Seeds, explained
A seed is a short Markdown note from which a rule or a piece of knowledge grows. Each page here covers one folder of seeds, with one section per seed: what it says, when it applies, examples, sources, and the rules or machine-readable blocks it defines.
| page | folder | what |
|---|---|---|
| Global-level seeds (global/seeds) | global/seeds | the global level: values, concepts and categories, verb classes, links, relations, measures, senses of words written alike; compiled into Rust tables by build.rs |
| English seeds: morphology (en/seeds/morph) | en/seeds/morph | English morphology: tags, lemmas, word formation |
| English seeds: lexicon and valency (en/seeds/lexicon) | en/seeds/lexicon | word classes the parser needs to know (copula-like verbs, indirect-object verbs, …) |
| English grammar seeds: the noun phrase (en/seeds/nominal) | en/seeds/nominal | noun phrases: determiners, pronouns, numbers, names |
| English grammar seeds: verbs and clauses (en/seeds/verbal) | en/seeds/verbal | verbs and clauses: auxiliaries, objects, complements |
| English annotation-error seeds (en/seeds/errors) | en/seeds/errors | typical annotation errors and the rules that catch them |
| The Mova dialect of Universal Dependencies (dialects/mova) | dialects/mova | Mova's dialect of Universal Dependencies and the converters to and from it |
Grammar seeds in en/seeds carry ```rule blocks that the rule engine en::expert loads at run time (en expert-check); a test gate checks that every seed parses. Global seeds are compiled at build time.
The pages were written for release 0.3 (2026-10-03). The grammar pages cover every seed in their folders as of 2026-10-15. On the global page, measures, selection and homographs were added on 2026-10-15.
Related: Worlds and knowledge (the absurdity matrix), Training Mova (rules the model induces itself), Training Mova, The English layer, Pragmatics: frames — who did what to whom, why, and what came next.
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/repo-docs/seeds/