Correspondence (Ingold)
Correspondence is Ingold's replacement for the information-transmission model of making. Where communication implies a message sent from a determinate source to a determinate receiver, correspondence implies a sustained mutual attentiveness between two parties, each adjusting in real time to what the other does. The potter corresponds with clay. The weaver corresponds with thread. The hunter corresponds with the landscape through which she moves. In each case, the knowledge that develops is not stored in the practitioner's mind and deployed when needed — it lives in the relationship itself. Sever the relationship and the knowledge does not migrate to the practitioner's brain. It ceases to exist. This framing, developed across four decades of ethnographic fieldwork and given its definitive statement in Ingold's 2021 book Correspondences, dissolves the conventional picture of skill as a personal possession and relocates it in the ongoing engagement between a living being and a responsive world.
The Material Infrastructure of Attunement — Contrarian ^ Opus
There is a parallel reading that begins not with the phenomenology of skilled practice but with the political economy of what makes correspondence possible. The potter's correspondence with clay depends on access to a kiln, a wheel, a workspace, time freed from wage labor, and a market or community that values handmade ceramics. The weaver's correspondence with fiber requires looms, materials, and buyers willing to pay more than industrial textile prices. These are not mere background conditions — they are the substrate without which correspondence cannot occur. Ingold's framework, by locating knowledge in the relationship between maker and material, naturalizes what is actually a highly contingent arrangement requiring specific economic and social conditions.
The AI moment does not break correspondence so much as reveal how rare genuine correspondence has always been under capitalism. Most workers have never corresponded with their materials — they operate machines according to protocols, follow scripts in service encounters, enter data into systems designed by others. The prompt-execute cycle that Ingold critiques as non-correspondence is continuous with the command-execute structure that has organized most labor for centuries. What AI threatens is not correspondence as such but the small protected zones where correspondence has been allowed to persist — the research lab, the design studio, the writer's study. These spaces were always exceptions, maintained by institutional structures that could afford inefficiency. As those structures dissolve under market pressure, we discover that correspondence was never a general human capacity but a luxury good, and the question is not whether AI enables correspondence but who can still afford it.
In the AI Story
The concept emerges from Ingold's refusal of the sender-receiver model that dominates information theory and cognitive science. In that model, a message is encoded at one end, transmitted through a channel, and decoded at the other end. Meaning is what survives the journey. Correspondence rejects the model entirely. There is no message, no channel, no decoding. There is only the ongoing, mutual adjustment of two parties whose responses to each other constitute the activity. The potter does not receive a message from the clay. She attends to its moisture, its plasticity, its tendency to sag, and the clay — a non-intentional material — responds to her pressure, her speed, her angle. The form emerges from the interaction. Neither party authored it.
The practical consequence is that skilled knowledge becomes a property of a relationship rather than a property of a person. This inverts the entire conventional account of expertise. Expertise, in the standard view, is something the expert has — years of training have deposited it, and she carries it with her from context to context. Ingold's view: expertise is something the expert does, in correspondence with a specific medium, and the doing cannot be abstracted from the medium without the expertise dissolving. The senior engineer who 'feels a codebase the way a doctor feels a pulse' has developed that feel in correspondence with a specific codebase over years. Move her to a codebase she has not lived with, and the feel does not migrate. It has to be cultivated again.
For the AI moment, the implication is sharp. The prompt-execute cycle is not a form of correspondence. It is a form of communication — a message sent to a machine that processes it and returns a result. The mutual responsiveness that constitutes correspondence has no foothold in the exchange. The user specifies, the machine produces, and the back-and-forth adjustments through which correspondence deepens are absent. This does not mean AI work produces no knowledge. It means the knowledge it produces is of a different kind — propositional rather than relational, connectable rather than grounded, portable rather than particular. Whether this kind of knowledge can sustain a practice over time is the question Ingold's framework most sharply poses.
The concept also reframes authorship. If making is correspondence, then what is made belongs to neither party alone. The pot belongs to the correspondence between potter and clay. The cloth belongs to the correspondence between weaver and fiber. The book belongs to the correspondence between writer and language, and — in the AI case — between writer, language, and machine. This dissolves the comfortable fiction of the autonomous creator while also refusing the anxious question of whether the AI 'really' contributed. The question is malformed. Contributions cannot be disentangled from the correspondence in which they emerged.
Origin
Ingold developed correspondence across a long arc of ethnographic and theoretical work, drawing on Merleau-Ponty's phenomenology of embodied perception, Gibson's ecological psychology, and his own fieldwork among the Skolt Sámi of northeastern Finland, where his encounter with reindeer herders' embodied engagement with a living landscape made the limitations of information-theoretic models of knowledge impossible to ignore.
The 2021 book Correspondences gives the concept its most explicit treatment, but the idea runs through his earlier major works — The Perception of the Environment (2000), Lines: A Brief History (2007), and Being Alive (2011) — as the theoretical through-line connecting his critiques of hylomorphism, his account of wayfaring, and his analysis of the meshwork.
Key Ideas
Mutual responsiveness, not transmission. Correspondence is the continuous adjustment between two parties attending to each other, not the sending and receiving of messages.
Knowledge lives in the relationship. Sever the correspondence and the knowledge does not relocate to the practitioner — it ceases to exist, because it was never a possession.
Authorship dissolves into the correspondence. What is made belongs to the mutual engagement, not to either party alone, which reframes every question about AI contribution.
AI breaks the mutual condition. The prompt-execute cycle is communication, not correspondence; what the machine lacks is not intelligence but the capacity to participate as a genuine correspondent.
The diagnostic standard for making. Correspondence provides the measure against which the quality of a making process — not its product — can be assessed.
Debates & Critiques
The concept is contested at two points. First, whether correspondence requires consciousness on both sides: Ingold insists it does not — the clay corresponds with the potter though the clay has no mind — but critics argue this stretches the concept to the point of metaphor. Second, whether the concept can travel from material practice to symbolic practice: can a writer correspond with language the way a potter corresponds with clay? Ingold is ambivalent; the present volume treats this as the central question his framework poses to the AI moment.
Appears in the Orange Pill Cycle
Correspondence as Luxury and Necessity — Arbitrator ^ Opus
The tension between Ingold's phenomenological account and the materialist critique depends entirely on which question we're asking. If we're asking what constitutes genuine skill development — how expertise actually forms, how knowledge becomes embodied — then Ingold's framework is nearly 100% correct. The potter really does develop knowledge through correspondence with clay, and this knowledge really does dissolve when the relationship ends. The AI prompt-execute cycle really is categorically different from this mutual responsiveness. On this descriptive question about the nature of skilled practice, the contrarian view adds context but doesn't challenge the core insight.
But if we're asking about the distribution and accessibility of correspondence — who gets to correspond, under what conditions, for how long — then the materialist reading dominates at about 80%. Most human work has long been structured to prevent correspondence, not enable it. The factory worker, the call center employee, the data entry clerk — none correspond with their materials in Ingold's sense. They follow procedures designed to eliminate the need for correspondence. Here the contrarian is right: AI doesn't introduce non-correspondence but extends it into previously protected domains.
The synthetic frame that holds both views recognizes correspondence as simultaneously a human necessity and a structural luxury. It is a necessity because without it, certain forms of knowledge and practice cannot exist — the knowledge that lives in relationships rather than representations, the expertise that emerges from sustained engagement rather than instruction. But it is a luxury because the conditions that enable correspondence — time, materials, freedom from immediate economic pressure — are unevenly distributed and increasingly scarce. The AI moment makes this contradiction acute: we have tools that can simulate the products of correspondence without the process, raising the question of whether we can afford to preserve the process when its products can be approximated more efficiently.
Further reading
- Tim Ingold, Correspondences (Polity, 2021).
- Tim Ingold, The Perception of the Environment (Routledge, 2000).
- Tim Ingold, Making: Anthropology, Archaeology, Art and Architecture (Routledge, 2013).
- Maurice Merleau-Ponty, Phenomenology of Perception (1945).
- James J. Gibson, The Ecological Approach to Visual Perception (1979).