A word enters a form, and something changes. Not the world — the form — but from that point on the word behaves differently: it can be counted, queried, denied, appealed. The history of classification is also a history of intervention, because categories do not merely organise an already-constituted world; they help decide what can be perceived, compared, administered and eventually transformed. Aristotle's categories, Kant's conditions of judgement, Wittgenstein's language-games and Foucault's analysis of discourse return again and again to the relation between naming and reality, yet what changes across this long conversation is the status each thinker assigns to classification itself. For Aristotle, the categories supply a grammar for saying what a thing is and how predicates properly attach to it: substance, quantity, quality, relation, place, time, position, state, action and affection are not an accumulation of words but distinct ways in which something can be said to exist. What endures from this gesture is not the taxonomy but the deeper claim that intelligibility depends on distinction — to think is already to divide, to relate, to assign a predicate — and that a category earns its analytical traction as much by what it excludes as by what it gathers in. Transdisciplinary research faces the same stake today, where terms such as network, system, ecology, platform, infrastructure or resilience can expand until they stop discriminating anything at all. The danger, as later commentators on Aristotle's legacy have often observed, is never conceptual mobility as such but conceptual inflation: a travelling concept needs boundaries firm enough to preserve whatever it was built to detect.
Kant radicalises the problem by making categories constitutive of intelligible experience rather than classifications discovered among things already given. In the first Critique, the categories are conditions through which the manifold of intuition becomes available to judgement in the first place, which displaces the question from what categories exist to what a given categorial apparatus allows us to recognise. Classification stops being innocent description and starts establishing a horizon of possible objects. An environmental dataset does not collect a pre-existing reality; it decides in advance what will count as an observation — temperature, particulate matter, canopy coverage, perceived safety, a biodiversity index, land value, household income — and each variable opens one dimension of the world while quietly suppressing the others. The epistemological problem, then, concerns not merely the accuracy achieved within a category but the architecture of the categories themselves, an architecture always partly arbitrary and always fully consequential.
Wittgenstein supplies a necessary correction to any theory that treats categories as self-contained objects floating free of practice. The later Philosophical Investigations relocates meaning in use: words operate within language-games whose rules cannot be separated from situations, practices and forms of life, and his notion of family resemblance matters above all for fields that cannot be organised through necessary and sufficient conditions — which describes a great many categories in contemporary spatial, cultural and environmental research. Public space, community, nature-based intervention, participation or artistic research rarely name objects sharing one essential property; they hold together through overlapping similarities, institutional histories and habits of use. One should study how a term functions before asking what definition it deserves. Use also generates drift: a concept can migrate between disciplines while keeping the same lexical surface even as its underlying mechanism changes — resilience in ecology, in psychology, in engineering and in urban policy is the obvious case — and while such translation can produce a genuinely productive mutation, it can just as easily produce a semantic smoothing in which incompatible meanings become interchangeable. Enough definition, differentiation, recurrence and declared limit can give a concept the resistance it needs to survive circulation without collapsing into generic language; call this SemanticHardening. The aim is never to immobilise meaning but to keep enough internal structure that a concept remains inspectable while it travels.
Foucault carries the argument beyond individual concepts toward the historical conditions that organise an entire field of what can be said. In The Archaeology of Knowledge, discourse is not reducible to vocabulary or to ideology, for statements belong to formations governed by rules concerning objects, concepts, enunciative positions and strategies. Certain things become describable only once a field has developed the institutional, technical and conceptual machinery required to recognise them; others remain hard to articulate because the available grammar offers them no purchase as objects of knowledge. A problem, on this reading, is never a mere reaction to external conditions — it must first become legible within a regime capable of naming its components, relating them to one another and recognising which evidence is allowed to count. Foucault also builds the bridge from language to infrastructure. Categories are reproduced through hospitals, schools, censuses, archives, professional disciplines and administrative routines, and contemporary digital systems intensify this same condition: databases, recommendation engines, platform taxonomies and automated decision systems depend on classification architectures whose categories may remain invisible to users while staying fully consequential for them. Content moderation translates culturally variable expression into operational classes whose thresholds execute at scale, so a classification does not merely describe content but decides whether it may circulate at all. Ideas, in turn, enter institutions not by being true or persuasive but by assuming a form the institution can accept — a research finding becomes a field in an application form, a keyword in a repository, a code in a planning regulation — a passage this genealogy fixes as FormatAdmission. And the same proposition, passed through another disciplinary medium, tends to come out bent in a new direction, a variation transdisciplinary work can treat as an object of analysis instead of dismissing it as noise.
Read together, these four positions yield more than a history of theories of language: four complementary operations, in which Aristotle establishes the necessity of distinction, Kant exposes the constitutive force of any categorial framework, Wittgenstein situates meaning firmly within use, and Foucault reveals the historical and institutional conditions under which objects and statements come to exist at all. Contemporary knowledge infrastructures add a fifth operation, as categories today are increasingly executable — encoded, queried, ranked, filtered, coupled directly to decisions. None of this amounts to linguistic determinism, nor to the claim that social reality simply reduces to discourse: bodies, ecological processes, buildings, technologies and economic distributions resist language continuously. Words become consequential only when they enter arrangements capable of acting through them. Everything else is commentary.
Aristotle (1984) Categories, in J. Barnes (ed.), The Complete Works of Aristotle. Princeton: Princeton University Press.
Foucault, M. (1972) The Archaeology of Knowledge. London: Tavistock.
Foucault, M. (1977) Discipline and Punish. New York: Pantheon.
Kant, I. (1998 [1781/1787]) Critique of Pure Reason. Cambridge: Cambridge University Press.
Wittgenstein, L. (2009 [1953]) Philosophical Investigations. Oxford: Wiley-Blackwell.
