nucleation

THEORY-032: The Yoneda lemma determines an object only up to isomorphism, from its whole hom-functor, among objects the category already has; as a formal version of "an object is determined by its relations" it presupposes the relata

4 min read

Properties6
StatusActive
Tagsphilosophy-of-mathematics, mathematics, individuation, metaphysics
SourceLIT-221 — Yoneda lemma, LIT-222 — No categorial support for radical ontic structural realism, LIT-154 — Criteria of Identity and Structuralist Ontology
SupportsCLAIM-132 — Characters classify finite-dimensional representations of a finite or compact group up to isomorphism only over a field of characteristic zero; they forget the isomorphisms, fix neither the group nor a permutation action, and classify kinds, not tokens or positions, CLAIM-135 — Recovering latent structure from relational data is three achievements, reconstruction, identification up to stated ambiguities, and constitution of a communicative kind, and identifiability theory reaches the second at most, CLAIM-136 — Communicative acts and their participants exist as relata but have no identity beyond what their relations differentiate: non-eliminative structural realism without primitive identity, CLAIM-137 — A communicative kind is identified by an invariant that does not depend on any one realization, as a group character identifies a representation up to isomorphism, once one symmetry group is identified across the realizations by its action on the situation and not by labels of a medium, so that invariant has classificatory priority over any particular realization, CLAIM-142 — Two communicative realizations can agree on every probe in a restricted family without being structurally identical, so observational equivalence is relative to the probes, and only independently justified probes evaluated on held-out cases are evidence that structure was preserved, CLAIM-150 — An ontology of intrinsically individuated objects with relational properties can reproduce the categorical and operational identifications, and the predictions, of a relational one, so identification underdetermines constitution, CLAIM-264 — Two objects can agree on every probe in a restricted family without being isomorphic, so equivalence under a probe family is coarser than identity unless the family is rich enough
Cited as “corroborated by” byCLAIM-136 — Communicative acts and their participants exist as relata but have no identity beyond what their relations differentiate: non-eliminative structural realism without primitive identity
Cited as “grounds” byARG-023 — The Yoneda lemma shows that relations constitute what a communicative act is

Graph View

Source

nLab, Yoneda lemma, revision of 2026-08-17, LIT-221, read in NOTE-195: Proposition and proof, Corollaries I–III, “Necessity of naturality”, “In semicategories”. Supporting: Lam & Wüthrich, LIT-222 §§4–5 (NOTE-194); Leitgeb & Ladyman, LIT-154 (NOTE-133), for the matching limit of structural identity criteria.

What was actually shown

The mathematics (proved, and checked in NOTE-195). For a locally small category C, a presheaf X and an object c, natural transformations Hom(−,c) ⇒ X correspond bijectively to X(c), by η ↦ η_c(id_c). Taking X to be a representable gives Corollary I: the Yoneda embedding y : c ↦ Hom(−,c) is fully faithful. Because fully faithful functors reflect isomorphisms, Corollary II follows: y(c) ≅ y(d) exactly when c ≅ d.

Three conditions the page’s own content exposes.

  1. The relational profile is the hom-functor with its action by composition, not a list of hom-sets. The page gives two non-isomorphic objects A and B with Hom(A,−)(Z) and Hom(B,−)(Z) in bijection for every Z (hom-sets ℤ≥0 and ℤ≥1 under addition). Only natural isomorphism of hom-functors fixes the object. Bare counts suffice only in special classes: finite relational structures (Lovász 1967), Pultr’s categories.
  2. Identities are needed. The proof evaluates at id_c, and the lemma fails for semicategories in general.
  3. The objects are presupposed. y(c) is indexed by every object of C, and the conclusion is an isomorphism between two of them.

Up to isomorphism, and no further. Two distinct but isomorphic objects, such as two singletons in Set, have isomorphic profiles. The lemma cannot tell them apart. Leitgeb & Ladyman meet the same limit for graphs: the two nodes of the edgeless two-node graph share every relational property, yet mathematicians count them as two (LIT-154). The categorical response is to deny that the distinction matters (identity = isomorphism, as in univalent foundations). It does not discern them.

Its unnamed use in the OSR literature. Lam & Wüthrich argue that generalized elements X → A exist in every category, and that morphisms “only distinguish isomorphism classes of structures” (LIT-222 §4 p. 10, §5 p. 11). These are consequences of the lemma, although the paper never names it. The record’s search found no core OSR paper that names it (NOTE-193, NOTE-194).

So, the record’s reading. The lemma is the precise form of “an object is determined by its relations” for objects that are already there. It is a relational criterion of identity up to isomorphism. It does not construct objects from relations that lack relata.

What this does not say

  • That the lemma supports radical (eliminative) ontic structural realism. On this reading it supports at most objects as positions in a structure. The argument about ROSR is a separate claim, filed beside this one.
  • That any author in the OSR debate reads the lemma this way. Bain (LIT-223) and Lam & Wüthrich do not mention it. The two works found that do name it were not filed: Adlam 2025 misreads its own source, and Hendren et al. 2026 assert rather than argue (NOTE-194).
  • Anything about enriched, bicategorical or (∞,1)-versions, whose statements differ. Nor anything about Riehl & Shulman’s directed dependent Yoneda (LIT-032).
  • That the nLab page is flawless. Its finite counterexample breaks the identity law as written, and its classical statement omits naturality in c and X (NOTE-195).

Connections

  • LIT-223 and LIT-222: the category-theoretic ROSR debate that this settles the formal side of.
  • LIT-196 (Chakravartty): inside a category the lemma is a partial counterexample to “relations cannot fix identity without intrinsic features”. It fixes identity only up to isomorphism, and the objects are presupposed.
  • THEORY-004 and THEORY-017 have the same shape in other settings: something is fixed only up to a symmetry group (orthogonal, unitary) by the data taken to be intrinsic. This is an analogy, not a shared mechanism.

Backlinks