Methodology.
EmergenceLexicon defines 1341 concepts by a single rule: a word's meaning is an explanation of what it emerges from. Every entry names its components, the interaction between them, and the properties that appear only in the whole. Each chain is followed down until it reaches one of four foundations — space, time, energy, pattern. Nothing here is asserted; everything here is a claim a validator can be pointed at.
Why an emergence definition
A conventional dictionary defines words with other words, and those with others again, in circles. The circularity is invisible because no one follows it. Here the graph is the definition: every content word in every entry must resolve to a node at a lower level, so a definition cannot lean on a word that has not itself been explained. Follow any entry far enough and it bottoms out in space, time, energy, or pattern — four sealed terminals — rather than looping back on itself.
How each entry is written
Entries are generated by a language model working from a fixed emergence point of view, then accepted or rejected by deterministic checks. The model supplies the reading; it does not get to decide whether the reading is admissible. An entry is kept only if it satisfies every rule below — the model is re-prompted with the specific violation until it does, or the entry is dropped. Curated strata (the window-ontology entries and folk phrases) are hand-written and are protected in code from being reworded by any model.
What the rules enforce
- Computed levels. A concept's level is
1 + max(level of its components), derived from the graph on every build and never stored. A node cannot assert its own depth. - Definitional closure. Every content word of every definition must resolve, through the lexicon, to a node lower in the graph. Function words come from a sealed primitive set.
- Sealed foundations. Space, time, energy and pattern are the only terminals, and their definitions may use nothing but primitives.
- No cycles. Two concepts cannot define each other; the lower one is written by hand.
- Sense-keyed identity. Nodes are
term.sense, so operation.math and operation.surgery never collide. - Coherence. The stated components must match the words the definition actually uses.
The full rule set, with the failure each rule was written to prevent, is in
schema/RULES.md in the
source repository.
Status, and honesty about it
Every entry carries a status. 28 entries have passed human audit; 1313 are machine-drafted and validator-checked but not yet individually reviewed. All 1341 satisfy the structural rules above — that is what the validator guarantees, and it is a narrower claim than being right. Human audit is a separate, slower pass.
Current corpus: closure 72.4%, deepest chain L26, 0 validation errors, 581 warnings. These counts are written by the build, not maintained by hand, so they cannot drift from the corpus they describe. Machine-readable: stats.json.
Corrections
If a definition is wrong, say so — every page carries a feedback control, and it goes straight to the maintainers. Reported errors are triaged against the rules above; if a definition violates one, it is regenerated or hand-written, so a fix becomes a better rule rather than a silent exception.
Reuse
Text is available under
CC BY-SA 4.0. Each concept is also
served as .json and .md, and
llms.txt describes the corpus for language
models. Attribution: EmergenceLexicon, The Emergence Machine, with a link
to the concept URL.
Track the project
New concepts, new strata, and what the graph starts saying as it fills in — it all lands on the newsletter first. Subscribe on Substack ↗
Last updated 2026-08-22.