Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE ( © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 27 February 2020

Quine, Kripke, and Putnam

Abstract and Keywords

The traditional descriptivist distinction between the sense and reference of a proper name came under attack in the twentieth century. Quine’s attack on synonymy (sameness of sense) undermined the analytic–synthetic distinction, but his argument, in the form he gave it, depended on equating analyticity and necessity. Kripke and Putnam attack the idea that the sense of a name determines its reference and serves as a mode of presentation for its reference. They generalize their arguments to natural kind terms too. Kripke articulates a different way to understand necessity that reveals the possibility of contingent apriori and necessary aposteriori truths, thus breaking the connection Quine saw between analyticity and necessity, while leaving much of substance in Quine’s criticisms of the notion of synonymy intact. We place Putnam somewhere between Quine and Kripke on meaning, necessity, and analyticity. We close with an examination of the central use of philosophical intuition in these arguments and think Quine would find much to like in the rise of experimental philosophy.

Keywords: W. V. O. Quine, S. Kripke, H. Putnam, sense, reference, analyticity, apriority, aposteriority, necessity, necessary aposteriori, contingent apriori, synonymy, analytic–synthetic distinction, philosophical intuitions

19.1 Introduction

The emergence of semantic externalism and its dominance is undoubtedly one of the most significant philosophical developments of the past fifty years. Saul Kripke and Hilary Putnam’s rejection of Fregean and Russellian accounts of names introduced profound changes not just in philosophy of language, but also in our thinking about mind, metaphysics, epistemology, and even ethics. Their key ideas—rigid designators, aposteriori necessity, division of linguistic labour, the twin earth thought experiment, etc.—have become the common tools of how we think about our discipline and conduct our craft. This momentous development, however, was preceded by an earlier and equally significant challenge to philosophical orthodoxy: Quine’s attempt at naturalizing philosophy, which centrally involved his questioning of the very idea of analyticity and hence the possibility of having a coherent conception of meaning as traditionally construed. The broad outline of these tectonic changes is well known and discussions of their significance abound. What, at least until recently, have been neglected are the metaphilosophical consequences of the changes these philosophers had wrought and their impact on philosophical methodology. In this chapter we examine the connections between these dual shifts in philosophy of language and explore a concern about the methodological assumptions underlying the externalist takeover in philosophy of language.1

(p. 595) For much of the twentieth century, philosophy of language was dominated by what is commonly known as ‘descriptivism’—the approach to meaning adopted by the founders of analytic philosophy, Gottlob Frege and Bertrand Russell.2 The views that in any way associate reference with conceptual content or descriptions have also come to be known as internalist theories of meaning. In these theories, with a few exceptions such as logically proper names for Russell, the meaning of a word is a quite separate matter from its reference. Our utterances about the world are mediated by ideas that we have in our head, where the relevant sense of ‘idea’ is spelled out quite differently, varying for example, from the straightforwardly Lockean mental images of Wittgenstein (1921), to more sophisticated construals invoking representational intentional states, as in Searle (1983).

19.2 Quine’s Troubles with Sameness of Meaning

Frege famously distinguished between two aspects of the pre-theoretical notion of the meaning of an expression: its sense and reference. He argued that names do not refer directly to their bearers, rather their reference is given to them through a mode of presentation of the object he called the ‘sense’ of a name.3 Russell’s Theory of Descriptions, while it was developed as a critique of Frege’s idea of sense, made descriptions the medium linking names with their bearers. In Russell’s version of descriptivism, ordinary names are disguised definite descriptions and these descriptions determine what names refer to, e.g. the name ‘Scott’ may be shorthand for ‘the author of Waverley’ and the referent of ‘Scott’ would be the object, whatever it is, that wrote Waverley. According to Russell only logically proper names such as ‘this’ and ‘I’, used under appropriate circumstances, can refer directly. What the two theories have in common is the positing of semantic intermediaries between the referent of a name and that name.

Putnam and Kripke’s revolutionary move is rightly seen as a reaction to this model of meaning, but what is frequently neglected in the telling of this familiar story is the role of another revolution in twentieth-century philosophy, Quine’s attempt at dismantling the traditional notion of analyticity and hence meaning.4

(p. 596) Frege distinguished between the sense and reference of an expression by drawing our attention to the cognitive difference between identity statements containing two occurrences of the same name and those containing two different names of the same thing (Frege 1892). The cognitive value of ‘Hesperus is Hesperus’ differs from the cognitive value of ‘Hesperus is Phosphorus’. Both sentences are true, but while the first, at least on the standard account, appears to be a tautology, the second represents an astronomical discovery. In Frege’s account, the truth of identity statements depends on the reference of the names involved and the difference in cognitive value is due to the different sense attached to the names. Frege’s distinction is grounded in a difference, but arguably it takes more than that to introduce a new notion. To make sense of the distinction, we also have to know when the sense of one expression is the same as the sense of another, i.e. to say when two coextensive expressions mean the same thing or to give a criterion of synonymy. One way to see Quine’s ‘Two Dogmas of Empiricism’ is as a challenge to the possibility of providing this criterion of sameness of sense. If his challenge succeeds, those seeking an explanation of Frege’s distinction must return to square one.

The distinction between analytic and synthetic judgements was explicitly introduced by Kant in the Critique of Pure Reason, where Kant defines analytic statements as those where ‘the predicate B belongs to the subject A, as something which is (covertly) contained in this concept’. Analytic judgements, e.g. ‘bachelors are unmarried men’ are ‘explicative’, while synthetic ones, e.g. ‘Kant is a bachelor’, are ‘ampliative’. Kant also believed that analytic truths are logically necessary because their denial leads to contradiction ([1781/1787] 1965: 48). The distinction came to play an important role in Frege’s logicist programme, albeit with different definitions, especially of analyticity; it also became central to the logical positivists’ criterion of cognitive significance. In this sense, it is a formative idea of the analytic tradition of philosophy in the twentieth century. Frege characterizes analytic judgements as those that can be reduced to general logical truths and definitions only. Truths which are not of a general logical nature, but belong to the sphere of some special science, are synthetic (Frege 1884: § 3). Carnap, who was the primary target of Quine’s article, follows Frege’s lead and identifies analytic truth with truth by definition and logical truth (Carnap 1937, 1947) but also adds a more explicit semantic dimension by identifying analyticity with truth by semantical rule (Carnap 1947).

Quine’s general strategy is to show that none of the attempts to characterize the analytic/synthetic distinction, including Kant’s but most notably Carnap’s, manages to give a non-circular or non-question-begging account of the distinction. He concludes that the belief ‘That there is such a distinction to be drawn at all is an unempirical dogma of empiricists, a metaphysical article of faith’ (Quine 1953: 37).

Quine’s official target in ‘Two Dogmas of Empiricism’, in the first instance, was not so much the notion of sameness of meaning as that of truth in virtue of meaning and independently of fact, that is, the very idea of an analytic truth, which in its traditional rendering relies on the notion of meaning. He notes that the statements people take (p. 597) to be analytic fall into two general classes: logical truths and statements that become logical truths when we substitute synonyms for synonyms within them. He uses the statements

  1. 1. No unmarried man is married

  2. 2. No bachelor is married

as examples of the first and second kind respectively. Statement 2 becomes the logical truth 1 when ‘unmarried man’ is substituted for ‘bachelor’ in 2. Quine grants that we have an acceptable account of what makes a statement a logical truth: given a fixed set of logical vocabulary (‘if’, ‘then’, ‘and’, ‘not’, etc.), a sentence is a logical truth if it is true and remains true under all reinterpretations of the sentence’s non-logical vocabulary. So, the good standing of the notion of analyticity will be assured if what it is for expressions to be synonyms of each other can be made clear. Unfortunately, Quine thinks, the notion of synonymy is just as much in need of clarification as analyticity itself. The reason, it later appears, for finding synonymy unclear is that we lack an independent criterion for when two expressions are synonymous. That two words apply to exactly the same things is not sufficient for synonymy, but we can say what more is required only by employing the notion of analyticity.

Quine next considers whether statements like 2 can be reduced to logical truths by definitions. There are three cases to consider. First, there are dictionary definitions of the kind put together by lexicographers. These, he claims, rely on the pre-existing belief in the synonymy of the definiendum and definiens. They don’t provide any independent support for the notion of synonymy, as presumably the ‘definitions’ will have to be reconceived as something else if the notion of synonymy is unclear. The second case concerns explicative definitions of the kind often provided by philosophers. Here, the philosopher fixes on a number of favoured cases in which the definiendum is employed, finds an apparently synonymous definiens, and extrapolates on that basis to provide the definiendum with a precisified and sharpened usage in those further contexts. While this form of definition does not presuppose synonymy across the whole range of the word’s employment, it does rely on the belief in a pre-existing synonymy for the initially favoured cases. As with dictionary definition, then, it doesn’t offer independent support for synonymy. Lastly, Quine considers the introduction of novel notation for the purposes of abbreviation. In this case, the new vocabulary is synonymous with the old because it has been expressly created for the purpose of being synonymous with the old. Quine grants that here we do have a case of synonymy established by definition, but denies that the lessons carry over to the general case. Quine’s admission, with the last case, that there are genuine examples of synonymous expressions is inconsistent with his overall conclusion, since his concern is not how synonymies get created but what synonymy is (Soames 2003: 364n8; the tension is also noted by Boghossian 1997). If his arguments derail the notion, then it applies to nothing, so the status of these ‘definitions’ will also have to be reconceived.

(p. 598) The heart of Quine’s critique is a discussion of the following necessary and sufficient condition for synonymy:

S x is synonymous with y if and only if x and y may be substituted for each other wherever they occur salva veritate (without affecting the truth value of the sentences concerned).

This is immediately qualified to exclude substitution within words or phrases (to forbid, for example, the substitution of ‘puss’ for ‘cat’ in the word ‘cattle’) and within quotes (to forbid, for example, the substitution of ‘unmarried man’ for ‘bachelor’ in ‘the word “bachelor” starts with a “b”’). Quine also specifies that the synonymy in question need not be perfect, encompassing all aspects of a word’s meaning including poetic and psychological associations or those aspects which a Fregean might call tone (Färbung). No two expressions are perfectly synonymous in this sense. But this does raise the question of precisely what ‘grade’ of synonymy is at issue. Quine elects to specify the relevant sense indirectly.

The sort of synonymy needed [here is] merely such that any analytic statement could be turned into a logical truth by putting synonyms for synonyms. Turning the tables and assuming analyticity … we could explain cognitive synonymy of terms as follows (keeping to the familiar example): to say that ‘bachelor’ and ‘unmarried man’ are cognitively synonymous is to say no more nor less than that the statement ‘all and only bachelors are unmarried men’ is analytic.

(1953: 28–9)

In other words, the synonyms we are after are intersubstitutables such that the statement of their equivalence is an analytic truth.

Now the necessity of this condition is not disputed, but the question is whether it provides a strong enough condition for, as he calls it, cognitive synonymy. This is the heart of Quine’s argument. Quine claims that the strength of the condition depends on what sorts of contexts are covered by the ‘wherever they occur’. He says two things. If the ‘everywhere’ where the two terms are intersubstitutable encompasses only extensional contexts, then the condition is not strong enough for synonymy. For in these cases, any two terms with the same extension, that is, any two terms that correctly apply to exactly the same things, will be intersubstitutable without affecting the truth-value of the resultant sentences. And there are plenty of examples to show that such pairs are not synonyms. For example, the phrases ‘creature with a heart’ and ‘creature with kidneys’ have the same extension. So they are intersubstitutable in an extensional context like the sentence fragment ‘For all x, if x is a —— then x has blood’. If the sentence is true when the blank is filled in by one of the pair, it is also true when filled in by the other. But no one takes these phrases to be synonyms.

On the other hand, if the ‘everywhere’ where the two terms are intersubstitutable encompasses modal contexts, such as that formed by prefixing sentences with the modal operator ‘necessarily’, then Quine concedes that the condition is strong enough. That is, Quine believes only terms which intuitively really are synonyms could satisfy (p. 599) the criterion. How could we show that only intuitive synonyms pass the test? We could assume that two terms are intersubstitutable and then ask whether they are synonyms in the relevant sense, allowing their status as synonyms to be settled by whether the statement of their equivalence or identity is an analytic truth.

Indeed this is how Quine argues: The statement ‘Necessarily all and only bachelors are bachelors’ is true. If ‘bachelor’ and ‘unmarried man’ are everywhere intersubstitutable, then this sentence is also true: ‘Necessarily all and only bachelors are unmarried men.’ And for Quine this last statement says all of and no more than that the statement ‘All and only bachelors are unmarried men’ is analytically true. So, the mutually intersubstitutable terms can transform an analytic sentence of the second class into a logical truth, so they are synonyms in the relevant, cognitive, sense.

Before evaluating the argument, consider what it would show if it is successful. If successful, it shows that intersubstitutivity everywhere up to and including modal contexts is sufficient for synonymy (if two terms are so intersubstitutable, then they are synonyms). Quine has rested his case on the unclarity of synonymy, so once a necessary and sufficient condition for synonymy is provided, his case falls. The argument, however, does not withstand scrutiny for long. For Quine the problem lies with the adverb ‘necessarily’, which is ‘so construed as to yield a truth when and only when applied to an analytic statement’ (29/30), and consequently, ‘intelligible only in so far as the notion of analyticity is already understood in advance’ (31). In other words, Quine objects that this explanation of synonymy in terms of necessity and analyticity is circular.

Nowadays, however, circularity isn’t always seen as a fatal defect. Especially where the explanandum and explanans enjoy strong intuitive support, as the notions of analyticity and synonymy do, circular explanations can be seen as perspicuously displaying mutually supportive relationships among concepts. But the work of Kripke and Putnam, to follow, provides independent reason to reject the explanation of synonymy in terms of analyticity and necessary truth. Kripke and Putnam argue that the domain of analytic truths and the domain of necessary truths are not coextensive. If this is right, modal contexts aren’t themselves strong enough to test for cognitive synonymy.

The intersubstitutivity criterion poses particular problems in the contexts of belief ascription. As Quine himself has pointed out, coextensive terms such as ‘creature with a heart’ and ‘creature with a kidney’ are not intersubstitutable salva veritate in belief contexts, which might suggest that only genuine synonyms will be. It was the observed differences in cognitive significance and in behaviour of terms in belief contexts that led Frege to distinguish meaning as sense from meaning as reference in the first place. Belief contexts, then, are an especially appropriate venue for examining the intersubstitutability criterion and can afford a more direct specification of the kind of synonymy at issue. If two expressions are intersubstitutable in belief contexts they are equivalent in what matters for meaning as sense. They are therefore synonyms.

Two questions arise: are putative synonyms intersubstitutable in these contexts, and is anything else also intersubstitutable? These are rather delicate questions. But the following considerations seem to show that any reading of ‘belief’ context that would permit the intersubstitutivity of putative synonyms would also permit the intersubstitutivity of necessarily coextensive terms, like ‘water’ and ‘H2O’, that are not synonyms. In his (p. 600) subsequent discussions of the topic, Quine distinguished between referentially transparent and referentially opaque senses of belief (1956 and 1960: §§ 30, 31). A belief, or more generally, a sentential context, is referentially transparent if and only if co-referring expressions can be substituted for each other without affecting the truth-values of the resultant sentences (and opaque otherwise). The opaque sense of belief, that disallows substitutions, seems called for by the fact that someone who believes Superman can fly may deny that Clark Kent can fly (if they’re unaware that Clark Kent and Superman are one and the same man). If we take their sincere avowals and denials as circumscribing the extent of their beliefs, we must say they don’t believe that Clark Kent can fly. The transparent sense of belief, on the other hand, seems to be required by the fact that it is also true that the man they believe can fly is Clark Kent. So they do, in another sense, believe he can fly.

Given the above, it seems that the two senses of belief dictate different answers as to whether putative synonyms are intersubstitutable in belief contexts. On the transparent sense, if John believes that all bachelors are bachelors, then he ipso facto believes that all bachelors are unmarried men. Of course the recognition that certain terms pass the intersubstitutivity criterion given a transparent sense of belief still doesn’t establish that the terms are synonyms unless we can make the connection between transparent belief sentences and analytic truths (a move analogous to the move we rejected between necessity and analyticity). But we can say for sure that the move should be rejected because other terms besides synonyms are intersubstitutable in transparent, de re, belief contexts. For exactly the same reasons motivate saying that anyone who believes that (liquid) water is wet ipso facto believes that liquid H2O is wet. Intersubstitutivity in transparent belief contexts isn’t sufficient proof that the intersubstitutables are synonyms.

On the referentially opaque, de dicto, sense, we cannot automatically infer from ‘John believes that all bachelors are bachelors’ to ‘John believes that all bachelors are unmarried men’ because John might for some reason not believe this. He might think there are exceptions, or he may just be unsure. Given that the extent of his beliefs is settled by what he is disposed sincerely to avow, he simply may not have the latter belief. A defender of synonymy might be tempted to hold that if someone such as John fails to believe that all bachelors are unmarried men, it can only be because they don’t really understand the expressions in question. If John believes that all bachelors are bachelors and understands the meaning of ‘bachelor’ and of ‘unmarried man’ then the inference must go through. Recent work by Timothy Williamson undermines this suggestion (2005 and 2007).5 John might, consistent with understanding them, believe that both ‘bachelor’ and ‘man’ are vague terms with borderline cases. Indeed they are vague terms. No one would call a young boy a bachelor no matter how raucous he was. John could hold that in such cases ‘all bachelors are bachelors’ is true, but that ‘all bachelors are unmarried men’ is indeterminate for want of a guarantee that the vaguenesses of the terms exactly match. And the inference fails because truth is not preserved. If (p. 601) the Williamsonian counterexample is accepted, the intersubstitutability criterion once more fails to sort synonyms from non-synonyms. Even putting the apparent counterexample to the side, it would seem that the only reason to insist that someone who believes that all bachelors are bachelors must for that reason also believe that all bachelors are unmarried men, if they understand it, is a pre-existing belief in the synonymy of the words in question. For that reason the criterion, even if it provides an acceptable statement of what it is to be a synonym, doesn’t provide an independent test of whether there are any pairs of synonyms (and consequently any analytic truths). Making an exception in the case of alleged synonyms to taking a subject’s sincere avowals as indicative of their beliefs deprives the criterion of its original simple clarity. What someone believes is no longer a matter of what they would be inclined to say. ‘The thinks idiom,’ as Quine would say, ‘is heir to all the obscurities of the notion of synonymy … and more’ (Quine 1986: 9).

Quine’s dismantling the traditional view of analyticity had wide-ranging consequences, not least of which was the dethroning of logical positivism as the dominant philosophical position of the 1930s and 1940s. A further, and more damaging, consequence was his questioning of the very possibility of a theory of meaning. Our inability to give a non-circular or empirically backed meaning to the notion of meaning undermined the very idea of philosophy of language. Quine’s positive proposal was to offer an account of meaning, which would be in line with his behaviourist naturalism. He tells us,

we now have before us the makings of a crude concept of empirical meaning. For meaning, supposedly, is what a sentence shares with its translation; and translation at the present stage turns solely on correlations with non-verbal stimulation.

(Quine 1960: 32)

Quine bases his argument from the outset on behaviouristic premises and thinks of problems about meaning as problems of disposition to behaviour; as he puts it, ‘There is nothing in linguistic meaning, then, beyond what is to be gleaned from overt behaviour in observable circumstances’ (Quine 1987: 5). Meaning is to be explicated in terms of manuals of translation constructed by observing the stimulus-responses of speakers engaged in verbal behaviour. ‘We can take the behavior, the use, and let the meanings go’ (Quine 1979: 1). But even this pared down empirical approach was beset with problems, most notably the indeterminacy of translation and the inscrutability of reference. Quine’s thesis of indeterminacy, possibly his most controversial philosophical position, states that ‘manuals for translating one language into another can be set up in divergent ways, all compatible with the totality of dispositions, yet incompatible with one another’ (1960: 27). The doctrine, however, is not about mere translation, but applies to attempts at interpreting a ‘home’ language as well. Language is irredeemably indeterminate and the indeterminacy permeates even the level of singular putatively referential terms. Neither meaning nor reference can be pinned down. All we are left with is with language as

(p. 602) [A] social art which we all acquire on the evidence of other people’s overt behavior under publicly recognizable circumstances. Meanings, therefore, those very models of mental entities, end up as grist for the behaviorist’s mill. Dewey was explicit on this point: ‘Meaning […] is not a psychic existence, it is primarily a property of behavior’.

(1969: 26–7)

Moreover, Quine’s naturalism further solidified the dismissive attitude taken towards metaphysics by the logical positivist. Philosophy was no more than science conducted at higher levels of abstraction. As with the positivists, the autonomy of philosophy as a discipline was once again denied.

The next big revolution in philosophy of language challenged the two main assumptions of this Quinean picture, while simultaneously calling into question the prevailing descriptivist views of reference. The changes it brought about also revolutionized the thinking about the proper domain of philosophy.

19.3 Kripke and Putnam

The outline of Kripke and Putnam’s arguments for externalism about meaning is well known. Quine had argued for an essential gap between meaning and reference. According to him,

When the cleavage between meaning and reference is properly heeded, the problems of what is loosely called semantics become separated into two provinces so fundamentally distinct as not to deserve a joint appellation at all. They may be called the theory of meaning and the theory of reference.

(1953: 130)

Even earlier, in the ‘Two Dogmas of Empiricism’, he had relied on Frege’s discussion of sense and reference to argue that meaning is not to be identified with naming or reference. The Evening Star is the planet Venus, and so is the Morning Star. The two singular terms name the same thing. But the meanings must be treated as distinct, since the identity ‘Evening Star = Morning Star’ is a statement of fact established by astronomical observation. If ‘Evening Star’ and ‘Morning Star’ were alike in meaning, the identity ‘Evening Star = Morning Star’ would be analytic (Quine, 1953: 21). Kripke makes essentially the same argument the centrepiece of his new non-descriptivist position.

If ‘Aristotle’ meant the man who taught Alexander the Great, then saying ‘Aristotle was a teacher of Alexander the Great’ would be a mere tautology. But surely it isn’t; it expresses the fact that Aristotle taught Alexander the Great, something we could discover to be false. So, being the teacher of Alexander the Great cannot be part of [the sense of] the name.

(1980: 30)

(p. 603) Where Quine thought of meaning and reference as ‘so fundamentally distinct as not to deserve a joint appellation at all’, Kripke drew a dramatically different conclusion. He, and Putnam, sought to establish the contrary, Millian view that, at least in some crucial cases, meaning involves a direct relationship between names and their objects of reference and more generally, that we can give sense to the idea of meaning by looking at the relationship between language and the natural and social world.

Kripke offers three key arguments against the theory that names are disguised descriptions. First, many names are not, in general, associated with any uniquely identifying descriptions. For example, many people would be hard pressed to associate anything more descriptive than ‘a famous physicist’ with the name ‘Richard Feynman’. Second, it is perfectly possible to use a proper name in appropriate circumstances without being aware of the descriptions that others have associated with it. It does not seem like a case where we should want to say the name literally has a different meaning for speakers with different associations. Third and most significantly, proper names and natural kind terms act as rigid designators, expressions that refer to the same object in all possible worlds in which those objects exist, whereas descriptions can be satisfied by different objects in different possible worlds.6 For example, it seems intuitive to say ‘Aristotle’, as we use it, refers to Aristotle in the actual world and continues to refer to him when we consider him in various counterfactual situations, whereas the description ‘Plato’s greatest pupil’ needn’t always refer to him (such as when we imagine Plato had an even greater pupil).

Kripke’s arguments regarding the meaning and reference of names often rely on thought experiments or possible world scenarios as test cases for deciding between the two competing philosophical hypotheses of descriptivism and his preferred option of names acting as rigid designators,7 and the plausibility of his account of them is established, not just through testing their consistency and coherence, but ultimately by reliance on shared intuitions or our common sense. To take a favourite example, we could envisage possible worlds where Aristotle was not the teacher of Alexander, but there cannot be a possible world in which Aristotle is not Aristotle. Later we consider the role of intuition in these arguments more closely.

(p. 604) According to semantic externalism, objects, as well as persons, are named through acts of linguistic baptism, which also fix their reference. Names are introduced into a linguistic community for the purpose of referring to an individual and they continue to refer to that individual as long as their uses are linked to the original act of dubbing or ‘baptism’ via a social historical chain of continuous use. Each new user acquires the name from others, who in turn have acquired it from their predecessors, and so on, up to the first user who introduced the name to refer to a specific individual. The reason that the name ‘Aristotle’, to revert to our previous example, picks out or refers to the man Aristotle is that there exists a long chain of connections between the name and the person Aristotle. Kripke explains,

Someone, let’s say, a baby, is born; his parents call him by a certain name. They talk about him to their friends. Other people meet him. Through various sorts of talk the name is spread from link to link as if by a chain. A speaker who is on the far end of this chain, who has heard about, say Richard Feynman, in the market place or elsewhere, may be referring to Richard Feynman even though he can’t remember from whom he first heard of Feynman or from whom he ever heard of Feynman.

(1980: 90)

Kripke’s social-historical picture of the fixation of reference for names has come to be known as the causal account of reference.8 Kripke applies it both to proper names and natural kind terms (names for kinds or types of things that occur in nature, e.g. ‘water’, ‘tiger’, ‘lemon’, ‘gold’). In this approach, natural kind terms are also rigid designators and they pick out certain objects through causal, communicative ties between the term and the object to which it refers rather than through a cluster of descriptions that may be associated with the term. Semantic externalism is motivated by an underlying metaphysics whereby the essential properties of each natural kind object are seen as microstructural rather than phenomenal. For instance, gold is anything which has the atomic number 79 and not whatever looks like gold. As we’ll see, the view has important consequences not just for theories of meaning but also for our understanding of the modalities of necessity, the a priori, etc.

Shortly after Kripke’s first statement of semantic externalism, Hilary Putnam set out to refute what he called a ‘grotesquely mistaken’ view of language (1975b: 271), a mistake rooted in our propensity to ignore the social nature of language and to neglect the contribution of external reality to meaning. In this, Putnam’s externalism has a wider scope than Kripke’s, which had little to say about the social determinants of meaning, or so-called social externalism. Putnam, like Kripke, argued that the description theory does not capture what our words mean because for proper names and natural kind terms the (p. 605) reference of a name is determined by a causal interaction between objects in the world and the speakers using those names. For both these philosophers, even in cases where a proper name has been introduced to language by means of a description, the description will not give the meaning of the name. Traditional theories of meaning, chiefly descriptivism, have tended to make two incompatible assumptions about language. First, that to know the meaning of a term is to be in an appropriate psychological state and second, that meaning determines reference. The obvious consequence was that speakers in the same mental state uttering the same word token would share both the intension and the extension of that word. Putnam argues against this conclusion, by pointing out that since he is unable to differentiate between a beech and an elm, his conception of the two trees is the same, while the meaning and the reference of ‘elm’ and ‘beech’ are quite different in his idiolect. Putnam, or any other speaker in a similar predicament, does refer to an elm when using the term ‘elm’ even if he is not able to identify an elm. Successful reference is possible because of what Putnam calls ‘the division of linguistic labour’. Experts, horticulturalists in this instance, can distinguish an elm from a beech just as metallurgists can distinguish gold from fool’s gold, etc. Ordinary users of language defer to these experts in deciding the cases where ‘elm’ and ‘beech’ should be used. Shared meaning does not require that every member of a linguistic community should have the ability to identify accurately the objects of reference for a given name. What is needed is a stereotype associated with the object of reference and shared by the members of that linguistic community.

Naming and Necessity and ‘The Meaning of “Meaning”’ not only provided trenchant criticisms of the then-prevailing descriptivist views of meaning but also ran counter to Quine’s scepticism about the possibility of a theory of meaning. Quine, as we saw, in addition to questioning the legitimacy of the division between analytic and synthetic statements also doubted the related distinction between necessary and contingent truths. Kripke and Putnam break from the traditional conception of necessity, making space for both a theory of meaning and a re-engagement with metaphysics, even in the context of a naturalist conception of philosophy.9 What they, in effect, try to show is that there can indeed be such a thing as a theory of meaning, but a correct theory will have to take the formative role of the world into account. Furthermore, while their view of meaning does not give analyticity quite the role formulated and criticized by Quine, necessity and identity, recast in terms of possible world semantics, do become central.

19.4 Necessity and the Analytic/ Synthetic Distinction

Since Kant it has become common practice to assume that a priori and necessary truths are coextensive, that analytic truths are both a priori and necessary, and that contingent (p. 606) truths are known only a posteriori, while necessary truths are known a priori. Quine, like the logical positivists, particularly Carnap who was the prime target of his arguments, identifies analyticity with necessity and by denying the legitimacy of the analytic/synthetic distinction also undermines the necessary/contingent distinction. To take one example, the true premises, ‘necessarily 9 is odd’ and ‘9 = the number of planets’ gives us the outrageous conclusion, ‘necessarily the number of planets is odd’ (Quine 1943, 1947).10 Quine blames the problem on the ill-defined operators of non-truth-functional logic and particularly the modal operator ‘necessarily’ which does not permit the substitution of coextensive expressions within its scope, thus delegitimizing any attempt at quantifying into their scopes (Quine, 1953, 1960). But once the core ideas of analyticity and necessity are called into question, given the context of Quine’s strong naturalism, the very possibility of a priori knowledge also becomes problematic. The rethinking of these core philosophical ideas underpins Putnam and Kripke’s revolutionary approach to meaning. The point at stake is most explicitly made in Putnam’s ‘“Two Dogmas” Revisited’ (1976), where Putnam affirms the historic importance of Quine’s original paper while pointing out its main failure: the failure to distinguish between analyticity and the a priori. Quine, Putnam maintains, targets two different accounts of analyticity. On the first account, the Kantian account, analytic truth is a true statement derived from a tautology by putting synonyms for synonyms. The second account, however, has a more epistemic thrust. Analytic truth here is defined as a truth confirmed come what may. This account, according to Putnam, is closer to the traditional account of the a priori. Putnam further argues that while the first four sections of Quine’s ‘Two Dogmas’ focus on analyticity in the first sense discussed above, the last two concern the a priori. The accuracy of the above interpretation is not at stake here, rather, the point we are making is that the externalist revolution was not only an attack on descriptivism but also a rethinking of the then prevailing views of the triumvirate analyticity/necessity/apriority.

Putnam, more explicitly than Kripke,11 sees a role for a rehabilitated version of the analytic/synthetic distinction. He insists that there is little doubt that Quine is wrong in rejecting the distinction out of hand. ‘This is not a matter of philosophical argument’, he says, for there is ‘as gross a distinction between “All bachelors are unmarried” and “There is a book on this table” as between any two things in the world, or, at any rate, between any two linguistic expressions in the world’ (Putnam 1975a: 36). However, Putnam is also in agreement with Quine’s concern about the over-use of the ill-defined notion of analyticity and believes that ‘overworking the analytic–synthetic distinction has created serious distortions in the logical positivists’ writing about (p. 607) science’ (1975a: 33). The distinction, he argues, could have a role provided that it is not treated as a sharp dichotomy but a continuum. Statements, in his approach, range from the clearly analytic (‘bachelors are unmarried men’) to the clearly synthetic (‘this piece of chalk is white’). For specific purposes and in specific contexts we may find it useful or appropriate to distinguish between analytic and synthetic sentences, but such situational distinctions are not fixed for all times and contexts and sentences don’t fall neatly under one or other rubric.

In his earlier writing on the topic, Putnam delineates a small group of concepts; he calls them ‘one-criterion concepts’, which have a single criterion for application. For instance, the only criterion for the application of ‘bachelor’ is that of being an unmarried man and we cannot abandon this criterion without depriving ‘bachelor’ of its use, hence the analyticity of ‘all bachelors are unmarried men’. A majority of concepts, on the other hand, are ‘law cluster’ concepts where ‘any one law can be abandoned without destroying the identity of the … concept involved’ (Putnam 1975a: 52). It is wrong to think of such concepts as being analytic or sentences containing them as true by virtue of the meaning of their terms, because the link between the term picking out this type of concept and the cluster of laws associated with it is contingent, contextual, and malleable. The Quinean arguments against analyticity apply to this type of statement.

Putnam also agrees with Quine that in principle, all our claims to knowledge are open to revision. Even laws of logic, he has argued, could be revised in order to accommodate evidence from quantum mechanics; indeed there could be purely theoretical reasons for abandoning beliefs that were previously treated as true a priori or necessary (1975a: 48). Yet, his position is markedly different from Quine’s in that Putnam is prepared to allow for the possibility of necessary statements in the empirical domain, such as physics, that are nevertheless open to revision in the light of new information. This was the situation with Euclidean geometry, a favourite example, where ‘a statement that was necessary relative to a body of knowledge later came to be declared false in science’ (1962: 662).

In his later work, Putnam draws an even sharper distinction between necessity and the a priori, and thus makes room for the synthetic a priori truth. He divides necessary truths into three main sub-areas: the analytic, the logical and mathematical truths, and the ‘synthetic apriori’ (Putnam 1975a). Quine’s arguments against a particular view of meaning applies to analytic sentences, but it is possible for some feature of a natural kind to be part of the meaning of a term by being part of the stereotype associated with that term, without rendering the definition of the term ‘analytic’. For instance, being striped is part of the meaning of ‘tiger’ as it is one of the stereotypes associated with tigers. However, albino tigers will still be tigers and upon their discovery, being striped will cease to be a stereotype associated with the name, without the meaning of the name changing. Some features are seen as central to our thinking about a particular natural kind but centrality is not the same thing as analyticity. All this would be in agreement with Quine who also explains our ‘intuitions’ regarding analyticity in terms of the centrality of certain beliefs. But Putnam, more controversially, also argues that although ‘all tigers are animals’ is not analytic it is necessary because the natural kind term ‘tiger’ is a (p. 608) rigid designator—it refers to tigers in all the possible worlds where tigers exist. However, this necessary truth is discovered empirically and is not known a priori. Moreover, even though no single empirical example will refute ‘all tigers are animals’, it would be falsified if it turned out all tigers were robots sent by aliens.

The gap between Kripke and Quine on the issues of necessity and the a priori is even wider than the one separating Putnam from Quine. Quine’s worries about the legitimacy of the analytic/synthetic distinction extended to the related issues of the a priori and necessity, since for Quine, the necessity operator depended on a clear sense of analytic truth. Kripke, on the other hand, not only allows for the a priori necessities of logic but also for a posteriori necessities such as ‘water is H2O’ which is not merely verbal or conventional but straightforwardly empirical. As with Putnam, rejecting Quine’s approach to the a priori and necessity has consequences for theories of meaning. In the Frege–Russell theory, names are abbreviated descriptions. As we saw, Kripke’s key idea is that names are rigid designators and not non-rigid descriptions. Descriptions attach themselves to objects contingently, while rigid designators refer to the same object in every possible world. The description ‘teacher of Alexander’ is true of Aristotle, but it could have been false. Aristotle may have chosen a different career. There is no possible world in which Aristotle is not Aristotle, while there are possible worlds where Aristotle is not a teacher. What is more, necessities can be established empirically. As Frege noted, the identification of Hesperus with Phosphorus was an empirical discovery, but once the identity was established then we are in possession of an a posteriori necessary truth as identity statements involving rigid designators are necessarily true.12 The same point holds with other frequently used examples of a posteriori necessities such as water is (=) H2O and temperature is (=) mean kinetic molecular energy. Kripke also argues that there are some contingent truths that are known a priori. For instance, we have stipulated that a certain platinum/iridium rod, kept in Paris, is the standard measure for ‘one metre’. Once we accept this stipulation, then we know a priori that particular rod in Paris is a metre long. But the claim is not necessary, the definition of a metre—one/ten-millionth of the distance from the equator to the north pole along a meridian through Paris—and the rod that was originally picked could have been different. Thus, contrary to the assumption dominating philosophy since Kant, the a priori/a posteriori distinction does not track the necessary/contingent distinction. And most significantly, while Quine runs apriority and necessity together, Kripke lays bare an essential difference between them showing that some kinds of necessity are metaphysical, rather than logical, while the a priori is epistemic. Moreover, contrary to Quine, genuine a priori knowledge is not limited to necessary truths. Kripke explains:

(p. 609) Philosophers have talked … [about] various categories of truth, which are sometimes called ‘apriori’, ‘analytic’, ‘necessary’ … these terms are often used as if whether there are things answering to these concepts is an interesting question, but we might as well regard them all as meaning the same thing.

… First the notion of aprioricity is a concept of epistemology. I guess the traditional characterization from Kant goes something like: apriori truths are those which can be known independently of any experience…. The second concept that is in question is that of necessity…. what I am concerned with here is a notion, which is not a notion of epistemology, but of metaphysics … We ask whether something might have been true, or might have been false. Well, if something is false, it’s obviously not necessarily true. If it is true, might it have been otherwise? Is it possible that, in this respect, the world should have been different than the way that it is? … This in and of itself has nothing to do with anyone’s knowledge of anything. It’s certainly a philosophical thesis, and not a matter of obvious definitional equivalence, either that everything apriori is necessary or that everything necessary is apriori…. at any rate they are dealing with two different domains, two different areas, the epistemological and the metaphysical.

(Kripke 1980: 33–5)

He goes on to point out that that even if the a priori and the necessary are conceptually distinct, it does not follow that they are extensionally distinct. There could indeed be truths that are necessary but knowable only a posteriori as well as a priori truths that are not necessary. Almost paradoxically, Kripke’s arguments could help to undermine analyticity. Kripke stipulates that ‘an analytic statement is, in some sense, true by virtue of its meaning and true in all possible worlds by virtue of its meaning. Then something which is analytically true will be both necessary and apriori’ (Kripke 1980: 39). But as Tim Williamson has noted, after Kripke distinguished necessity from apriority, it’s not clear what work is left for the concept of analyticity to do (2007: 51ff.). But putting aside this particular caveat, it is widely accepted that Kripke’s novel treatment of necessity rescues meaning, at least when it comes to names and natural kind terms, from Quinean despair.

Kripke and Putnam’s role in shifting the debate on questions of language is well known. What is less frequently discussed, at least until recently, are the metaphilosophical changes that their approach to meaning and reference generated. Their work, Kripke’s in particular, not only led to a resurgence of interest in metaphysical questions but also gave new rigour to the view that philosophy could be an autonomous discipline, independent of the natural sciences.

19.5 The Role of Intuition

One significant and influential feature of both Kripke’s and Putnam’s approaches is their use of thought experiments in driving their arguments. Putnam’s Twin-Earth water and XYZ thought experiment, the aluminium–molybdenum science fiction story, and the (p. 610) less other-worldly elm–beech illustration, and Kripke’s Gödel/Schmidt and the biblical Jonah stories are by now classics in discussions of meaning and reference. The thought experiments attempt to show that descriptions associated with a name are not sufficient for establishing its reference, rather it is features of the environment that determine both the meaning and reference of proper names and natural kind terms.13 Kripke and Putnam used these thought experiments to demonstrate that our intuitive judgements about meaning and reference give us verdicts that are inconsistent with traditional descriptivism. Since then a growing number of philosophers of language, including most anti-externalists, have come to share Putnam and Kripke’s intuitions, finding them convincing enough to try to accommodate within a more sophisticated descriptivist framework (e.g. Evans 1973; Jackson 1998). Over the last thirty years it has become well entrenched that our semantic intuitions support the social-historical view; however, while logicians, mathematicians, linguists and philosophers rely heavily on intuitions in the justification of their theories, at least in philosophical debates, not much justification is provided for the authority assigned to such intuitions.

The prominence and respect philosophers ascribe to intuitions have ebbed and flowed, partly because of developments in sciences and cognate fields. The discovery of non-Euclidean geometries undermined what were once seen as strong spatial intuitions; the same is true of the theory of relativity and the revisionary views of time and motion it introduced. Furthermore, not all philosophers have advocated the use of intuitions as a methodological device. For example, John Stewart Mill in discussing theologians and other religious commentators says, ‘So they either dislike and disparage all philosophy, or addict themselves with intolerant zeal to those forms of it in which intuition takes the place of evidence, and internal feeling is made the test of objective truth’ (Mill 1874/2008: 72). Frege and Russell also mistrusted common intuitions in questions of mathematics and logic and Russell particularly disparaged any reliance on our stock of common-sense beliefs. Quine too is suspicious of over-reliance on them. On the role of intuitions in establishing the legitimacy of the analytic/synthetic distinction he says:

The intuitions are blameless in their way, but it would be a mistake to look to them for a sweeping epistemological dichotomy between analytic truths as by-products of language and synthetic truths as reports of the world.

(Quine 1960: 67)

The question of the role intuitions in philosophy is thorny, to say the least. At its most extreme, the appeal to intuitions in philosophy of language assumes that language users have an implicit theory of reference that becomes the sources of speaker’s intuitions on the connections between names and their objects (Mallon et al. 2009). Such appeals to intuitions about reference are frequently conceived by analogy to Chomsky’s project (p. 611) in linguistics. Chomsky, as is well known, maintains that all language users rely heavily upon ‘intuitions’ of grammaticality, or the innate linguistic abilities to make instant judgements about correct syntax in sentences such as ‘Who did you speak to Noam and?’ vs. ‘Who did you and Noam speak to?’ Ordinary speakers of language simply know what a well-formed linguistic expression is and any native or even non-native but competent speaker of English, including very young children who have had very little exposure to language, would be able to see (intuit) the difference between the two sentences and pick the grammatical one. To have intuitions about the correct use of language is constitutive of being a language user. Linguists, on the other hand, make use of these intuitive judgements as data for their theories, but also rely on their own intuitions to further these theories.

There may indeed be a historical connection between the use of intuitions in linguistics and philosophy of language. Hintikka (1999) points out that the use of the term ‘intuition’ in contemporary philosophy became much more common as philosophers became acquainted with Chomsky’s work in linguistics. The thought was that theories of reference could be constructed in analogy with the Chomskyan project in linguistics and philosophers of language would be able to rely on ordinary people’s intuitions about reference in their reconstruction of the implicit theories that were part of speakers’ cognitive endowment (Mallon et al. 2009: 339). Interestingly, Chomsky believes that there is no such thing as ‘philosophical intuitions’ about language and that linguistic intuitions and the so-called ‘philosophical intuitions’ are not comparable.

A good part of contemporary philosophy of language is concerned with analyzing alleged relations between expressions and things, often exploring intuitions about the technical notions ‘denote’, ‘refer’, ‘true of’, etc., said to hold between expressions and something else. But there can be no intuitions about these notions, just as there can be none about ‘angular velocity’ or ‘protein’. These are technical terms of philosophical discourse with a stipulated sense that has no counterpart in ordinary language.

(Chomsky 2000: 130)

Chomsky’s objections, however, have not had much impact on the use of intuitions as a methodological tool. Devitt and Hanley’s Blackwell Guide to the Philosophy of Language, confirms that:

The dominant method in semantics is to consult ‘intuitions’ about what an expression means, refers to, and so on, intuitions that are usually elicited in ‘thought experiments’…. Whichever account of intuitions one adopts, there is almost uniform consensus on the role of intuitions in philosophy.

(Devitt and Hanley 2006: 1–2)

One problem with the official ideology is a lack of consensus on what philosophical intuitions are and what role they play in philosophical argument. Are semantic intuitions expressions of our a priori knowledge of concepts, or simply unreflective first (p. 612) judgements with the same evidential status as the type of empirical intuitions scientists use in building and assessing their theories? Indeed at times, the official ideology seems more akin to a sleight of hand than a reliable methodology. As Tim Williamson puts it,

When contemporary analytic philosophers run out of arguments, they appeal to intuition. Intuitiveness is supposed to be a virtue, counter-intuitiveness a vice. It can seem, and is sometimes said, that any philosophical dispute, when pushed back far enough, turns into a conflict of intuitions about ultimate premises: ‘In the end, all we have to go on is our intuitions’. Yet analytic philosophy has no agreed or even popular account of how intuition might work, no accepted explanation of the hoped-for correlation between our having an intuition that P and its being true that P. Since analytic philosophy prides itself on its rigour, this blank space in its foundations looks like a methodological scandal.

(2004: 109)

Accounts of intuitions range from apriorist views, on which intuition is a sui generis intellectual faculty (Bealer, 1992), to those that compare intuitions to perception—a strategy that has it roots in Kant where one ‘sees’ the force of a philosophical argument (Sosa 1999), to comparisons with imagination (Chalmers 2002), to the so-called naturalist view of intuitions where intuitions are seen as judgements reached quickly (Devitt 2006; Williamson 2004).

In Naming and Necessity Kripke appeals to philosophical intuitions in a number of fundamental ways. Intuitions are used to decide the connection between names and their referents, they are also arbiters of meaning, but even more importantly they are aids for establishing the general plausibility of an argument. For instance, he writes: ‘When you ask whether it is necessary or contingent that Nixon won the election, you are asking the intuitive question whether in some counterfactual situation, this man would in fact have lost the election’ (1980: 41). Furthermore, he thinks that meaningful sentences have ‘intuitive content’, which is evident to the ‘ordinary man’ and not just to philosophers. He goes on to add,

Some philosophers think that something’s having intuitive content is very inconclusive evidence in favor of it. I think it is very heavy evidence in favor of anything, myself. I really don’t know, in a way, what more conclusive evidence one can have about anything, ultimately speaking.

(Kripke 1980: 42)

Such a stance assumes people don’t have conflicting intuitions (individually and between themselves). If one’s direct intuition in one case clashes with the consequences of one’s intuitions about other cases, as is arguably the case in the semantics of names, it’s not clear which should win out.

In addition to the central role assigned to intuitions in general, Kripke, more specifically, seems to be relying on two types of intuitions. The first are conceptual or linguistic intuitions—or what Margolis and Laurence (2003) have called ‘Socratic intuitions’, (p. 613) which are pretheoretical dispositions to apply concepts to some particular cases and scenarios and refuse to apply them to others. The second are so-called ‘modal intuitions’, which seem to boil down to a direct or unmediated knowledge of metaphysical necessity and possibility. It is this second form of intuition, in particular, that creates serious philosophical worries.

In arguing for names as rigid designators, Kripke relies on modal and counterfactual reasoning, but he seems to believe that the ‘ordinary man’ is as well placed to grasp the intuitive content of propositions as any philosopher. He argues,

Suppose that someone said, pointing to Nixon, ‘that’s the guy who might have lost’. Someone else says ‘Oh no, if you describe him as “Nixon”, then he might have lost; but, of course, describing him as the winner, then it is not true that he might have lost.’ Now which one is being the philosopher, here, the unintuitive man? It seems to me that obviously the second. The second man has a philosophical theory.

(1980: 41)

It is not clear why the famous ‘man on the street’ should possess any modal intuitions, but if his intuitions do have special evidential value, surely this would require a move away from the philosophical armchair in favour of the polling and sampling methods common in the rest of social science. Especially since, as we shall see, these intuitions are readily contested by both philosophers and non-philosopher.

Putnam, on the other hand, assigns a more modest scope to philosophical intuitions. In mathematics and science, intuition, he tells us, is a fallible but valuable guide, and a fallible guide is better than no guide at all. In particular, intuition plays a significant but not decisive role in mathematics and natural sciences to help with the formation of hypotheses (Putnam 1979b: 67). In ‘Possibility and Necessity’ and ‘Is Water Necessarily H2O?’ (1990) he expresses his initially implicit and then more clearly articulated reservations about Kripke’s view that, in addition to linguistic intuitions, there could be such a thing as metaphysical intuitions. What Putnam initially had done, he now confesses, was to try to give a ‘minimalist’ (re)interpretation to Kripke’s fuller conception of intuitions by assimilating his ‘metaphysical intuitions to the linguistic intuitions, intuitions about how we speak, that other analytic philosophers talk about’ (Putnam 1990: 64). But this is what he now thinks cannot be done. According to Putnam, Kripke’s notion of metaphysical intuition points to a fundamental capacity of reason to discover metaphysical necessities. This is the view of intuition that Putnam now fully rejects but initially, he confesses, he was trying to whitewash out of Kripke’s text. For Putnam, intuitions, in general, are a mode of access to a culture’s inherited picture of the world. And linguistic intuitions could be understood in terms of ‘sortal identity’ which gives speakers criteria for substance-identity, person-identity, table-identity, and so on. To decide on questions of identity regarding these categories, what we need to do is ‘to consult our intuitions and lay down a set of conventions which seem reasonable in the light of those intuitions’ (Putnam 1990: 64). The convention could vary in different contexts, and so would be (p. 614) context-dependent; it could also vary depending on the ‘point’ of the counterfactuals under consideration. Putnam explains:

When philosophers disagree about what are reasonable criteria of, say, person-identity across possible worlds, then (unless one of them thinks the other has a ‘tin ear,’ that is, no ear for the way we actually speak at all) they may well agree ‘one can do it either way.’ In this view, the criteria for person-identity across possible worlds are, to some extent, to be legislated and not discovered.

(Putnam 1990: 64)

Kripke, as we saw, argued that there could be empirically discovered identities that hold necessarily or in all the possible worlds where the entities exist. The thought experiments Kripke has used to elicit common intuitions on behalf of causal externalism relate to proper names such as ‘Gödel’ and ‘Jonah’; he then extends the argument to natural kind terms and empirically discovered identities. As Putnam points out, Kripke is a realist about identity conditions across possible worlds. Yet, it is not at all clear that common-sense intuitions follow his on this matter. Kripke is aware of the discrepancies and conflicts of intuitions in this area. He attempts to explain the contrary intuition that empirically discovered, but necessary, identities might not have held, by proposing the idea of ‘qualitatively identical epistemic situations’ or situations that produce the ‘same sensory evidence’ and yet which are connected with different natural kinds. For instance, there could be entities that produce gold type sensory evidence, the experience of shininess, yellowness, etc., but are not causally linked to gold. The metaphysical possibility of such a scenario explains our intuition that necessary a posteriori identities might not have held. But Kripke’s defence of this appeal to intuitions, in this particular case, demonstrates the weakness of his approach. Kripke is introducing a complex theoretical apparatus in order to overcome what seems intuitively appealing. And once we allow that intuitions are trumped by more theoretical constructs in one domain then it behoves us to question their authority in others.

What is at stake between Quine and Kripke, in particular, is not just a question of the very possibility of a theory of meaning, but the much more significant and profound issue of the nature and scope of philosophy as a discipline. Kant famously had restored philosophy to its supposedly rightful place as an autonomous and even foundational subject by arguing that the domains of analytic and synthetic are separate, that a priori truths belong uniquely to the domain of philosophy, and that metaphysics is made possible by our having access to the domain of synthetic a priori truths. Quine’s rejection of the legitimacy of these categories was a direct challenge to this loftier conception of knowledge and his naturalism reinstates philosophy as continuous with the empirical sciences. Kripke reassigns a unique role to philosophy by allowing for metaphysical intuitions;14 (p. 615) by introducing the idea of metaphysical necessity, and through his emphasis on the central role of intuitions, once again he makes it possible for philosophers to think seriously about their discipline as both rigorous and independent of the natural sciences.

The accuracy of our intuitive or common-sense judgements has been questioned in various arenas; for instance, we now know a great deal about the systematically irrational features of our instantaneous judgements concerning probability (see Tversky and Kahneman 1974). The role of intuitions has also come under considerable scrutiny recently through the work of so-called experimental philosophers who through empirical experiments, using test cases from philosophy, have attempted to cast doubt on the use of intuitions in philosophy.15 Their attack on the established view of intuitions by now covers large areas of philosophy. We will concentrate only on those that touch directly on the work of Quine, Kripke, and Putnam on the topic of meaning.

19.6 The Analytic/Synthetic Distinction and Intuitions

Philosophers who acknowledge the authority of intuitive evidence find a straightforward justification for the necessary/contingent distinction: we have a very wide range of robust modal intuitions (for example, the intuition that it is contingent that the number of planets is greater than seven—there could have been fewer), and when such intuitions are taken as evidence, the best theory is one which accepts the distinction at face value. Early criticism of Quine appealed to this idea of widespread agreement over which sentences were analytic. Grice and Strawson claimed:

We can appeal, that is, to the fact that those who use the terms ‘analytic’ and ‘synthetic’ do to a very considerable extent agree in the applications they make of them. They apply the term ‘analytic’ to more or less the same cases, withhold it from more or less the same cases, and hesitate over more or less the same cases. This agreement extends not only to cases which they have been taught so to characterize, but also to new cases. In short, ‘analytic’ and ‘synthetic’ have a more or less established philosophical use; and this seems to suggest that it is absurd, even senseless, to say that there is no such distinction. For, in general, if a pair of contrasting expressions are habitually and generally used in application to the same cases … this is a sufficient condition for saying that there are kinds of cases to which the expressions apply; and nothing more is needed for them to mark a distinction.

(1956: 143)

Similarly, J. J. Katz claimed that empirical tests would demonstrate the presence of the distinction (1967: 50). He suggested compiling lists ABCD comprising analytic, synthetic, contradictory, and anomalous sentences respectively. Subjects were to be shown (p. 616) a classification of initial range of cases into ABCD and told to classify further sentences. Katz did not carry out the procedure but he was as confident as Grice and Strawson that it would reveal habitual and general uniformity in usage. Putnam also thinks that Strawson and Grice are right to claim that those who use the terms ‘analytic’ and ‘synthetic’ do to a very considerable extent agree in the applications they make of them, but believes that the weakness of their argument is in not providing a positive account of what that distinction is: ‘It is the task of the methodologist to explain this special status, not to explain it away’ (Putnam 1979a: 92).

Rather surprisingly, this may have been a mistake. Geoffrey Sampson (1980: 70–1, 2001: 203–4) reports the results of an experiment by Morley-Bunker (1977), to determine the consistency of judgements of analyticity between and within groups of philosophically trained and untrained subjects. Morley-Bunker’s results did not support Katz’s or Grice and Strawson’s contentions that speakers by and large draw the distinction in the same way. If Morley-Bunker’s work is correct, then to the extent to which expectations of the distinction were based on a philosophical intuition, such intuition and the intuitions driving analyticity judgements themselves are shown to be fallible.16

In recent years, experimental philosophers have challenged the universality of Kripkean intuitions regarding rigid designators.17 One such experiment makes use of the (p. 617) Gödel/Schmidt story in cross-cultural contexts. Here is a brief account of the thought experiment:

Suppose that Gödel was not in fact the author of [Gödel’s] theorem. A man called ‘Schmidt’ … actually did the work in question. His friend Gödel somehow got hold of the manuscript and it was thereafter attributed to Gödel. On the [descriptivist] view in question, then, when our ordinary man uses the name ‘Gödel’, he really means to refer to Schmidt, because Schmidt is the unique person satisfying the description ‘the man who discovered the incompleteness of arithmetic’…. But it seems we are not. We simply are not.

(1980: 83–4; emphasis added)

Kripke assumes that most readers of the passage would agree with his intuition that the name ‘Gödel’ refers to the man Gödel regardless of the descriptions associated with the name. He tells us that our externalist judgements regarding the correct assignment of a name are ‘simple’ and straightforward. However, experimental data collected by Machery et al. (2004) indicate that this is not the case. They claimed to have shown that Westerners are significantly more likely to respond to Kripke’s thought experiment in accordance with social-historical accounts of reference, while East Asians are more likely to respond in accordance with descriptivist accounts of reference.18 Their data were strongly supported by experiments that Maria Baghramian conducted in Guiyang University in China in August 2010.19 The 2004 experiment, its methodology, and the accuracy of the results have been questioned in a number of ways;20 however, regardless of their shortcomings, what the results obtained by experimental philosophers, and the extensive discussion arising out of them demonstrate is that the reliability of philosophical intuition could not simply be assumed but ought to be justified.

A major difficulty in taking intuition as an essential component of the methodology of philosophy is that the term ‘intuition’ in effect acts as a placeholder for a number of vaguely defined moves in philosophy. What the different intuitive gambits in ethics, epistemology, metaphysics, philosophy of language, etc., have in common is the absence of well-specified steps in reasoning and the largely unreflective character of the resulting judgements. Intuitions are often the starting point of philosophical arguments, because (p. 618) arguments have to start somewhere. This is the role of intuition emphasized by Devitt and Williamson. They are also the ultimate tribunal where we seem to come to the end of explicit procedure for philosophical argumentation. They are used as handy and easy decision procedures for some of the most important and all-embracing philosophical questions, realism vs. anti-realism, for instance, as well as detailed issues in sub-areas of philosophy, e.g. the trolley problem or the innocent person hooked up to the famous violinist (Thomson 1971). It is a common experience that some statements or philosophical positions just seem right to us. But the metaphor of seeming should not be taken too seriously. It is not so much a particular faculty of mind’s eye that gives us the impression that we have a deep and immediate philosophical insight, but our inability to spell out some crucial steps of the argument clearly or carefully. Faced with such a situation, the use of the term ‘intuition’ should carry a warning sign as it signals an inability rather than providing a short cut to philosophical truth.

And yet, intuitions played a formative role in the revolution in philosophy of language wrought by Putnam and Kripke. If the experimental philosophers are correct in their view that our intuitions vary with cultural and socio-economic background or gender then the authority assigned to them by Kripke and other philosophers of his ilk begins to flounder. Putnam’s more modest position, which attempts to stay clear of metaphysical aspirations, fares better. His advocacy of a more context sensitive and changeable conception of both analyticity and necessity, where they are seen as contextual or relative to a particular body of belief, could accommodate the worries raised by experimental philosophers. Above all, the increasing willingness of philosophers to subject premises supplied by intuition to empirical examination harkens back to Quine’s call for a ‘naturalized epistemology’ and his vision of a profound rapprochement between philosophy and the other sciences. Such a trend, if fully established, could ultimately give rise to an empirically informed conception of meaning and the a priori.21


Baghramian, Maria (forthcoming). ‘Semantic Intuitions Chinese Style’.Find this resource:

Bealer, George (1992). ‘The Incoherence of Empiricism’, Proceedings of the Aristotelian Society, Supplementary Volume 66: 99–138.Find this resource:

Boghossian, Paul (1997). ‘Analyticity’. In Bob Hale and Crispin Wright (eds.), A Companion to the Philosophy of Language. Oxford: Blackwell, pp. 331–68.Find this resource:

Burge, Tyler (1978). ‘Belief and Synonymy’, Journal of Philosophy 75 (3): 119–38.Find this resource:

Carnap, Rudolf (1937). The Logical Syntax of Language, tr. A. Smeaton. London: Kegan, Paul. Originally published in German in 1934.Find this resource:

Carnap, Rudolf (1947). Meaning and Necessity. Chicago: University of Chicago Press.Find this resource:

Chalmers, David J. (2002). ‘Does conceivability entail possibility?’. In Tamar S. Gendler & John Hawthorne (eds.), Conceivability and Possibility. Oxford: Oxford University Press.Find this resource:

(p. 619) Chomsky, N. (2000). New Horizons in the Study of Language and Mind. Cambridge: Cambridge University Press.Find this resource:

Devitt, M. (2006). ‘Intuitions in Linguistics’, British Journal for the Philosophy of Science 57 (3): 481–513.Find this resource:

Devitt, M. and R. Hanley (eds.) (2006). The Blackwell Guide to the Philosophy of Language. Oxford: Blackwell.Find this resource:

Dummett, M. (1973). Frege: Philosophy of Language. London: Duckworth.Find this resource:

Evans, G. (1973). ‘The Causal Theory of Names’, Proceedings of the Aristotelian Society, Supplementary Volume 47: 187–208.Find this resource:

Frege, G. (1884). Die Grundlagen der Arithmetik. Breslau: W. Koebner. Tr. J. L. Austin as The Foundations of Arithmetic. Oxford: Blackwell, 1950.Find this resource:

Frege, G. (1892). ‘Über Sinn und Bedeutung’, Zeitschrift für Philosophie und philosophische Kritik 100: 25–50. Tr. in M. Beaney (ed.), The Frege Reader. Oxford: Blackwell, 1997, pp. 151–71.Find this resource:

Grice, H. P. and P. F. Strawson (1956). ‘In Defence of a Dogma’, Philosophical Review 65(2): 141–58.Find this resource:

Hintikka, J. (1999). ‘The Emperor’s New Intuitions’, Journal of Philosophy 96: 127–47.Find this resource:

Jackson, F. (1998). From Metaphysics to Ethics. Oxford: Oxford University Press.Find this resource:

Kant, I. (1781/1787). Critique of Pure Reason, tr. Norman Kemp Smith. Basingstoke: Macmillan, 1965.Find this resource:

Katz, J. (1967). ‘Some Remarks on Quine on Analyticity’, Journal of Philosophy 64 (2): 36–52.Find this resource:

Knobe, J. and S. Nichols (2008). ‘An Experimental Philosophy Manifesto’. In J. Knobe and S. Nichols (eds.), Experimental Philosophy. Oxford: Oxford University Press, pp. 3–16.Find this resource:

Kripke, S. (1980). Naming and Necessity. Oxford: Basil Blackwell.Find this resource:

Machery, E., R. Mallon, S. Nichols, and S. Stich (2004). ‘Semantics, Cross-Cultural Style’, Cognition 92: B1–B12.Find this resource:

Maitra, I., B. Weatherson, and J. Ichikawa (2012). ‘In Defense of a Kripkean Dogma’, Philosophy and Phenomenological Research 85(1): 56–68.Find this resource:

Mallon, R., E. Machery, S. Nichols, and S. Stich (2009). ‘Against Arguments from Reference’, Philosophy and Phenomenological Research 79: 332–56.Find this resource:

Margolis, E. and S. Laurence (2003). ‘Should We Trust Our Intuitions? Deflationary Accounts of the Analytic Data’, Proceedings of the Aristotelian Society 103(3): 299–323.Find this resource:

Marti, G. (2009). ‘Against Semantic Multi-Culturalism’, Analysis, 69 (1): 42–8.Find this resource:

Mill, J. S. (1874). ‘The Utility of Religion’. In Three Essays on Religion. New York: Cosimo, 2008, pp. 69–124.Find this resource:

Morley-Bunker, N. (1977). ‘Speaker’s Intuitions of Analyticity’, unpublished BA dissertation, University of Lancaster. Available online at: <>, accessed 3 December 2012.

Nagel, J. (2012). ‘Intuitions and Experiments: A Defense of the Case Method in Epistemology’, Philosophy and Phenomenological Research 85(3): 495–527.Find this resource:

Putnam, H. (1962). ‘It Ain’t Necessarily So’, Journal of Philosophy 59(22): 658–71.Find this resource:

Putnam, H. (1975a). ‘The Analytic and the Synthetic’. In Mind, Language and Reality: Philosophical Papers, Volume 2. Cambridge: Cambridge University Press, 1979, pp. 33–69.Find this resource:

Putnam, H. (1975b). ‘The Meaning of “Meaning”’, Minnesota Studies in the Philosophy of Science, 7: 131–93; repr. in Mind, Language and Reality: Philosophical Papers, Volume 2. Cambridge: Cambridge University Press, 1979, pp. 215–71.Find this resource:

Putnam, H. (1976). ‘“Two Dogmas” Revisited’. In G. Ryle (ed.), Contemporary Aspects of Philosophy. London: Oriel Press, pp. 202–13.Find this resource:

(p. 620) Putnam, H. (1979a). ‘Philosophy of Physics’. In Mathematics, Matter, and Method: Philosophical Papers, Volume 1. Cambridge: Cambridge University Press, pp. 79–92.Find this resource:

Putnam, H. (1979b) ‘What is Mathematical Truth?’ In Mathematics, Matter, and Method: Philosophical Papers, Volume 1. Cambridge: Cambridge University Press, pp. 60–78.Find this resource:

Putnam, H. (1990). ‘Is Water Necessarily H2O?’ In Realism with a Human Face. Cambridge, MA: Harvard University Press, pp. 54–79.Find this resource:

Putnam, H. (2002). ‘Quine’, Common Knowledge 8 (2): 273–9.Find this resource:

Quine, W. V. O. (1943). ‘Notes on Existence and Necessity’, Journal of Philosophy 40(5): 113–27.Find this resource:

Quine, W. V. O. (1947). ‘The Problem of Interpreting Modal Logic’, Journal of Symbolic Logic 12 (2): 43–8.Find this resource:

Quine, W. V. O. (1953). From a Logical Point of View. Cambridge, MA: Harvard University Press; 2nd edn. 1961.Find this resource:

Quine, W. V. O. (1956). ‘Quantifiers and Propositional Attitudes’, Journal of Philosophy 53; repr. in The Ways of Paradox and Other Essays. New York: Random House, 1966, pp. 185–96.Find this resource:

Quine, W. V. O. (1960). Word and Object. Cambridge, MA: MIT Press.Find this resource:

Quine, W. V. O. (1969). Ontological Relativity and Other Essays. New York: Columbia University Press.Find this resource:

Quine, W. V. O. (1979). ‘Use and its Place in Meaning’, Studies in Linguistics and Philosophy 3: 1–8.Find this resource:

Quine, W. V. O. (1986). Philosophy of Logic. Cambridge, MA: Harvard University Press.Find this resource:

Quine, W. V. O. (1987). ‘Indeterminacy of Translation, Again’, Journal of Philosophy 84 (1): 5–10.Find this resource:

Sampson, G. (1980). Making Sense. Oxford: Oxford University Press.Find this resource:

Sampson, G. (2001). Empirical Linguistics. London: Continuum.Find this resource:

Searle, J. R. (1983) Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press.Find this resource:

Soames, S. (2003). Philosophical Analysis in the Twentieth Century, 2 vols. Princeton, NJ: Princeton University Press.Find this resource:

Sosa, E. (1999). ‘Minimal Intuition’. In M. R. DePaul and W. Ramsey (eds.), Rethinking Intuition: The Psychology of Intuition and its Role in Philosophical Inquiry. Lanham, MD: Rowman & Littlefield, pp. 257–69.Find this resource:

Thomson, J. J. (1971). ‘A Defense of Abortion’, Philosophy and Public Affairs 1 (1): 47–66.Find this resource:

Tversky, A. and D. Kahneman (1974). ‘Judgment Under Uncertainty: Heuristics and Biases’, Science 185: 1124–31.Find this resource:

White, M. (1950). ‘The Analytic and the Synthetic: An Untenable Dualism’. In John Dewey: Philosopher of Science and Freedom. New York: Dial Press; repr. in Leonard Linsky (ed.), Semantics and the Philosophy of Language. Urbana: University of Illinois Press, 1952, pp. 272–86.Find this resource:

Williamson, T. (2004). ‘Philosophical “Intuitions” and Scepticism about Judgement’, Dialectica 58 (1): 109–53.Find this resource:

Williamson, T. (2005). ‘Armchair Philosophy, Metaphysical Modality and Counterfactual Thinking’, Proceedings of the Aristotelian Society 105(1): 1–23.Find this resource:

Williamson, T. (2007). The Philosophy of Philosophy. Oxford: Blackwell.Find this resource:

Wittgenstein, L. (1921). Tractatus Logico-Philosophicus, tr. B. F. McGuinness and D. Pears. London: Routledge and Kegan Paul, 1975.Find this resource:


(1) The first wave of externalism about meaning has been followed by a second where Tyler Burge and others attempted to externalize not just meaning but mental content. For an account of Frege’s influence in this story, see Burge’s chapter in the present volume.

(2) The story of Frege and Russell’s descriptivism is much more complicated than it is commonly presented, however, but a detailed discussion of their position is beyond the scope of this chapter.

(3) Michael Dummett has questioned this interpretation. He thinks that Frege did not necessarily identify sense with descriptions or see them as intermediaries between thinker and the object of reference. But Dummett does admit that Frege’s examples of the sense of a name are definite descriptions (Dummett 1973: 110).

(4) Morton White, a sometime colleague of Quine at Harvard, had also produced criticisms similar to Quine’s in 1950.

(5) Tyler Burge anticipated Tim Williamson’s point in his 1978 article.

(6) At least retrospectively, this last of the three arguments offered by Kripke has become less significant as post-Kripkean descriptivists, including two-dimensionalists, now employ rigidified descriptions using the modifier ‘actually’.

(7) Possible worlds, for Kripke, are not obscure metaphysical objects but ways the world might have been. In the preface to Naming and Necessity, he explains: ‘I will say something briefly about “possible worlds”…. In the present monograph I argued against those misuses of the concept that regard possible worlds as something like distant planets, like our own surroundings but somehow existing in a different dimension, or that lead to spurious problems of “transworld identification”. Further, if one wishes to avoid the Weltangst and philosophical confusions that many philosophers have associated with the “worlds” terminology, I recommended that “possible state (or history) of the world”, or “counterfactual situation” might be better. One should even remind oneself that the “worlds” terminology can often be replaced by modal talk—“It is possible that…”’ (15).

(8) We refer to Kripke’s preferred picture of naming as his ‘social-historical’ account, and not the ‘causal account of reference’, as is conventional, because what is distinctive about the view is the transmission of knowledge of reference from one speaker to another, which is certainly social and historical, but only causal in so far as social and historical connections are causal. The initial baptism by which the reference of the name is fixed may, though it need not, indeed be causal in a more direct way.

(9) A move that, as Putnam admits, turned him, in the eyes of the ardent Quinean Burton Dreben, into a Girondist to Quine’s revolutionary aspirations (Putnam 2002).

(10) At the time Quine was writing, Pluto was considered a planet in its own right. So there were considered to be nine solar planets.

(11) What does not emerge from our focus on Quine, Putnam, and Kripke is that Kripke’s work, unlike Putnam’s, is part of a significant tradition of formal intensional semantics going back to Carnap, through people like Montague and David Lewis, and now flourishing in linguistics as well as philosophy. We owe this point, and many more than we could mention, to Tim Williamson.

(12) Identity statements involving rigid designators are necessarily true, if true. Simplifying things a little, suppose ‘a’ and ‘b’ are rigid designators for the same individual. Then a = b is true (in the actual world). But since the names both refer to the same individual and that individual continues to be that same individual wherever they are, the names refer to the same individual in every possible world. So the identity is necessary.

(13) For instance, in the case of the Twin Earth thought experiment—the most celebrated of all of them—through the telling of a science fiction story, we are led to the conclusion that what tastes, smells, or looks like water would not be correctly labelled water unless it has the underlying molecular structure of H2O.

(14) Putnam’s role in this story is more complicated, at least partly because of his willingness to rethink his philosophical positions. As we saw, Putnam is to some extent in agreement with Quine on the analytic/synthetic distinction, or so Quine thought. He also rejects the Kripkean talk of metaphysical intuitions, but Putnam in his later work reintroduces the role of intuition in philosophy by giving pride of place to ‘common sense’.

(15) For instance see Knobe and Nichols, ‘An Experimental Philosophy Manifesto’ (2008).

(16) The experiment has not been published so it is worth giving a slightly more detailed summary than is conventional. Morley-Bunker’s subjects, 28 non-philosophy students and faculty and 10 philosophy faculty, were given cards displaying clear cases of analytic and synthetic sentences respectively and asked to sort these into piles. Their allocation was corrected by the experimenter as necessary, according to the generally perceived status of the sentences as analytic or synthetic. Thereafter, the subjects were asked to sort 18 further samples. The experiment assumed no fact of the matter as to the status of these further sentences. The goal was simply to determine the consistency of the different groups. Sampson reports the results in detail:

The results were as would be expected by sceptics of the analytic/synthetic distinction, in that neither group was internally consistent. Two sentences were assigned to a single category by a highly significant majority (p < 0.005) of each group of subjects, and for each group there were three more sentences on which a consistent classification was imposed by a significant majority (p < 0.05) of members of that group; on ten sentences neither group deviated significantly from an even split between those judging it analytic and those judging it synthetic. (To give examples: Summer follows Spring was judged analytic by highly significant majorities in each group of subjects, while We see with our eyes was given near-even split votes by each group.) Overall, philosophers were somewhat more consistent than the non-philosophers. But the global finding conceals results that manifested the opposite tendency. Thus Thunderstorms are electrical disturbances in the atmosphere was judged analytic by a highly significant majority of non-philosophers, while a (non-significant) majority of philosophers deemed it synthetic. […] Nothing can be completely red and green all over was judged analytic by a significant majority of philosophers but only by a non-significant majority of non-philosophers.

(2001: 203)

(17) Similarly there have been claims that there are gender differences in intuitive responses to Putnam’s Twin Earth thought experiments between men and women and that women are more likely to believe that Oscar and Twin-Oscar mean different things when they say ‘water’. The results, however, are controversial and have been questioned by Jennifer Nagel (2012), among others.

(18) Machery et al. (2004) found that 56.5% of Western participants gave a causal historical response compared to 31.5% of East Asian participants. The difference between the scores for the two groups is statistically significant, and Machery et al. conclude that there is cross-cultural variation in the responses to this probe.

(19) In a repeat of this experiment in Guangzhou University in China Maria Baghramian obtained even more dramatic results than those reported by Machery et al. Ninety-two per cent of the participants in the experiment, consisting of Ph.D. students and lecturers with some familiarity with philosophy of language, opted for a descriptivist, Russellian account of meaning and reference (Baghramian forthcoming).

(20) For instance, the experiment has been criticized by Genoveva Marti and others for inadequacies in its design and methodology. Marti argues that the experiment tests ‘people’s intuitions about theories of reference, not about the use of names’ (2009: 43). I believe Marti misstates the aim of the experiment, but will not argue the point here. For a discussion of various criticisms of the experiment see Maitra et al. 2012.

(21) We would like to thank Tim Williamson and Michael Beaney for their extensive and very helpful comments on earlier drafts of this chapter.