Contribution
Enaction had an account of individual cognition as sense-making by autonomous systems, and none of social cognition. Embodied critics of theory-of-mind and simulation theory acknowledged interaction but stayed “methodologically individualistic”. The paper supplies a definition of social interaction on which the interaction process is itself an autonomous organization, and the notion of participatory sense-making, so that social understanding can be partly a property of the interaction rather than of either individual.
Key insight
Coordination between coupled systems needs no clever mechanism; pendulum clocks do it. What is special in social encounters is a double influence: the pattern of coordination affects whether the encounter continues, and the encounter affects the likelihood of coordination. That loop gives the interaction an identity of its own, and since movements are the tools of sense-making, coordinating movements coordinates sense-making.
Assumptions
- The enactive framework, in five ideas: autonomy, sense-making, embodiment, emergence, experience (p. 487). An autonomous system is a network of processes with operational closure under precarious conditions (isolated processes would run down).
- Coordination is “the non-accidental correlation between the behaviours of two or more systems that are in sustained coupling, or have been coupled in the past, or have been coupled to another, common, system” (p. 490). Synchrony is one kind; mirroring, anticipation and imitation are others.
- Scope: face-to-face, micro-level, dyadic interactions, “not restricted to linguistic or even human ones” (p. 486); larger groups are expected to extend the account but are not treated.
Key results
- Coordination is easy and often hard to avoid (p. 490–491). Schmidt & O’Brien’s pendulum-swinging pairs phase-lock when they can see each other. Coordination can be absolute (phase-locked) or relative, with phase attraction (Kelso 1995, p. 98).
- Autonomy of the interaction (pp. 492–494). The double influence makes the encounter “an emergent and autonomous process”, “typically a fleeting one”. Interactors “co-emerge as interactors with the interaction”. Fig. 1: interactional coordination (interaction shapes coordination) and functional coordination (coordination serves the interaction).
- Definition of social interaction (p. 493), quoted in the LIT. Body heat at a bus stop is coupling without regulation, so not social; bumping into someone can become social once regulated. Coercion or torture that destroys an interactor’s autonomy is, “at that point”, no longer social (p. 495).
- Participatory sense-making (p. 497), on a spectrum (Fig. 2). Low end: a video-link delay changes the meaning of a pause and A downgrades her praise. Middle: orientation (pointing; Stern’s “I’m gonna getcha”; Currie’s Janet at the window). High end: joint meaning-making (charades).
- Perceptual crossing (pp. 501–502). In Auvray et al.’s minimal virtual-tape experiment, recognition of the partner happens because both partners search for each other: mutual scanning is the only globally stable configuration. “The sense-making that solves the task only happens at the level of the collective dynamics.”
- Implications (pp. 502–504). Social skills are relational; individual incompetence should not be inferred from interactional failure (autism is named). The other is experienced as neither transparent nor opaque, but as shifting with coordination and breakdown.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Social interaction can constitute an autonomous organization in the relational domain | moderate as a conceptual claim | definition and corridor example (pp. 492–494) |
| C2 | Interaction is social only while the interactors remain autonomous | moderate | conceptual, p. 492–495 and footnote 1 |
| C3 | Coordination of movement in interaction coordinates sense-making, and can create meanings no individual had | moderate | examples on a spectrum (pp. 497–501) |
| C4 | Recognition of a partner can be a property of the collective dynamics rather than of individual detection | moderate | perceptual-crossing results reported from Auvray et al. 2006 and Di Paolo, Rohde & Iizuka 2007, not new data |
| C5 | Mirror-neuron and other individual-mechanism explanations re-describe rather than explain social coordination | weak | argument p. 495 |
Concepts
- Sense-making. An autonomous agent’s creation and appreciation of meaning, grounded in the norms its self-maintenance sets.
- Coordination; absolute and relative coordination. As above, after Kelso.
- Interactional and functional coordination. Interaction shaping coordination, and coordination serving the interaction.
- Participatory sense-making. As defined in the LIT.
- Orientation. One interactor modulating the other’s sense-making toward a domain of significance.
Connections
It applies the enactive autonomy of Varela, Thompson and Di Paolo (and the same group’s LIT-566) to social encounters, and takes its dynamics from coordination studies (Kelso 1995, LIT-578, read here through Kelso 2021, LIT-601; Winfree; Kuramoto). It positions itself against theory theory and simulation theory and beyond Gallagher’s, Hutto’s and Thompson’s embodied accounts, and alongside conversation analysis, Trevarthen and Tronick.
Bearing on the record
- Social ontology. The record’s social-ontology holdings are about collective intentionality and institutions. This paper offers a different route to a social entity: a dynamical organization with its own closure and precariousness. It does not discuss collective intentionality.
- Coordination dynamics. Its “relative coordination” is the metastable regime of the extended HKB model in LIT-601; the record can now trace the concept from the hand-movement model to social interaction.
- Psychopathology. The remark on autism (p. 503) points to a relational account of social difficulty; it is a single sentence, and the paper cites De Jaegher 2006 for it.
- No THEORY filed: one was declined, because the source does not carry a claim beyond the definition itself; the evidence is illustrative.
- No instruction for machine-learning practice.
Limitations
- Nothing is measured; examples are chosen to illustrate, and the authors call the implications “speculative at this stage”.
- “Autonomous” for an interaction lasting minutes stretches precarious operational closure; how a fleeting relational pattern has processes that “would run down” in isolation is not spelled out.
- Groups larger than two are set aside.
Open questions
- What measurement would show that an interaction, rather than its participants, sustains a pattern? The paper points to perceptual-crossing paradigms and to measures of coordination but proposes no test.
- How do breakdowns and repairs over many encounters change individual capacities, the developmental route the authors name?