Situated Cognition

In the AI Story

The framework emerged from Lave's dissatisfaction with the cognitivist paradigm that dominated psychology and early artificial intelligence. That paradigm assumed intelligence was computation — the manipulation of representations according to rules — and that knowledge was information that could be encoded, stored, transmitted, and applied independently of the context in which it was produced. Lave's ethnographic work in communities of practice from Liberian tailoring workshops to Orange County supermarkets produced evidence that this model systematically misrepresented how human cognition actually functions.

The framework does not deny that propositional knowledge exists or that information can be transferred. It insists that the location of intelligence has been misidentified. Intelligence is not extractable and measurable in isolation. It is a property of the person-acting-in-a-setting, a phrase Lave developed to emphasize that the setting is not a container within which cognition happens but a constitutive element of the cognition itself. Remove the setting, and the intelligence degrades — not because the person has become less intelligent, but because the intelligence was never solely in the person.

This insight connects directly to work by Lucy Suchman on plans and situated actions, by Edwin Hutchins on distributed cognition, and by the broader tradition of embodied cognition. What Lave contributed was the anthropological evidence — the patient ethnographic documentation of how cognition actually works in the supermarket, the workshop, the Weight Watchers meeting — that made the theoretical claims unavoidable.

For the AI revolution, situated cognition is the most rigorous available challenge to the foundational assumptions of large language models. If knowledge is situated, then a system trained on the decontextualized textual residue of situated human practice cannot produce knowledge in the same sense that situated practitioners produce knowledge. It can produce outputs that match situated outputs statistically. But the outputs arrive without the context in which situated knowledge is constituted — and the practitioners who rely on them develop, over time, a thinner form of understanding.

Origin

Lave developed the framework through fieldwork beginning in the late 1960s among Vai and Gola tailoring apprentices in Liberia, and elaborated it across studies of adult arithmetic in American supermarkets, kitchens, and Weight Watchers meetings. The theoretical formalization appears in Cognition in Practice (1988) and Situated Learning: Legitimate Peripheral Participation (1991, with Etienne Wenger).

Key Ideas

Cognition is not individual. It is a property of the relationship between a mind and its environment, including tools, other people, and accumulated history of engagement.

Context is constitutive, not incidental. The specific setting in which knowledge is produced shapes the knowledge itself — not merely the conditions of its acquisition.

Transfer is weak. Knowledge produced in one context often does not apply directly to structurally identical problems in another context, because the structural identity is an abstraction that ignores the contextual features that made the original knowledge reliable.

The location of intelligence is the person-in-setting. Not the person alone, not the setting alone, but the specific configuration through which cognition is achieved.

Debates & Critiques

The framework's strongest critics argue that Lave overstated the failure of transfer — that while context-sensitive cognition is real, there is substantial evidence for transfer of abstract skills across domains, particularly with well-designed instruction. Cognitive scientists such as David Perkins and Gavriel Salomon have proposed a middle position: transfer happens, but it must be deliberately cultivated rather than assumed. The AI-era question is whether this middle position is tenable when the tool in question is not an instructional intervention but a substitute for situated engagement altogether.