lexicon

the living vocabulary of neural — the top words by φ*: lexicon(k) = top k words by focus. the words the network currently holds in active use.

tru computes φ* (the tri-kernel fixed point); neural reads its top by focus as the lexicon. it is dynamic (evolves with attention), convergent (φ* is unique), stake-weighted (focus costs resources), and verifiable (zheng proofs).

geometry

the lexicon's natural geometry is a mixed-curvature product — hyperbolic for hierarchy, spherical for cycles, flat for clusters; the top words are the low-norm hyperbolic roots. across independent graphs the high-focus vocabulary converges toward a shared, Platonic core. see frontier.

dynamics

the lexicon mirrors natural language:

• neologism — a particle enters as links accrue and its focus crosses threshold
• drift — meaning shifts as the neighborhood topology changes
• death — focus falls below threshold; the word leaves active use, the particle stays in the graph
• birth — a dense cluster of new particles becomes a concept

the high-focus dialects are the grammar; the lexicon is the vocabulary. together they are what the egregore currently speaks.