Skip to content

Triangle of Reference

An Ogden–Richards semiotic model in which a symbol reaches a referent through a user's thought or act of reference rather than an ordinary direct word–thing tie.

Version
v1 · 2026-10-31 · History
Domain-specific #
14592
Domain group
Humanities
Origin domain
Philosophy
Subdomains
Semiotics, Philosophy of Language → Philosophy
Aliases
Semiotic Triangle, Triangle of Meaning, Ogden Richards Triangle

Core Idea

Ogden and Richards' triangle of reference distinguishes a symbol, a user's thought or act of reference, and a referent.[1] In ordinary word use, the connection from the symbol to the referent runs through that reference; the original diagram calls the base relation imputed.[1] A speaker uses the symbol but is not one of the three corners, and the referent is distinct from the thought or concept.[1]

The model is a tool for asking how this use of a sign points to what is meant. It does not claim that every symbol has exactly one thought and one object. The authors also note exceptional resemblance-linked signs, including images and gestures, so the ordinary indirect word–thing relation is not an unrestricted rule for every sign.[1]

How would you explain it like I'm…

Word, Thought, Thing

When you say the word 'dog,' the word doesn't touch the real dog. Your thinking connects them: you think of the dog, and the word points there through your thought. It's like a triangle with the word, your thought, and the dog at the corners.

The Meaning Triangle

The Triangle of Reference is an idea from Ogden and Richards about how words connect to things. It has three corners: the symbol (like a word), the thought in someone's mind, and the referent (the actual thing being talked about). The word connects to the thing through the thought, not directly, so the bottom side of the triangle is drawn as only an assumed link. The person using the word takes part, but isn't a fourth corner. Some signs, like pictures or sounds that imitate things, do look or sound like what they stand for, and Ogden and Richards noted those cases.

Symbol–Thought–Referent Model

The Triangle of Reference is Ogden and Richards' model of meaning. A symbol (such as a word) relates to a referent (the thing talked about) through a thought or act of reference. In the original diagram, the line between symbol and referent is shown as 'imputed' rather than a direct bond, because the connection runs through the act of referring. The person interpreting takes part in using the symbol, but they aren't a separate fourth corner or the referent. The model asks what understanding makes a symbol point to what someone means. It doesn't claim every sign has exactly one concept and one object, and Ogden and Richards noted exceptions like images, gestures, drawings and imitative sounds, which resemble what they stand for.

 

The Triangle of Reference is Ogden and Richards' model of how a used symbol relates to a referent through a thought or act of reference. The original diagram labels those three corners and depicts the everyday symbol-referent connection as imputed, an indirect relation mediated by the act of reference, rather than a direct tie between word and thing. The interpreter participates in the use of the symbol but is neither a fourth corner nor the referent. The model's question is which act or understanding makes a symbol pick out what someone is talking about. It does not promise that every sign has exactly one concept and one object, nor that sign relations never involve resemblance. Ogden and Richards explicitly qualify the usual indirect base with cases such as images, gestures, drawings and imitative sounds.

Scope of Application

The original illustration concerns an ordinary word used to refer to animals.[1] A later research application uses the triangle to analyze biomedical terms and codes in the Unified Medical Language System (UMLS).[2] The model belongs to sign and language analysis; another field needs to show the actual symbol, mediating reference, and represented target before claiming an instance.

Clarity

A word does not identify its referent merely by existing on a page. In Ogden and Richards' dog example, someone uses the word to refer to animals in the street.[1] Ask separately what the word is, what reference the user makes, and what the reference picks out. This keeps “what someone has in mind” from being confused with the animal being discussed.

Manages Complexity

The three corners prevent the loose word Meaning from covering a sign form, a user's understanding, and a target all at once. Campbell and colleagues show why that separation matters in UMLS: several source expressions can share CUI C0004057 while source-specific conceptual distinctions remain.[2] A shared identifier alone does not settle every interpretive difference.[2]

Abstract Reasoning

To use the triangle, name the symbol first. Then ask what thought or act of reference mediates its use, and finally what counts as the referent in that setting.[1] If the middle step is erased, the result is a simple label–object pair, which cannot answer the model's characteristic question. If the referent is replaced by the thought itself, the distinct third corner has also disappeared.

Knowledge Transfer

The dog example and the UMLS aspirin analysis share the same diagnostic separation even though one concerns ordinary speech and the other a designed terminology.[1][2] In the UMLS paper, source terms or codes fill the sign role, source intensions and interpretation qualify the middle role, and CUI C0004057 represents an extensional concept or class of things in a possible-world model.[2] The CUI is not a single physical tablet or a person's private thought.[2]

This is a qualified adaptation, not an assertion that computers literally think. It shows how keeping word, conception, and represented class distinct can reveal problems in terminology integration.

Examples

Ordinary word use. The authors' word dog is a symbol; someone using it makes a reference; animals in the street can be referents.[1] Mapped back: word → thought or reference in use → animal, with the ordinary word–animal base understood through that use.[1] This is an illustrative argument, not an experiment on speakers.

Biomedical terminology. Campbell and colleagues analyze aspirin expressions and CUI C0004057 in UMLS.[2] Mapped back: source words or codes → source intensions and interpretive context → a represented extensional class in the system's possible world.[2] The class role is not one physical aspirin item, and different source conceptions should not be silently treated as one private thought.[2]

Relationships to Other Abstractions

Local relationship map for Triangle of ReferenceParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Triangle of ReferenceDOMAINDomain-specific abstraction: Sign — presupposesSignDOMAIN

Current abstraction Triangle of Reference Domain-specific

Parents (1) — outgoing typed relations

  • Triangle of Reference presupposes Sign Domain-specific

    The triangle analyzes a sign's reference relation; without a sign standing for or indicating something to an interpreter, the model has no subject.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Triangle of Reference sits in a sparse region of the domain-specific corpus (67th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Language, Mind & Meaning-Making (73 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-31

Not to Be Confused With

  • Signifier–Signified Duality: a two-part account of a sign's form and concept that brackets the external referent; the Ogden–Richards model requires a referent corner.

  • Sense and Reference: a Fregean distinction involving an objective mode of presentation, with different commitments from the interpreter-involving triangle.

  • Denotation: a formal assignment of semantic value, which does not by itself establish the triangle's mediating thought or act of reference.

  • Direct naming: a word–object lookup may be useful, but it does not establish the original model's indirect ordinary base.[1]

References

[1] C. K. Ogden and I. A. Richards, The Meaning of Meaning (1923), printed pp. 10–12, especially the diagram on p. 11 and the dog discussion and resemblance qualification on p. 12. Original scan hosted by MIT: https://courses.media.mit.edu/2004spring/mas966/Ogden%20Richards%201923.pdf . The cited printed pages were checked through indexed primary-page text; local facsimile extraction was unavailable. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k

[2] Keith E. Campbell, Diane E. Oliver, Kent A. Spackman, and Edward H. Shortliffe, “Representing Thoughts, Words, and Things in the UMLS,” Journal of the American Medical Informatics Association 5, no. 5 (1998): 421–431, https://doi.org/10.1136/jamia.1998.0050421 . Full text and figures: https://pmc.ncbi.nlm.nih.gov/articles/PMC61323/ . Relevant locators: Figure 1; “The UMLS as a Possible World”; “Words,” Table 1; “Thoughts,” Figures 3–4; and “Things,” Figure 5. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i