Set — the deep split lives in the count, not the material

Deep splits are read in the inventory: a vertebrate reads 3+ OTX switches, an invertebrate 1, under one grammar. The difference is the count, not the material.

SET is the presence/absence and copy count of the bistable switches a genome carries. Where γ is nearly invariant, the deep splits are read in the inventory: a vertebrate reads 3+ OTX-family switches, an invertebrate 1 (each γ in 1.34–1.48, all R19-bistable). The difference is the count, not the material.

What SET is

SET is the most basic Layer-1 read: which bistable switches a genome carries, and in how many copies. A switch “exists” when a readable switch structure is present and its γ clears spinodal (it can turn on). SET is discrete; deep splits (kingdom/phylum) are decided here, while shallow splits (within species) share the SET.

The inventory reads

Vertebrate vs invertebrate — the OTX/otd homeodomain switch.

A vertebrate reads 3 OTX-family switches (+DMBX1 → ~4); an invertebrate reads 1. Same homeodomain grammar (every γ in 1.34–1.48, all R19-bistable) — what differs is the count.

Plant vs animal — which switch families are present. Both carry eukaryotic switch grammar (bHLH·MYB·homeodomain TFs on both sides). The SET-level difference is which families a genome carries: an animal genome carries a gut-formation + cell-migration + neural-crest regulatory network; a plant genome does not (a cell wall makes migration impossible, so morphogenesis runs by oriented division + a morphogen field).

The deep split is in the inventory

Because γ is nearly invariant across taxa, it cannot be where plant/animal or vertebrate/invertebrate part. They part in which switches the genome carries. More switches under the same grammar afford finer Layer-1 control (neural crest, placodes, a more elaborate brain). Whether the count co-varies with any readable complexity index is left as an open question, asserted with no causal claim.

Article IV boundary — count alone is a pointer

Locating an element by similarity is a pointer. So the admissible statement is “a switch that can turn on is here” — an intact structure with γ above spinodal — while a bare count of homologous copies is a pointer until tied to that structure. The OTX count above is therefore reported tied to each copy's measured γ; a bare count is never used as evidence.

Failure log

• Count alone = pointer: the OTX count is reported only tied to each copy's measured γ.
• γ blind to SET: γ does not carry the count; fly oc γ 1.340 vs human OTX 1.42–1.48 differ by GC, not by count.
• Toolkit not directly measured here: the plant/animal toolkit presence/absence is principle-demonstration, not measured in this session — open for a direct read.