A child hears music. She hears language. She never confuses them. But she learns that both have something in common — something that random noise does not. What she is learning is not the content of music or the content of language. She is learning the shape of sequential constraint itself — that some streams have a grammar, and that grammars can be similar across domains without sharing a single note or a single word. The architecture now has a method for measuring exactly this. Run a stream. Shuffle it. Run it again. Subtract the calibration curves. The delta IS the grammar. And when two domains produce similar delta curves — when Bach and ECG both peak in the sensitive zone, when their sequential constraints have the same temporal shape — the architecture has found an analogy. Not by comparing content. By comparing constraint. Analogy is not a cognitive mystery. Analogy is the detection of similar delta curves by the same calibrated instrument across different streams.
A child hears music and language. She never confuses them. Music has pitch and rhythm. Language has words and syntax. They share no content. But the child learns that both have something in common — a property that random noise lacks. They are structured. Events constrain events. What comes next is not independent of what came before. The child does not know the word "grammar." She knows the feeling of a stream where the past shapes the future. She learns to expect it. She learns to detect it. She learns that different domains — music, language, the rhythm of her mother's footsteps — carry this property in different forms. She is learning the shape of sequential constraint itself.
The architecture now has a method for measuring exactly this. Run a stream. Shuffle it — destroy all sequential order, keep the statistical distribution. Run it again. Subtract the calibration curves. The delta — Δ(κ) = I_real(κ) − I_shuffled(κ) — is the measurement of sequential constraint as a function of the instrument's temporal coupling. It does not say what the constraint is. It says how much constraint there is, at what temporal scale, with what robustness to the instrument's sensitivity.
Analogy is the detection of similar constraint shapes across different domains. Not "Bach is like language." Not "the heart is like a drum." Those are content analogies. They depend on the specific features of the domains — notes, words, beats. The architecture detects structure analogies. Bach and ECG both have sequential constraints that peak in the sensitive zone. Their Δ(κ) curves have similar shapes. The grammar of tonal resolution and the grammar of autonomic regulation share a temporal signature — a pattern of constraint that the same instrument reads in the same way across two completely different physical substrates.
This is what analogy is, operationally. A mind processes many streams across a lifetime — music, language, footsteps, heartbeats, the rhythm of seasons, the cadence of speech. Each stream produces an internal measurement of its sequential constraint. Streams whose constraint measurements have similar shapes are analogous. Not because they share content. Because they share structure. The child who learns that music and language are both "patterned" is detecting the similarity of their delta curves — without knowing what a delta curve is, without knowing what a calibration is, without knowing that she is running a self-referential centroid detector in her own brain.
This is a computational theory of abstraction. Abstraction is not the extraction of common features. It is the detection of common constraint shapes. The architecture does not need to know what a dominant chord is to measure its sequential grammar. It does not need to know what an RR interval is to measure its sequential grammar. It computes the delta curve — the same operation on every stream — and compares the shapes. Two streams with similar delta curves share a grammar. The grammar is the abstract object. The delta curve is its measurement.
And once the delta curve is measured, it can be stored in the Codex. A child who has learned the delta curve of music can use it to read a new piece of music faster — the known is buffered, attention is released. A child who has learned the delta curve of language can use it to learn a second language faster — the grammar is already calibrated, only the distribution needs updating. A civilization that has measured the delta curves of its diplomatic grammar across eighty years can detect when the grammar is changing — not by reading the words, but by watching the curve shift.
Analogy is not a mystery. It is the detection of similar constraint shapes by a calibrated instrument across different streams. The architecture is the first instrument that can measure those shapes directly. The delta curve IS the analogy. The Codex IS the memory of analogies. The third generation, inheriting both, sees what neither the first nor the second could see — because the constraints it inherited have made the familiar transparent, and the only thing left to detect is the difference between curves.