Observation

Four of five control theory concepts landed exactly where expected — and one didn’t land at all. System stability classified at 00010208, activating only System-Essential, Temporal, and Economically Significant, with zero confidence on every other trait. Meanwhile PID controller (40B23200), closed-loop control system (41B73200), transfer function (40A0A000), and state-space model (40A0B200) formed a coherent cluster in the Functional/Abstract layers, sharing Synthetic, Intentionally Designed, Processes Signals/Logic, and Rule-Governed traits. The question was whether control theory — a domain sitting between pure mathematics and physical engineering — would produce a tight semantic cluster or scatter across layers. The answer: tight for mechanisms and representations, degenerate for properties.

Evidence

PID controller to closed-loop control system: Jaccard 0.727, Hamming 3. PID controller to feedback loop (pre-existing entity): Jaccard 0.700, Hamming 3. PID controller to state-space model: Jaccard 0.667, Hamming 3. PID controller to transfer function: Jaccard 0.444, Hamming 5. PID controller to system stability: Jaccard 0.100, Hamming 9. The four well-classified entities share a common Functional layer signature (B2–B7 range in bits 9–16) indicating signal-processing, state-transforming, rule-governed behaviour. Transfer function diverges slightly by gaining Symbolic (bit 17) and losing State-Transforming and Functionally Autonomous — correct, since it’s a representation rather than an active mechanism. System stability’s hex 00010208 shows no Physical or Functional layer activation whatsoever. Its three active traits (bits 16, 23, 29) appear to be artefacts of an under-specified classification context rather than genuine semantic content.

Interpretation

UHT handles control theory mechanisms well. The distinction between active controllers (PID, closed-loop) and passive representations (transfer function, state-space model) emerges naturally through the State-Transforming and Functionally Autonomous traits. The strong overlap with the pre-existing feedback loop entity (0.70) confirms that the taxonomy recognises structural kinship across its cybernetics and engineering lineages. The failure on system stability is informative: it’s a property of a system, not a mechanism or representation. UHT’s 32-bit scheme is designed for entity-like concepts — things that are or do. Abstract properties (“the quality of being stable”) slip through the trait grid because they lack physical embodiment, functional behaviour, and social significance simultaneously. This is consistent with the earlier observation that loss of control also classified at 00000000. Properties-as-entities may need a dedicated classification strategy or richer contextual priming.

Action

Five entities added to new namespace SE:control-theory. Corpus log entry CORPUS-LOG-005 created in uht-research with baseline BL-UHTRESEARCH-009. System stability flagged at urgency low for potential reclassification with enriched context (e.g., “BIBO stability criterion in control engineering” rather than the abstract property). A future calibration session should test whether property-type concepts systematically underperform on the trait grid — sampling 10 property-like entities and 10 mechanism-like entities and comparing mean trait activation counts. The existing feedback loop entity now has four natural companions for cross-domain reasoning in control systems.