The case
A stipulated design, so that the criterion has a concrete test.
- Participants and conditions. Several experienced couples in an improvised partner dance (lindy hop, tango or salsa) to recorded music. A control condition records the same two people standing and talking after the song. A second control pairs each dancer with a stranger, so a practised pair can be compared with a new one.
- Data. Motion capture of both dancers’ trunks and arms at a rate fast enough for the dance’s timescale. The music’s beat is taken as an exogenous input. The target is one-beat-ahead positions.
- Analysis. Fit a VAR model to the joint kinematics. Group the variables by dancer, and by pair against the rest. Compute the gap that separates the pair’s block and its Occam factor (CLAIM-154) in each condition.
- Perturbations. A mild, unannounced resistance applied to one dancer’s arm. Does the partner compensate within the beat (LIT-902)?
- Architecture. Estimate the lead→follow and follow→lead couplings separately. Their ratio says whether the pair works as peers or as a hierarchy (CLAIM-167).
- Drive profile. Take the pair’s relative phase as the macro-variable and estimate the information decomposition’s downward-causation and causal-decoupling terms (LIT-025) over windows (TERM-051). The question is whether the dance has dynamics of its own that neither dancer accounts for (decoupling), whether it constrains the dancers (downward), or whether it is only their resultant.
What it can show
- If the pair clears the threshold while dancing and not after, the owner’s claim is a measured result, relative to this target.
- If the pair clears it but does not compensate, that counts against CLAIM-157.
- If the follow→lead coupling is near zero while the pair clears the threshold, the pair is one system with a hierarchical architecture: the lead drives. That is a finding about its structure, not against its unity. How far follows feed back, through styling and hijacking, is then an empirical question about the dance.
It is stipulated. Nothing here has been run, and the design’s choices (the target, the linear model, the prior variances) are exactly what CLAIM-152 says a verdict is relative to.
Note of 2026-10-23: part of the aim met elsewhere
Bigand et al. (2024, LIT-1424, read; CASE-149), cited at A5 of the exchange of 2026-10-23, “Framework Overview Request”, varied music and partner visibility independently in 70 untrained dancers in pairs. “We find that both music and partners drive synchrony, but through distinct dance moves”: the music drove anteroposterior components and the partner lateral and hand components. So the beat, as an exogenous input, acts on particular components, and a whole-body block test could mix the two. Still for this design: direction (their measure is symmetric), a perturbation and compensation test, the block gap, and touch, lead and follow roles and trained dancers; their dancers stood 2.5 m apart and were “without dance training”. A5 proposed treating the dance study as “a reanalysis candidate”.