M3 — Socially Scaffolded Regulation via Attachment.

Human emotional regulation is constitutively socially scaffolded. The default architectural expectation of the organism is proximity to trusted co-regulators - not as a preference, but as a structural condition the regulation system was built to operate inside. When that scaffolding is absent, the system does not adapt around its absence; it produces the well-documented downstream costs of social isolation, including elevated baseline cortisol, degraded HRV recovery, increased all-cause mortality, and persistent activation of threat-detection circuits in conditions that should not warrant them. The evidence is converged across five independent literatures: Coan's Social Baseline Theory fMRI program, Bowlby's attachment work, Hrdy's cooperative-breeding primatology, Dunbar's social-brain hypothesis, and Holt-Lunstad's mortality meta-analyses. The full citation chain is in the atlas under C5.

Three of these literatures explicitly engage Dunbar's parameter in their own frameworks: Tomasello and Hrdy build on his layered social architecture directly (shared intentionality and cooperative-breeding respectively), and Henrich cites his social-brain parameter while reframing the underlying mechanism in terms of cumulative cultural learning rather than Machiavellian intelligence. The parameter survives the disagreement, which strengthens rather than weakens its load-bearing role across traditions.

The 150 parameter has direct empirical validation: Facebook friend-count studies on UK national samples (N≈2000 and N≈1375) returned mean friend counts of ~155 and ~183, statistically indistinguishable from the prediction, with only ~14% of users exceeding 300 friends (Dunbar 2016).

The 150 ceiling is the outer cognitive limit on coherent face-to-face relationships, but the layers below it are the structure M3 actually operates on day-to-day: 5 close confidants providing emotional and physical support, 15 in the sympathy group (jury-scale), 50 at the typical hunter-gatherer overnight camp scale, 150 at the clan ceiling, with looser layers extending to 500 and 1500. M3 resolution conditions are typically met or unmet at the 5 and 15 layers; the 150 ceiling matters most for team/floor design, urban planning, and institutional scale, where it sets the upper bound on coherent group cohesion.

This is the first operationalization of a single mechanism, drafted from the converged evidence base and not yet calibrated against pilot data. It is the first concrete example of what the Cor specification - the operational layer being built on top of the atlas - will look like at scale. The full specification will contain a worked operationalization of this shape for every mechanism in the atlas, with thresholds calibrated against actual data and inter-rater reliability tested across independent coding teams. This page exists so that a reader can see, today, what the path from atlas to specification looks like.