Sedimented experience names the mechanism through which expertise accumulates in the embodied organism. The experienced programmer's sense that a codebase is fragile, the surgeon's sense that a tissue plane is not where it should be, the editor's sense that a sentence is wrong before she can articulate the rule it violates — each is drawing on sediment: the accumulated residue of thousands of hours of structural coupling with the domain, deposited in the body as a set of dispositions that orient perception and action pre-reflectively. The felt sense is not a hunch or an irrational impulse. It is the organism's deepest form of cognition, refined through repeated engagement, and constitutive of the skilled practice that distinguishes the expert from the novice.
The concept is Merleau-Ponty's, from Phenomenology of Perception (1945), and Thompson adopts it as the mechanism through which affective framing is constructed. The expert's affective response to a situation is not arbitrary — it is shaped by her sedimented history with that kind of situation. The sedimentation is specifically embodied: it lives in the hands of the skilled craftsperson, in the ear of the musician, in the visual system of the radiologist, in the full organism-scale pattern of responses that the expert has developed through years of practice.
The concept provides Thompson with his most precise diagnosis of what is at risk in the AI transition. AI tools cannot develop sedimented experience, because they have no bodies to sediment into. More consequentially, AI tools can disrupt the development of sedimented experience in their human users by substituting for the structural couplings through which sedimentation occurs. The developer who debugs through Claude rather than through manual engagement does not sediment the tactile sense for code in the same way. The loss is not dramatic; it is incremental and invisible, revealed only when the practitioner discovers, months or years later, that she can no longer feel the wrongness in a system through the embodied sense she once possessed.
This does not entail that AI mediation necessarily destroys sedimented expertise. It entails that the substitution of AI-mediated engagement for direct engagement has consequences for sedimentation that must be examined. The senior engineer who developed her expertise through decades of manual practice brings that sediment to her AI collaborations and gets more out of the tool because of it — a phenomenon that The Orange Pill observes without fully explaining. The junior engineer who uses the tool from the beginning develops a different sediment, adapted to the new couplings. Whether the new sediment supports the kind of judgment the senior sediment supported is an empirical question that the practice has not yet had time to answer.
The concept of sedimentation is introduced in Merleau-Ponty's Phenomenology of Perception (1945). Thompson develops its cognitive implications in Mind in Life and subsequent work.
Expertise is sediment. The expert's judgment is the deposited residue of thousands of hours of embodied engagement with the domain.
The body remembers what the mind cannot articulate. Sedimented experience operates pre-reflectively, shaping perception and action below the threshold of conscious awareness.
AI cannot sediment. It has no body in which experience could deposit; it has only training data, which is a different kind of accumulation.
AI mediation changes sedimentation. Substituting AI-mediated engagement for direct engagement produces different sediment, with consequences the practice is only beginning to perceive.