Correspondence is Ingold's term for the process by which maker and material grow together — not through information exchange but through mutual responsiveness. The potter does not impose form on clay but corresponds with it, responding to resistance with pressure adjustments, following tendencies while guiding them, arriving at an outcome neither party predetermined. The correspondence is asymmetric (the potter has intentions, the clay does not) but genuine — the clay's behavior shapes the potter's decisions as decisively as hands shape clay. Both are transformed: clay becomes pot, potter becomes incrementally more skilled. This is the fundamental structure of skilled making, and it requires material resistance — the medium through which the correspondence becomes generative.
The concept draws on traditions of epistolary correspondence — the letters scholars exchanged over months and years, growing together through the practice of responding to each other's thought. The correspondence shaped both writers. It was not information transfer but mutual transformation. Ingold extends this structure to making: the maker's relationship to material is a correspondence in which both parties answer to each other's presence. The birch talks back to the carver through grain direction, knot deflection, density variation. The carver responds by adjusting technique. The form that emerges is the product of this dialogue, conducted through hands, in which material is a genuine interlocutor. The wood contributes information — through resistance, deflection, holding — that shapes decisions and produces outcomes neither intention nor properties alone could generate.
Material correspondence is characterized by three features: mutual transformation, material resistance, and temporal growth. The maker is changed by the encounter — skill develops, perception is educated, the body learns. The material is changed — it takes form, becomes something it was not. The transformation is mutual. Second, the material resists — it has its own logic, tendencies, insistence on certain behaviors. Resistance is not obstacle but medium. Third, the correspondence unfolds continuously over time — the pot grows moment by moment on the wheel, each moment dependent on the previous, the maker's hands always on the material, feedback continuous. This continuous flow distinguishes correspondence from transactional exchange (prompt-response-prompt-response) which introduces gaps where thinking happens outside the correspondence.
When Edo Segal describes collaboration with Claude as conversation, the metaphor suggests correspondence — iterative refinement, mutual responsiveness, insights emerging from the space between. Ingold's framework asks: what kind of correspondence is this? The human is transformed (Segal testifies to genuine learning). The machine 'resists' (Claude's interpretations differ from Segal's intentions, forcing reconsideration). But the resistance is interpretive, not material — Claude pushes back with alternative readings drawn from training data, not with the physics of matter asserting independent logic. And the exchange is sequential, not continuous — prompt, gap, response, gap, evaluation. The gaps are where human thinking occurs, but they mean the correspondence is transactional rather than flowing. The knowledge produced is representational (evaluating specifications) rather than enacted (depositing traces through continuous material engagement).
Ingold developed the concept across his late-career works, with fullest articulation in Correspondences (2021). The term recovers an older sense of correspondence as mutual answering, distinct from modern communication-as-information-transfer. Philosophical roots include Martin Buber's I-Thou relation (genuine meeting rather than instrumental use) and Merleau-Ponty's chiasm (the reversible intertwining of perceiver and perceived). Ingold grounds these abstractions in ethnographic observation: the Sámi herder's relationship to reindeer, the Cree hunter's to caribou, the Scottish builder's to stone. In each case, the relationship is not one of dominance but of mutual responsiveness — the human learns from the animal or material as much as shaping it.
Correspondence is mutual transformation. Both parties to the encounter are genuinely altered — the maker's skill develops, the material takes form, neither is unchanged by the process.
Resistance makes correspondence generative. The material pushes back with its own logic and tendencies; correspondence is productive because the maker must adjust to conditions she did not create and cannot fully control.
Continuity distinguishes correspondence from transaction. The potter's hands never leave the clay — feedback is continuous, each moment flows into the next, the correspondence grows without interruption.
AI correspondence is sequential and interpretive. Human-AI collaboration produces discrete exchanges with gaps between; the machine's 'resistance' draws from linguistic patterns, not from material physics — a different form of correspondence with different epistemic consequences.
Defenders of AI collaboration argue it constitutes genuine correspondence — that the human grows through the encounter, that the machine's responses are often unexpected (a form of resistance), and that insights emerge from the space between that belong to neither party alone. Ingold's framework would acknowledge these features while insisting on the categorical difference: correspondence with a linguistic system produces representational knowledge, correspondence with material produces enacted knowledge, and the two are not equivalent. The debate becomes empirical when asking whether a generation of practitioners trained primarily through AI correspondence will develop the judgment that previous generations built through material correspondence — a question answerable only through longitudinal observation.