Enskilment

In the AI Story

The concept challenges a deep assumption in contemporary educational and corporate practice: that knowledge can be codified, transmitted, and acquired efficiently through well-designed instructional materials. Ingold's ethnographic observation is that the most important kinds of professional knowledge — the hunter's reading of tracks, the surgeon's feel for tissue, the engineer's intuition for where a system will fail — cannot be codified, because they are not information. They are perceptual capacities, developed through practice, that exist in the practitioner's trained sensorium rather than in her stored memory.

The practical implication is that shortcuts to skill are illusory. The novice who reads a manual, watches videos, or receives an AI-generated summary of expert practice does not acquire the skill; she acquires information about the skill. The skill itself requires the novice to undergo the friction-rich process of practice — the mistakes, the corrections, the slow cultivation of perceptual sensitivity — that alone deposits the capacities that constitute expertise. Ingold's framework is not hostile to instruction; teachers and texts play crucial roles in structuring the practice. But the instruction cannot replace the practice.

AI intensifies the stakes. The tool offers outputs that appear to be the products of expertise — working code, competent analyses, serviceable designs. If the market rewards the output, the process appears dispensable. Why go through the friction-rich process of enskilment when the output is available by prompt? The answer Ingold's framework provides: because the output is not the expertise. The expertise is the cultivated capacity to evaluate the output, to recognize when it is wrong, to adapt it to novel situations that no training data anticipated. Without enskilment, the practitioner cannot perform these meta-level judgments. She has the artifact without the understanding.

This connects directly to the debate about AI and education. If AI produces the output, and the market rewards the output, the incentive structure pushes students and institutions toward producing outputs rather than developing capacities. The long-term consequence is a workforce that can produce but cannot judge — that can direct tools but cannot evaluate their outputs with the depth that enskilment alone cultivates. The framework thus joins the growing literature on deskilling, but sharpens it: what is lost is not merely a specific technique but the perceptual infrastructure on which all technique-use depends.

Origin

Ingold developed the concept of enskilment in The Perception of the Environment (2000), drawing on his fieldwork on Sámi reindeer herding and the broader anthropological literature on skilled practice. The term itself is Ingold's coinage, though the underlying idea draws on Gibsonian ecological psychology, phenomenology, and Jean Lave and Etienne Wenger's work on situated learning and legitimate peripheral participation.

The broader intellectual context is the critique of cognitivism in anthropology and cognitive science — the insistence that mind cannot be understood apart from body, and that skill cannot be understood apart from the environment in which it develops.

Key Ideas

Learning is doing. Skill develops through the learner's own sustained practice, not through the transfer of information from expert to novice.

The teacher structures but cannot substitute. Instruction shapes the environment for practice; it cannot replace the practice itself.

Skill is perceptual. What the expert knows lives in her trained sensorium, not in a stored representation that she retrieves and applies.

Shortcuts are illusory. Information about expertise is not expertise; the friction-rich process of practice is the mechanism by which capacities are cultivated.

AI threatens enskilment. The tool offers outputs without the practice, producing a generation that can direct but cannot judge.

Debates & Critiques

The most serious challenge to the enskilment framework is the claim that some skills — especially in abstract domains — may be genuinely acquirable through instruction alone. Mathematicians debate whether mathematical intuition requires embodied practice. Programmers debate whether debugging skill can be taught without hours of actual debugging. Ingold's framework is strongest for skills with an irreducibly perceptual or bodily component; it is more contested for skills that are primarily cognitive.