary measurement trades away. In quantum mechanics it's the uncertainty principle. In galactic rotation it's the residual the NFW profile can't fit. In oncology it's the pre-transition cellular state that fires before malignancy is classifiable.
The gap isn't noise. It's information. MCORE is a measurement framework built to instrument it.
What We See
Acoustic Signal Structure
You cannot simultaneously know when a sound occurs and what frequency it contains. That traded-away information is S₃. The Gabor limit σ_t × σ_f ≥ 1/(4π) is the uncertainty principle made audible — and it governs everything.
Epistemic Propagation
How ideas spread, mutate, and destabilize systems. 47M Reddit comments. SIR epidemiological modeling. Granger causality at p < 10⁻⁶. The S₃ of social signal — what narrative dynamics discard.
Cosmological Structure
143 SPARC galaxies. Binary NFW vs. ternary NFW + S₃ Gabor envelope. BIC model selection. 21 galaxies show massive ternary wins (ΔBIC as low as −667). The residual the standard model cannot fit.
AI Interpretability
What does a model actually know — internally, structurally? MCORE instruments the S₃ term in weight geometry. The gap between model representation and ground truth is not noise. It's measurable.
Current Research
Sound and Uncertainty: Why Perfect Knowledge of Audio Is Physically Impossible
The clearest proof that S₃ exists. σ_t × σ_f ≥ 1/(4π) — you cannot know both when and what simultaneously. The cochlea, the bat, the MP3 codec, the quantum phonon: all governed by the same traded-away term.
Memetic Volatility Propagation: Cross-Subreddit Sentiment Contagion as a Leading Indicator of Market Volatility
SIR epidemiological modeling applied to 47M Reddit comments across 54 finance subreddits. 25/1,144 Granger causality tests survive Benjamini-Hochberg FDR correction at p < 10⁻⁶.
MCORE-1: A Ternary Weight Algebra for Narrative Structure in Language Model Representations
Formal algebraic engine with ternary state space T = {−1, 0, 1}, TME-6 encoding, and provably convergent attractor. 63/63 unit tests passing. GF(3) arithmetic backbone.
Work with us.
Or just talk.
Symonic is an independent quantitative research platform pursuing SBIR and DARPA-track funding. We're interested in collaborators, researchers, and institutions working at the edge of AI interpretability, epistemic systems, and formal signal theory.
jacob@symonic.com