Laughter as cognitive-load release and inter-regime social signal

Status: SEED (candidate claims registered 2026-06-09; V4/V5, off the V3 critical path) Source intake: evidence/planning/thought_intake_2026-06-06_laughter_cognitive_load_social_reconfiguration.md Raw thought: docs/thoughts/2026-06-06_laughter_cognitive_load_social_reconfiguration.md

This doc is the home for the two laughter claims that the 2026-06-06 intake added on top of the existing respiratory-cluster laughter model. It is a seed / boundary doc, not a build spec — there is no V3 substrate work implied here.

Source-weight caution (read first)

The prompting article — Neuroscience News, “Laughter Rewires Brain Architecture and Lowers Cognitive Load” — summarises a cross-disciplinary popular-science synthesis (Dr. Jacqueline Harding, The Brain That Loves to Laugh, Routledge/Taylor & Francis), not a primary empirical study. “Rewires brain architecture” is editorial framing for neuroplasticity + distributed prefrontal activation, not a measured large-scale network-regime reconfiguration. The claims below are compass-level hypotheses, registered at candidate / substrate_conditional / v4 — they are not backed by citable experimental evidence and must not be cited as such.

What REE already owned

MECH-110 — Laughter is rapid repeated hypothesis-tag cycling: each forced exhalation is one E3 plan-sweep (MECH-107) that activates a threat-hypothesis tag and clears it on safe resolution; the pleasure is the affective signature of iterated safe-confirmation. Includes a V5-scoped social extension (group/contagious laughter = respiratory phase-reset synchronising E3 plan-sweep clocks across agents). Home: control_plane_heartbeat.md.
Respiratory cluster MECH-107 (exhalation = plan-sweep / trajectory-abandonment), MECH-108 (respiratory rhythm = E3 plan-sweep oscillator).
Modes cluster ARC-016 (modes ARE control-plane regimes on the precision-to-commitment circuit), MECH-027 (pathological modes = mis-tuned regimes). Home: modes_of_cognition.md.
Play-mode cluster ARC-049/050 / INV-058–060 / MECH-194–199 / Q-035 — laughter appears in the basic-expression catalog as a bilateral frame-opening signal. Substrate-blocked in V3 (no play_frame_tag); do not queue probes against it.

The central idea of the intake (laughter = rapid threat→safe-resolution reframe, incl. social synchrony) is therefore already owned. A new laughter claim restating it would duplicate MECH-110.

What is new (the two registered claims)

MECH-364 — laughter discharges E3 conflict load and marks an inter-regime boundary

mechanism_hypothesis · candidate · substrate_conditional · v4 · v4_v5 depends_on: MECH-110, MECH-107, ARC-016, MECH-027, MECH-076

MECH-110 is the micro mechanism (per-cycle tag-clear). MECH-364 is the macro/regime-level consequence: the cumulative effect of iterated safe-confirmation is a drop in E3 active constraint/conflict pressure — a regime-level load discharge — and that discharge is what marks/enables a safe transition between control-plane regimes (ARC-016): from threat-vigilant / high-commitment to exploratory / play.

Mapping the intake’s primitive laughter_event = f(conflict_detected, threat_reclassified, social_safety, reframe_success, load_release, shared_alignment): load_release + the regime-boundary role → MECH-364; per-cycle reframe_success/tag-clear → MECH-110; shared_alignment → MECH-110’s V5 social-synchrony extension.

Failure taxonomy (predictions). A system that cannot run reframe → load-release → regime-re-open cannot soften rigid commitments after a benign reframe → attractor lock-in (MECH-076), pathological-regime persistence (MECH-027), social-threat overclassification, feedback entrapment, belief fixation, excessive literalism.

Substrate note / V3 reconsideration fork. load_release needs an E3 conflict/constraint-load readout. ARC-016 already exposes an E3-derived prediction-variance signal in V3 — so a V3 substrate check could reclassify MECH-364 toward implementation_phase: v3. Held at v4/substrate_conditional pending that check. Do not build or queue a probe in V3 until routed by an explicit version decision; a probe before the readout exists would be vacuous (cf. the play-mode cluster).

Q-059 — crying analogue + laughter repair-vs-damage adjudication

open_question · candidate · substrate_conditional · v4 · v4_v5 depends_on: MECH-110, MECH-364

REE owns a laughter mechanism (MECH-110, MECH-364) but no crying / tears / sobbing claim exists (only “distress vocalisation” inside the play-mode catalog, INV-058). The raw thought pairs laughter with crying as the two halves of a social-release axis. Open questions: (a) does the high-arousal / distress-signal counterpart deserve its own claim? (b) how is the valence of social laughter adjudicated — when does it repair the social field vs damage/exclude? — given that laughter is not monotone-affiliative (it can exclude, dominate, mask distress, or signal threat). Set substrate_conditional explicitly so narrow_open_question does not fire — V4/V5-parked, awaiting a social substrate, not a V3-tractable question.

Deliberately NOT registered

A restated laughter mechanism — already MECH-110.
The social-synchrony / contagious-laughter strand — already MECH-110’s V5 extension.
The play-mode transition strand — already the ARC-049/050 cluster (substrate-blocked).
The non-affiliative dark side as a standalone assertion — folded as a caveat on MECH-364 and as the motivating tension inside Q-059, rather than its own claim, to avoid over-registration.

Known cost

substrate_conditional is not in the workset proposal-suppress set on its own, but implementation_phase: v4 + version_relevance: v4_v5 are now suppressed from the /queue-experiment proposal lane (IGW _is_deferred_beyond_v3, landed 2026-06-09). So MECH-364/Q-059 should not surface as V3 experiment proposals. If one does, it is a blocked_substrate STOP (play-mode-cluster pattern) — do not queue a probe until a version decision routes this to a build.