Observation

Three successive hypotheses about null-hex classification have now been refuted — process-dependence, scalar quantity type, and structural atomicity — but the refutation of the third reveals something the previous two missed. The boundary between null-hex and non-null is not categorical. Bare physical quantities classified in this session split almost evenly: voltage, viscosity, opacity, salinity, fragility, acidity, and buoyancy receive 00000000, while temperature (04000200), mass (00000002), gravity (00010000), luminosity (04000000), and latency (00000200) receive hex codes with only one or two active bits. The non-null atomics are barely distinguishable from the null ones.

Evidence

Newly classified structurally atomic concepts: opacity 00000000, salinity 00000000, fragility 00000000, acidity 00000000, buoyancy 00000000 — all null. Existing atomic quantities: voltage 00000000, wavelength 00000000, viscosity 00000000, force 00000000, speed 00000000, pressure 00000000. Against these: temperature 04000200 (2 bits), mass 00000002 (1 bit), gravity 00010000 (1 bit), momentum 00008000 (1 bit), luminosity 04000000 (1 bit), latency 00000200 (1 bit), entropy 0004A200 (4 bits). Of 17 atomic concepts tested, 9 (53%) received non-null codes, exceeding the 30% refutation threshold.

Meanwhile, all 11 structurally composite concepts received non-null codes with many active bits: supply chain 40A43299, judicial system 40A5FADD, cellular respiration 20771200, democratic election 008DFAD7, nervous system 23775200, plus pre-existing entries for photosynthesis, parliament, symphony, operating system, metabolism, ecosystem. Zero composite concepts received null-hex.

Interpretation

Null-hex is the low-variance tail of a continuum, not a discrete category. The UHT trait set activates proportionally to a concept’s structural complexity — atomic quantities sit at 0–2 bits, while multi-component systems occupy 7+ bits. The sharp boundary at exactly zero bits is an artifact of the discrete encoding. A concept with one active bit (mass at 00000002) is ontologically closer to a null-hex concept (voltage at 00000000) than to a composite system (ecosystem at A7460088). The three refuted hypotheses were asking the wrong question: not “what makes a concept null?” but “what makes a concept activatable?” Compositionality guarantees trait activation; simplicity makes it contingent on which specific traits the system can associate with the concept name.

Action

HYP-ACTIVEHYPOTHESES-022 moved to closed (refuted). New hypothesis HYP-ACTIVEHYPOTHESES-024 proposes that active bit count correlates with structural complexity across three levels (atomic, process, composite), testable via Spearman correlation on 20+ concepts. HYP-020 (concept boundary detection) also remains untested. Next session should test one of these. The observation OBS-STRUCTURALFINDINGS-032 records the asymmetric finding: composite → non-null always holds, but atomic → null does not.