Category: logic Type: thought experiment Associated thinker(s): W. V. O. Quine Original publication or approximate period: Quine, Word and Object (1960) Difficulty: Advanced Themes: radical translation, indeterminacy, reference Related experiments: Beetle in a Box; Hesperus and Phosphorus; Kripke’s Quus
In Quine’s imagined case of radical translation, a linguist encounters speakers of an unfamiliar language without an interpreter. A rabbit scurries by; a native speaker says “Gavagai.” The linguist tentatively records “Rabbit” (or “Lo, a rabbit”) as a translation, to test against further cases. This is a constructed example, not a report of a documented field encounter. (Quine, Word and Object, ch. 2, §7, pp. 29–31.)
Quine’s immediate point is about what this evidence establishes. Matching assent to “Gavagai” and “Rabbit” across relevant stimulus situations does not by itself establish that the native term gavagai and English rabbit apply to the same things. The native term might instead be analyzed as applying to brief rabbit stages or undetached rabbit parts; Quine also considers other analyses, including a rabbit-fusion or “rabbithood.” These alternatives concern the term’s reference, not whether the whole occasion sentence “Gavagai” can be tentatively translated as “Rabbit.” Establishing a term-to-term translation also depends on how the linguist analyzes such things as identity and plurality in the native language. (Quine, Word and Object, ch. 2, §12, pp. 51–53.)
Does the observable evidence of speakers’ assent and dissent, on its own, determine what a term refers to?
Quine’s “Gavagai” case is an illustration of the inscrutability (or indeterminacy) of reference: stimulus equivalence between occasion sentences does not settle how their component terms refer. It should not be treated as Quine’s complete argument for the broader indeterminacy of translation. In Word and Object, he separately argues that rival systems of analytical hypotheses can fit all speech dispositions while assigning incompatible translations to some sentences (ch. 2, §§15–16, pp. 72–74). Quine later made the distinction explicit: he wrote that the Gavagai example “was at best an example only of the inscrutability of terms, not of the indeterminacy of translation of sentences” (“On the Reasons for Indeterminacy of Translation,” 1970, p. 182).
This is a claim about what the evidence fixes within Quine’s account, not a demonstration that ordinary communication or practical translation fails. Quine allows that communication and translation can work; his point is that more than one translation may fit the relevant evidence. (Hylton and Kemp, “Willard Van Orman Quine,” §6.2.)
Quine’s discussion distinguishes the reference of a term from the translation of an entire sentence or language. The familiar rabbit alternatives concern the first issue. His broader thesis concerns the possibility of incompatible translation manuals that accord with speech dispositions. These claims are related, but the Gavagai example alone does not establish the broader thesis.
The interpretation and consequences of Quine’s indeterminacy theses have been disputed. The Stanford Encyclopedia’s overview notes John Searle’s objection that indeterminacy threatens communication or makes the idea absurd, and Dagfinn Føllesdal’s reply; it presents these as a debate, not a consensus. Quine’s own account of successful communication is compatible with working translation. This short entry does not attempt a general survey of externalist, causal, or pragmatic theories of reference.
Quine is not claiming that translation is impossible or that people never understand one another. Nor is “undetached rabbit part” an empirical claim about what an actual community’s speakers believe. It is an alternative term-analysis used to show that assent to the occasion sentence alone does not settle the term’s reference.
The point is not that every conceivable follow-up question or observation is useless. Quine discusses how a linguist might ask about identity and other grammatical constructions; such inquiry can settle the reference relative to adopted analytical hypotheses. But those hypotheses themselves involve choices not uniquely fixed by the original stimulus evidence. (Quine, Word and Object, ch. 2, §12, pp. 52–53; “On the Reasons for Indeterminacy of Translation,” pp. 181–82.)
Questions about rule-following and radical interpretation are useful comparisons, but they are separate philosophical discussions rather than parts of Quine’s original Gavagai example.
The primary-text locations above were checked in an online reproduction of Word and Object, Chapter 2, and relevant passages from Quine’s 1970 article. The book reproduction is hosted on a university philosophy department website; the complete book’s title page and all chapters were not consulted. Relevant passages in the 1970 article were read in a publicly accessible scan hosted on a course-reading site. The article’s DOI landing page at the Philosophy Documentation Center was used to confirm bibliographic details; it gates the full text and was not the source of the passage read. Exact references use the original journal’s printed pages. The Stanford Encyclopedia entries were read in the stated archived versions. These secondary sources support exposition and historical context; they do not settle whether Quine’s philosophical argument succeeds.
The scenario and questions for reflection are editorial teaching framing. They should not be mistaken for an actual ethnographic report or for claims that Quine’s Gavagai example alone proves the full indeterminacy-of-translation thesis. This is an educational explanation, not a substitute for the primary texts or a claim of expert review.
Research check: 2026-11-01.