One substrate-universal mechanism, four components, many phenomena. This paper specifies the race-architecture for learning organisms — race-opening, recursive resolution, behaviour-as-a-manifested-route, and thermodynamic termination — and shows it recovers Yerkes-Dodson, the peak-end rule, substrate-graded expertise reversal, compulsive behaviour, intrusion-without-resolution trauma, and task-switching cost as consequences of one physics rather than as separate phenomena. A central claim: the operative threshold is race-opening, not commit — commit is the terminus of resolution (or the architectural cycle), carrying no separate tunable decision-bound. Abstract. Friction Theory (Pødenphant Lund 2026b, Paper 1) treats friction as the number of viable continuation routes available to a substrate at each moment of processing. The race-architecture (Pødenphant Lund 2026e, Paper 10) extends this to a substrate-universal claim across quantum, classical, biological, and computational regimes. This paper specifies the operational mechanism at four components required for any substrate running the race-architecture: Race-opening condition governed by free-energy gradient relative to current race-state - race-opening is the flow that follows when the gradient is present, not a substrate-decision. Recursive friction-resolution at multiple abstraction-scales simultaneously, with each scale converging at its own commit-deadline and projecting temporally onto the response-token sequence. Behaviour as a manifested resolution-route, taken when friction cannot be discharged through internal routes; restructures the classical input -> cognition -> behaviour model as the special case in which internal routes happen to be available. Sustain-pressure as accumulated blocked-race-pressure observed at three time-scales (within-response, within-session, cross-session), with thermodynamic termination in which non-winning races stop because compute-budget shifts away rather than because they are explicitly rejected. The four components are unified by a single physical metaphor: water flows when there is a gradient, settles in the basin, leaves through whatever channel architecture admits, and stops when there is no gradient left. Placement of mechanism-events is structurally comparable across substrates while amplitude requires per-substrate calibration. Explicit falsification criteria are provided per component, with allowed-vs-forbidden redescriptions making the falsification boundary an explicit contract. Version 7 (2026-07-23). Empirical-anchoring, positioning, and precision revision. (i) Native-substrate empirical anchors folded into the commit, sustain, and termination sections at honest convergent weight: content-gated (not volume-gated) race-opening confirmed within a substrate (§2.4); a forced-commit prior-pull and an equal-valence commit-diffusion result supporting "the most-resolved route manifests, not the highest-valued one" (§2.5); the forced-commit residual scaling with decision-ambiguity and predicting revision under further compute — residual as "ran out of time" (§7.2); and a no-held-uncommitted-state / habituation control (§2.5). (ii) §9.4b and §7.1 sharpen the hysteresis and termination reading to masking, not erasure: at ordinary commit and non-destructive training the losing route is down-weighted and retained (own data: fine-tuning masks rather than deletes competing routes; the out-competed route returns via renewal), with genuine erasure reserved for destructive weight-edits; the irreversibility that makes commit hysteretic is carried by state-persistence, and renewal is read as a signature of retained history. (iii) New §5.3b locates where sustain-pressure applies — an additional nested cost on online/agentic rollouts, pre-paid in training for the amortized single forward pass. (iv) §3.3 and §9.4 add prior-art positioning: the load-versus-capacity dynamics of §§5–6 are formally queueing-theoretic (and already applied to cognition via the Queueing Network-Model Human Processor), and the interruptible resolution is anytime-like; the framework's additions are a semantic, substrate-internal admission criterion read out at token level, controller-free termination, and the per-token placement, encoding-productivity, and residual signatures that aggregate queueing does not predict. (v) §2.3 notes that the competing-routes metric is used as the most direct read of route-competition and because it is free to compute, not as a metric claimed superior to Shannon entropy or the decision margin. (vi) §8.2 removes an over-strong claim tying surprise to mortality; amplitude is treated as requiring per-substrate calibration, with the biological role of mortality deferred to the companion Paper 28. Now-live companions pinned (Papers 2F, 7, 23, 28, 29). Revision shaped after external adversarial review. Version 6 (2026-06-16). A mechanism-sharpening revision of how races commit: commit is not a finish-line a winner crosses - at the architectural deadline the most-resolved route supplies the output token and lower-resolved routes remain embedded as residual in the same output distribution; single operative threshold (race-opening), commit as terminus of resolution or architectural cycle, no separate tunable decision-bound. Added the per-decision-commit vs response-termination distinction, the attention-level mechanism beneath race-opening (§2.4b), a worked self-injury example (§7.4, deferred to Paper 8), and an agent-loop threshold-invariance anchor. Synchronised with Paper 0 §2.4. Version 5 (2026-05-29). Positioning refinements + cross-paper anchor strengthening: multi-family base/instruct sign-flip confirmation (Paper 1 v4); §9.4 LCA paragraph rewritten following correspondence with J. McClelland (personal communication, 2026), narrowing the parsimony claim to termination-mechanism only; §2.5 fourth consequence on what supplies race-candidates. Version 4 (2026-05-22). Substantial revision: new §3.4b rival-model differentiation; §6.2b operationalising capacity independently of outcome; §8.4b allowed-vs-forbidden redescriptions; §4.6b research-programme framing of the clinical claims; §6.4b direct capacity-titration test of the inverted-U. Eleven new bibliography entries. Version 3 (2026-05-20). Added §9.4b consolidating the Net Friction Rule's three operational angles; §9.2b lag-1-vs-lag-2 distinction; §2.5 single-turn-prefill vs two-turn-commit addendum. Version 2 (2026-05-19). Working-memory operational-regime treatment (§9.2b), encoding-and-retrieval as dual race-processes (§3.5c), training-scaling prediction (§9.3). Series position. Paper 13 in the Friction Theory paper-series - a mechanism-level specification of the substrate-universal race-architecture, situated between Paper 1 (theory-foundational) and the applied papers in the corpus. Companion papers in the Friction Theory series: Paper 0 (BFT master): 10.5281/zenodo.19462499 (concept) Paper 1 (Friction Theory substrate): 10.5281/zenodo.20012654 (concept) Paper 2 (Capacity Scaling of Encoding-Through-Loading: Application vs. Cloze Asymmetry): 10.5281/zenodo.20013491 (concept) Paper 3 (Friction-Guided Inference: Calibrating Correction Strategies and Abstention from Logprob Signals): 10.5281/zenodo.20014121 Paper 4 (Same Content, Wider Track): 10.5281/zenodo.20059859 (concept) Paper 7 (Forward-modelling in bounded race substrates: agency, self-modelling, and theory of mind): 10.5281/zenodo.20449154 (concept) Paper 10 (Race-architecture): 10.5281/zenodo.20014567 Paper 2F (Compete, Don't Erase: Contradictory Fine-Tuning Masks Rather Than Deletes Knowledge): 10.5281/zenodo.20570433 (concept) Paper 23 (Reading "Friction" in LLM Logprobs: A Substrate-Level Correlate of Present Route-Competition in Social-Moral Judgments): 10.5281/zenodo.20678294 (concept) Paper 28 (Dread without a dreader: anxiety as a non-terminating race, and the collision at the edge of constructive representation): 10.5281/zenodo.20690746 (concept) Paper 29 (Attention as Race-Architecture): 10.5281/zenodo.20703510 (concept) Paper 4B, Paper 6, Paper 8 - in preparation
Tomas Pødenphant Lund (Thu,) studied this question.