Word and Object

Context

Word and Object, published in 1960 by MIT Press, is the magnum opus of Willard Van Orman Quine and one of the most important works of analytic philosophy in the twentieth century. Quine had been developing the ideas in the book for more than a decade — the seeds are visible in his 1951 paper "Two Dogmas of Empiricism," which rocked the philosophical establishment by attacking the analytic-synthetic distinction and reductionism, the two pillars on which logical positivism had been built. Word and Object is the full-scale elaboration of the philosophical system that "Two Dogmas" had sketched: a systematic account of meaning, reference, translation, and ontology grounded in a radically empiricist and behaviorist conception of language.

To understand the book's significance, one has to appreciate the intellectual landscape it transformed. By the mid-twentieth century, analytic philosophy was dominated by the legacy of logical positivism, a movement that had taken shape in Vienna in the 1920s and 1930s and had been transplanted to the English-speaking world by emigres like Rudolf Carnap. The positivists held that meaningful statements are either analytic (true by definition, like "all bachelors are unmarried") or synthetic (empirically verifiable, like "water boils at 100 degrees Celsius"). This distinction was foundational: it separated the truths of logic and mathematics, which hold in all possible worlds, from the truths of empirical science, which depend on observation. It also underwrote the positivist program of reducing meaningful discourse to statements about immediate sensory experience — a program that promised to clarify, and eventually dissolve, the traditional problems of philosophy.

Quine's "Two Dogmas" argued that both pillars were illusory. The analytic-synthetic distinction, he contended, could not be drawn in any non-circular way: every attempt to explain what makes a statement analytic (true by meaning alone) ultimately appeals to notions like "synonymy" or "necessity" that are no clearer than the notion they are supposed to explain. And reductionism — the idea that each meaningful statement can be translated into a statement about immediate experience — was equally untenable, because our statements about the world face the tribunal of experience not individually but as a corporate body. Word and Object takes these destructive arguments and builds something positive on their ruins: a comprehensive theory of how language relates to the world that dispenses entirely with the notions of analytic truth, synonymy, and meaning as traditionally understood.

The book was also shaped by Quine's deep engagement with behaviorism and with the empiricist tradition stretching from Hume through the logical positivists. Quine was convinced that any respectable account of meaning had to be grounded in publicly observable behavior — in the dispositions of speakers to assent to or dissent from sentences in response to sensory stimulation. This commitment to behaviorism was not merely methodological but metaphysical: Quine believed that meaning is not an inner mental entity that we somehow "grasp" but a pattern of dispositions that can, in principle, be observed and measured. This is the engine that drives the book's most radical and famous thesis: the indeterminacy of translation.

Summary

Indeterminacy of Translation

The indeterminacy of translation is the central thesis of Word and Object, and it is also the most controversial. Quine's claim is that there is no fact of the matter about whether two translation manuals — two systematic schemes for translating one language into another — are correct or incorrect. Given a language we don't speak, there will always be multiple translation manuals that are equally compatible with all the behavioral evidence (all the observed patterns of assent and dissent in response to stimuli), but that yield systematically different translations of the same utterances. And there is no further fact — no hidden meaning, no inner mental state, no Platonic form of meaning — that could decide between them.

This is a startling claim. It says that the question "What does this word mean?" has, in a deep sense, no determinate answer. We can translate, of course — we do it all the time — but our translations reflect a choice among equally valid manuals, not the discovery of a pre-existing fact about meaning. The thesis bears a family resemblance to the uncertainty principle in physics, and Quine sometimes encouraged the comparison, but the indeterminacy of translation is not a limitation of our knowledge. It is, Quine insists, a limitation of the facts. There is no fact of the matter to be known.

The argument proceeds through a thought experiment that Quine calls "radical translation" — the translation of a completely unknown language by a linguist who has no prior contact with the speakers and no bilingual interpreters to rely on. Quine asks us to imagine a field linguist encountering a previously unknown tribe. The linguist observes the tribe's verbal behavior — the circumstances under which they utter various sounds, the patterns of assent and dissent — and tries to construct a translation manual that maps the tribe's utterances onto English. The indeterminacy thesis says that the behavioral evidence will always underdetermine the translation manual: multiple manuals, yielding systematically different translations, will be equally consistent with all the evidence.

Radical Translation

The radical translation scenario is not merely illustrative; it is the methodological core of the book. Quine uses it to probe the nature of meaning by stripping away all the crutches that ordinarily conceal the difficulties. When we translate between languages we already know, we rely on shared history, shared culture, shared knowledge of the world — all of which make translation seem transparent. Radical translation removes these props and forces us to ask: what exactly is the evidence on which translation rests? What observable facts determine that a word in one language means a particular thing in another?

Quine's field linguist begins with the most basic case: a native utters the word "gavagai" while a rabbit runs past. The linguist tentatively translates "gavagai" as "rabbit." But how does the linguist confirm this? By observing the circumstances under which the native assents to or dissents from "gavagai" — pointing at rabbits, at other animals, at pictures, at rabbit parts. The problem is that all these tests are themselves mediated by language. To ask the native "Is this a gavagai?" the linguist must have already established a way of asking yes-or-no questions, which requires having already translated enough of the language to construct interrogatives. And even if the linguist can ask the question, the native's assent or dissent is a behavioral response that is compatible with multiple interpretations.

The deepest difficulty emerges when we consider what "gavagai" might mean. It might mean "rabbit" — an intact, enduring animal. But it might also mean "undetached rabbit part" — a temporally continuous slice of rabbit flesh. Or it might mean "rabbit stage" — a momentary temporal slice of a rabbit. Or it might mean "the rabbithood that is manifested here" — a universal instantiated by this particular rabbit. The behavioral evidence — the native's assent to "gavagai" in the presence of rabbits and dissent in their absence — cannot distinguish between these interpretations, because in every observable situation, all of them are present or absent together. To distinguish them, the linguist would need to ask analytical questions ("Is this the same gavagai as the one we saw yesterday?"), but those questions presuppose a translation of terms like "same" — and the translation of "same" is itself subject to indeterminacy.

This is the heart of the argument. The indeterminacy is not a practical limitation that better data could overcome. It is a structural feature of the relationship between behavior and meaning. No amount of behavioral evidence — no matter how exhaustive — can determine a unique translation, because the evidence is always compatible with multiple competing manuals. And since behavior is all there is (Quine's behaviorism rules out any appeal to inner mental states as the hidden determinants of meaning), the conclusion follows: translation is indeterminate.

Ontological Relativity

The indeterminacy of translation has a startling consequence for ontology — the study of what exists. Quine had long argued that our ontological commitments are determined by the variables in our best theories: "to be is to be the value of a bound variable." If our best physical theory quantifies over electrons, then we are committed to the existence of electrons. If it quantifies over numbers, we are committed to numbers. This criterion — ontological commitment via quantification — seemed to provide a clear, objective way of determining what a theory says exists.

But the indeterminacy of translation undermines this clarity. If the translation of a theory's terms is indeterminate, then the theory's ontological commitments are also indeterminate. The native who says "gavagai" might be committed to rabbits, or to undetached rabbit parts, or to rabbit stages — and there is no fact of the matter about which. In a later essay, "Ontological Relativity" (1968), Quine pushed this point to its conclusion: the question "What are we really talking about?" has no absolute answer. We can say what our terms refer to relative to a background language — English, say — but we cannot say what they refer to absolutely, because the background language is itself subject to the same indeterminacy. Ontology, Quine concluded, is relative: there is no fact of the matter about what our words refer to except relative to a translation manual, and the choice of manual is not dictated by any fact.

This is a deeply unsettling conclusion. It seems to undermine the idea that we can ask, in a meaningful way, what really exists. Quine accepted this consequence with characteristic equanimity. The relativity of ontology, he argued, doesn't make the question meaningless — it just means that the answer is always relative to a conceptual scheme, and that there is no privileged scheme from which to give an absolute answer. We can still do ontology; we just have to accept that our ontological commitments are relative to the language we use, and that the choice of language is a pragmatic one, not one dictated by the nature of reality.

Key Arguments

The Argument Against the Analytic-Synthetic Distinction

Word and Object extends the argument of "Two Dogmas" by showing in detail how the analytic-synthetic distinction collapses once we take seriously the empiricist and behaviorist constraints on meaning. The traditional distinction holds that some statements are true by virtue of meaning alone ("all bachelors are unmarried") while others are true by virtue of fact ("bachelors are less than 5% of the adult male population"). Quine's argument is that the notion of "meaning" on which this distinction depends cannot be made intelligible within a behaviorist framework. If meaning is determined by behavioral dispositions — by the circumstances under which speakers assent to and dissent from sentences — then there is no sharp line between statements that are true by meaning and statements that are true by fact. The boundaries of assent and dissent are fuzzy, shifting, and sensitive to the totality of the speaker's beliefs and experiences.

What replaces the analytic-synthetic distinction, on Quine's view, is a continuum. At one end are statements that are deeply entrenched in our web of belief — the truths of logic and mathematics, which we would revise almost anything before revising. At the other end are statements that are peripheral — empirical generalizations that are readily abandoned in the face of recalcitrant experience. But the difference is one of degree, not kind. No statement is immune to revision (we could, in principle, revise the law of non-contradiction if experience demanded it), and no statement is immune to retention (we could, in principle, retain any statement, no matter how empirically discredited, by making compensating adjustments elsewhere in our web of belief). This is Quine's confirmation holism: our statements face the tribunal of experience not individually but as a corporate body.

The Web of Belief and Confirmation Holism

The image of the "web of belief" is one of the most influential metaphors in twentieth-century philosophy. Quine invites us to think of the totality of our beliefs as a web, with the most central beliefs (logic, mathematics) at the center and the most peripheral beliefs (observations of particular events) at the edges. Experience impinges on the web at its edges — when we observe something unexpected, it creates a tension that must be resolved. But the resolution is underdetermined: we can resolve the tension by revising the peripheral belief (deciding that we misobserved), or by revising a more central belief (deciding that our theory is wrong), or by adjusting any number of beliefs in between. The choice is governed not by logic but by pragmatic considerations — which revision would do the least damage to the overall structure of our beliefs, which would be simplest, which would be most conservative.

This picture has profound consequences for epistemology. It means that there is no foundation of certainty on which our knowledge rests — no layer of "given" experiences or self-evident truths that anchors the entire structure. Every belief, no matter how central, is revisable in principle, and every belief, no matter how peripheral, can be retained if we are willing to make sufficient adjustments elsewhere. Knowledge is not a building with foundations; it is a web, and the strength of any particular strand depends on its connections to the rest. This is the view that Quine would later call "naturalized epistemology" — the project of understanding knowledge not as something justified by foundational truths but as a natural phenomenon, to be studied by the same empirical methods that we use to study everything else.

The Inscrutability of Reference

A close cousin of the indeterminacy of translation is what Quine calls the "inscrutability of reference." Even if we fix the translation of a sentence — even if we know that "gavagai" is used in the same situations as "rabbit" — the reference of the term remains inscrutable. "Gavagai" might refer to rabbits, or to rabbit parts, or to rabbit stages, and no behavioral test can tell us which. The reason is that reference is determined not by observable behavior alone but by the background of auxiliary assumptions that we bring to the interpretation of that behavior. To determine what "gavagai" refers to, we need to know what the native means by "same" — but the translation of "same" is itself indeterminate.

The inscrutability of reference has an unsettling implication: it seems that we cannot say, in any absolute sense, what our own words refer to. If the native's "gavagai" might refer to rabbits or to rabbit parts, and if there is no fact of the matter that decides between these, then the same is true of our own word "rabbit." We might be referring to rabbits, or to undetached rabbit parts, or to rabbit stages — and there is no fact of the matter that could settle the question. Quine accepted this consequence. The most we can say is that our terms refer to what they refer to relative to our own language — but this is trivial, and it doesn't give us the absolute notion of reference that traditional philosophy of language sought.

The Rejection of Meanings as Entities

Underlying all of these arguments is a more fundamental one: the rejection of meanings as entities. The traditional view, inherited from Aristotle and refined by Frege, held that words have meanings — abstract entities that they express and that speakers grasp when they understand the word. Quine's behaviorism leads him to reject this picture entirely. There are no meanings in the sense of abstract entities attached to words; there are only behavioral dispositions — patterns of assent and dissent in response to sensory stimulation. The notion of "meaning," if it is to be retained at all, must be reconstructed in terms of these dispositions, and the reconstruction turns out to be too coarse-grained to support the traditional philosophical uses of the notion.

This is why Quine is sometimes described as the philosopher who killed meaning. He didn't deny that words are meaningful — he denied that there are meanings in the sense of determinate, individuable entities that words express. The indeterminacy of translation is the proof: if there were meanings, translation would be determinate, because the meaning of a foreign word would be a fact that the right translation manual would capture. But translation is indeterminate, so there are no such facts, so there are no such entities. What there are, instead, are patterns of behavior — and these patterns underdetermine meaning in just the way that the indeterminacy thesis describes.

Legacy

Word and Object transformed virtually every area of analytic philosophy. Its impact has been so pervasive that it is difficult to imagine what the discipline would look like without it.

In the philosophy of language, the indeterminacy of translation shattered the assumption — shared by virtually all parties to the debates of the mid-twentieth century — that meaning is a determinate, individuable something that attaches to words and sentences. The assumption had been common ground between the positivists (who sought to analyze meaning in terms of verification conditions) and the ordinary language philosophers (who sought to analyze it in terms of use). Quine showed that both projects rested on an illusion: there is no determinate thing to be analyzed. The response to Quine was varied. Some philosophers, like Donald Davidson, accepted the indeterminacy and built theories of meaning that took it into account. Others, like Saul Kripke and Hilary Putnam, developed theories of reference (the causal theory, semantic externalism) that attempted to secure the determinacy of reference even if the determinacy of meaning could not be saved. The debate continues, but it is conducted on a terrain that Quine reshaped.

In epistemology, the web of belief and the rejection of the analytic-synthetic distinction undermined the foundationalist picture that had dominated the field since Descartes. If no belief is immune to revision and no belief is self-justifying, then the project of finding an indubitable foundation for knowledge is misconceived from the start. Quine's naturalized epistemology — the proposal that epistemology should be absorbed into empirical psychology, studying how humans actually form beliefs rather than how they ought to justify them — grew directly out of the arguments of Word and Object. This proposal was controversial (some philosophers accused Quine of confusing description with justification, of abandoning the normative dimension of epistemology), but it has been enormously influential. Contemporary naturalized epistemology, Bayesian approaches to belief revision, and the whole enterprise of studying cognition as a natural phenomenon all owe a debt to Quine's reorientation of the field.

In metaphysics, ontological relativity and the criterion of ontological commitment reshaped the debate about what exists. Quine's slogan — "to be is to be the value of a bound variable" — gave philosophers a precise tool for determining what a theory is committed to, and the relativity of ontology forced them to acknowledge that the answer is always relative to a conceptual scheme. This didn't eliminate metaphysics (as the logical positivists had hoped to do), but it changed its character. Post-Quinean metaphysics is more self-conscious about the relationship between language and ontology, more careful about the inferences from how we talk to what exists, and more willing to acknowledge that ontological questions may not have absolute answers.

In the philosophy of science, confirmation holism became a central theme. Quine's picture of the web of belief — in which no statement is tested in isolation but only as part of a larger theoretical structure — resonated with the work of Thomas Kuhn and other philosophers of science who were arguing that scientific theories are tested as wholes, not statement by statement. The holist picture also influenced the philosophy of science debates about the underdetermination of theory by evidence — the thesis that the empirical evidence always underdetermines the theory, because multiple theories can be compatible with the same data. This thesis, which has become a central concern of contemporary philosophy of science, was given its clearest and most influential formulation by Quine.

The enduring significance of Word and Object lies in its radicalism and its rigor. Quine took the empiricist tradition — the tradition that ties all knowledge to experience — more seriously than any philosopher before him, and he followed its implications wherever they led. They led to the dissolution of meaning, the relativity of ontology, the holism of confirmation, and the naturalization of epistemology. These are not comfortable conclusions, and many philosophers have resisted them. But the resistance has to engage with Quine's arguments, and the arguments are powerful. More than sixty years after its publication, Word and Object remains a work that no serious student of analytic philosophy, philosophy of language, or epistemology can afford to ignore — a text that reshaped the questions philosophers ask about knowledge, truth, and the understanding of reality.