Knowledge becomes a field when its internal relations acquire enough coherence for other people to enter, navigate, criticise and extend them. Warburg demonstrates through the Mnemosyne Atlas that relational arrangement can itself produce knowledge without requiring a single linear hierarchy; Kuhn explains how shared problems, exemplars and standards make cumulative inquiry possible; Bourdieu reveals that intellectual fields are simultaneously structures of positions, legitimacy and institutional power; Haraway demands accountability for the situated conditions from which knowledge is produced; Luhmann shows how fields reproduce themselves through recurring communication. Taken together, these perspectives distinguish a field from a collection. Accumulation alone produces an archive; a field requires recurring distinctions, genealogical relations, methods of entry, teaching structures and mechanisms that allow one intellectual operation to generate another. A corpus therefore precedes a canon because sufficient density is needed before patterns, absences and meaningful hierarchies can become visible. Vocabulary must then discriminate mechanisms rather than merely brand them; references must expose intellectual dependencies; pedagogy must force tacit structures into communicable form; publication must make propositions available for retrieval and criticism; archives must preserve their transformations through time. Teaching becomes especially decisive because a concept that cannot be explained, compared or operationalised may conceal incoherence behind apparent sophistication. A field is thus not established by declaration, scale or institutional recognition alone. It emerges when concepts, evidence, practices and infrastructures begin to reproduce themselves with differentiation across research, teaching and publication. Its architecture should provide orientation without closure: sufficiently structured to permit cumulative work, sufficiently open to admit new cases, disciplinary perspectives, corrections and conceptual mutations.