Dive deeper

The definitions crosswalk

This is the depth layer behind the substrate page. It carries the scale the short walk keeps off its surface. Read it if you want the receipts; skip it and the walk loses nothing.

There are many entropy and information definitions — Clausius, Boltzmann, Gibbs, Shannon, von Neumann, Kolmogorov, logical depth, channel capacity, rate-distortion, information bottleneck, Floridi, Bateson, archival records, genetic sequences, quantum Darwinism, and dozens more. That is why there is no single entropy measure the Infotropy theory points at: each of those measures is a local formalism that captures one side of an irreversible physical process — usually the dispersion, uncertainty, cost, code-length, or correlation side. Infotropy is the discipline of asking, for each measure, what the flipped record-side / downstream-persistence half looks like — what durable structure survives the irreversible step, what reconstruction it supports downstream, and what test would let you see it. A recent crosswalk of ~96 definition families against the Infotropy theory found zero contradictions. The closest native fits — places where the field already does Infotropy's work, just under its own name — are Landauer's principle, the Sagawa–Ueda generalized second law, Jaynes' maximum-entropy principle, rate-distortion theory, the Tishby information bottleneck, Bennett's logical depth, Zurek's quantum Darwinism, and the biological / cultural record substrates (genetic sequence, epigenetic memory, dual inheritance). The "unclear" cases — differential entropy's coordinate-dependence, computational HILL entropy's adversary-relative pseudorandomness, the active-inference free-energy framing, Floridi's truthfulness clause, negative quantum conditional entropy — become tests, language-boundary calls, or productive falsifiers, not negators. The textbook false friends — algebraic entropy, Tsallis q-entropy, ecological Shannon-Wiener diversity, Lempel-Ziv complexity, Fisher information, Carnap-Bar-Hillel semantic information, Wheeler's "It from Bit," and a handful of others — share a word, a formula, or a metaphor without sharing the underlying ontology. The honest reason they stay out is not that they run on a reversible or costless substrate; it is that nothing in their mathematics generates new, world-laid-down distinctions over time — they count elements of a frozen structure, or the shape of a single snapshot, and so carry no record. (That is also why, in this reading, a measure that does track distinction generated over time can belong even on a fully reversible system — the arrow is carried by the record, not by the heat.) Infotropy is careful not to claim the false friends as support.

The names behind the record side, one line each

The substrate page keeps the roll-call off its surface so the walk can breathe. Here it is. First the lineage the word "entropy" actually passed through:

• Boltzmann — read entropy as the count of micro-arrangements a bulk state hides; the bridge that turned heat-entropy into counting. The macrostate you choose to record is already a record-side choice.
• Gibbs — the statistical entropy of a probability distribution; algebraically the same object as Shannon's. The distribution is fixed by what someone recorded.
• von Neumann — carried the same entropy into quantum states. Tracing out the environment — the partial trace — is the formal stand-in for a record forming: what the kept system inherits about what it was entangled with.

Then the closest native fits — places a field already does the record-side work under its own name:

• Landauer's principle — erasing one bit has a real physical cost. The record side, billed in energy.
• The Sagawa–Ueda result — a feedback system can only do work by paying, in its own books, for the memory it keeps.
• Jaynes' maximum entropy — assume nothing beyond what you actually recorded; the constraints you feed it are themselves record-side facts.
• Rate–distortion — the fewest bits that still rebuild a thing well enough: a record-quality bound the field already had.
• The Tishby information bottleneck — squeeze the input down to just what predicts the thing you care about; the compressed representation that survives is a record built for downstream use.
• Bennett's logical depth — what tells a real record from a random string that only looks full: not its length, but how much work it took to build.
• Zurek's quantum Darwinism — a quantum system leaving redundant copies of itself across its surroundings; the record side surfacing in physics under another century's name.

These are convergences, not corroborations — fields that arrived at the record side on their own terms, which is evidence the lens is useful, not that it has been proven.

If someone asks "where is the entropy equation?"

Infotropy doesn't add a new equation. It adds a reading discipline: for each existing entropy-or-information measure, what is the flipped record-side half? Landauer already showed it for bit erasure; Sagawa–Ueda showed it for feedback systems; Zurek showed it for quantum decoherence; Bennett showed it for algorithmic depth; biology shows it in DNA. The contribution is a unifying lens that reads these as one process viewed from different positions relative to constraint and record formation — not a competing formalism.

The receipts

The full crosswalk — every definition family with its source handles, the worked record-side readings, and the false friends that only share a word — is in the raw files:

• The lay crosswalk summary — the passage this page is built from, with the counts and the counter-question in full.
• The master definition explainer — every definition read from the record side, one at a time.
• The full family inventory — and the false friends set aside, with the reason each only shares a word.