Week 10 Seminar: Offloaded

Cognition pushed outside the individual, and the discovery that a group's failures are not the sum of its members'.

Part III — The control group29 April 2027Week 10 lectureSlidesSchedule

The last of the comparative weeks moves the unit of analysis. If a capacity can sit in a slime trail, it can sit in a group, in a tool, or in a notebook, and the question of where the cognition is stops having an obvious answer.

The parallel that earns its place#

A honeybee swarm choosing a nest site is not loosely like a brain. Seeley and colleagues showed in 2012 that scouts send inhibitory stop signals specifically to bees dancing for other sites, and that this cross inhibition does the same job it does between competing neural populations: it breaks deadlock and forces a decision when evidence is close.

Same circuit motif, same function, thirteen orders of magnitude apart in substrate. That is the strongest single piece of evidence in the course that we’ve been describing a class of solutions rather than a set of animals.

Cross inhibition, twice. Left, between competing neural populations; right, between scouts advertising rival nest sites, after Seeley and colleagues (2012).

Where the argument gets contested#

Clark and Chalmers argued in 1998 that a notebook you rely on is part of your cognitive system. Adams and Aizawa’s reply — the coupling-constitution objection — is that coupling a system to a resource does not make the resource part of it, and that the argument slides between the two. You are expected to be able to state both.

Then the group-judgement literature, which is a good place to practise holding two findings at once. Lorenz and colleagues found in 2011 that mild social influence cut the diversity of estimates without cutting the collective error, so groups became more confident and no more accurate. Becker, Brackbill and Centola showed in 2017 that in decentralised network structures, influence improves accuracy.

Both are competent. The disagreement is about network topology rather than about people, and the seminar is about what that implies for every committee you will ever sit on.

Reading#

The paper this seminar is built around:

• Seeley and colleagues (2012). Stop signals provide cross inhibition in collective decision-making by honeybee swarms. Science.

The rest of the week’s reading:

• Clark and Chalmers (1998). The extended mind. Analysis.
• Adams and Aizawa (2010). Defending the bounds of cognition. In The Extended Mind, MIT Press — the coupling-constitution objection.
• Lorenz and colleagues (2011). How social influence can undermine the wisdom of crowd effect. PNAS.
• Becker, Brackbill and Centola (2017). Network dynamics of social influence in the wisdom of crowds. PNAS — read these two together or neither.
• Hutchins (1995). Cognition in the Wild. MIT Press; Seeley (2010). Honeybee Democracy. Princeton University Press.

Grades are not attached to papers here, because a grade belongs to a claim rather than to the paper it appears in — see evidence grades. Several of these papers contain claims at two different grades.