Carnap, Quine and the Fate of Metaphysics

Huw Price

0. The Carnap Case

[1] Imagine a well-trained mid-century American philosopher, caught in a rare traffic jam on the New Jersey Turnpike one warm summer afternoon in the early 1950s. He dozes in his warm car ... and awakes in the same spot on a chill Fall evening in the late 1990s, remembering nothing of the intervening years. It is as if he has been asleep at the wheel for almost half a century!

[2] Let's suppose that he sees the up side of his peculiar situation. Phenomenologically, it is on a par with time travel, and what self-respecting philosopher could fail to be excited by that? Of course, he realises that it is far is more likely that he is suffering from amnesia than that he has actually been transported more than forty years into the future, or survived for that long on the Turnpike -- but all the more reason to savour the experience while he can, lest his memory should soon return.

[3] Indeed, he soon becomes a celebrity, written up by Oliver Sachs in The New York Times. Irreverent graduate students call him (with apologies to (Beth 1963: 478)) the Carnap* of contemporary philosophy, and everyone is interested in his impressions of modern life. What will surprise him about the society in which he finds himself? Any Australian philosopher who knows contemporary New York will find it easy to imagine some of the things that might stand out: the number of people who ask for change for a cup of coffee, the mind-numbing range of options available when he buys his own cup of coffee, the fact that all the options are mediocre, and so on. But let's suppose that Carnap* has the true philosopher's ability to ignore all of this. He wants to know what has happened to his own beloved discipline. "Never mind the beggars and the Starbucks!", he says, "Where are the big strides in philosophy this past half century?"

[4] At this point, I think, Australian intuitions are less reliable. Australian philosophical audiences find familiar one of the features of late-century philosophy that Carnap* will find most surprising, namely the apparent health of metaphysics at the end of century. Back in 1950, Carnap* recalls, metaphysics, like poverty, was supposed to be on its last legs. Yet everywhere that Carnap* turns these days, there's a philosopher espousing a metaphysical position -- someone claiming to be a "realist" about this, or an "anti-realist" about that. Out in the college towns of New Jersey and New England, Carnap* finds, there are more ontological options than kinds of coffee, more metaphysicians than homeless people. And metaphysics isn't simply an affliction of the aged, infirm and mentally ill. Like the Great War of Carnap*'s own childhood, it seems to have claimed the best and brightest of a generation. "When will they ever learn", he hums to himself -- a sign perhaps to us, if not to him, that his memory of the intervening years is beginning to return.

[5] If Carnap* were to enquire where the battle against metaphysics had been lost in 20th century philosophy, he might do well to turn his attention to a skirmish between his namesake and Quine in the early 1950s. In philosophy, as in less abstract conflicts, single engagements are rarely decisive, but this particular clash does seem of special significance. By the late 1940s, Carnap's views seem to represent the furthest advance of the anti-metaphysical movement, at least on one of its several fronts. The fact that the position was never consolidated, and the ground lost, seems to owe much to criticism of Carnap's views by Quine in the 1950s. Ironically, Quine's criticism was friendly fire, for Quine too opposed traditional metaphysics. The attack was no less damaging for the fact that it came from behind, however, and its effect seems to have been to weaken what -- at that time, at any rate -- seems to have been Quine and Carnap's common cause.

[6] However, this paper isn't an attempt at historical reconstruction. Nor am I concerned, at least directly, with the relevance to contemporary metaphysical positions of Carnap's anti-metaphysical views. (After all, it may be that -- perhaps due to Quine's influence -- what passes for metaphysics these days is not what Carnap attacked.) My interest in the Carnap-Quine debate stems from the fact that elsewhere (Price 1997) I make use of Carnap's view in order to block certain objections to a position which, as I argue in that paper, doesn't seem to have been given its due in contemporary metaphysics. The position is one I call descriptive or functional pluralism. Its key idea is that different bits of descriptive discourse serve different functions in language, in a way which, once recognised, undercuts certain kinds of reductionist moves. As I note in that paper, Carnap himself seems to have held such a view, at least in a limited form. The present paper stems from my attempts to clarify in my own mind the nature, basis and viability of Carnap's views, and their relation to this kind of pluralism in metaphysics more generally.

[7] Hence, in particular, my interest in Quine's criticisms of Carnap's view of the nature and place of metaphysics. It seems widely believed that Quine showed that Carnap's position is untenable, but I want to show that this conclusion is too strong. It turns out that there are two distinct strands to Carnap's position -- in effect, weaker and stronger anti-metaphysical doctrines. Quine himself accepts a version of the weaker doctrine, under the label "ontological relativity". He rejects the stronger doctrine, but I want to argue that the issue is much more open than Quine allows. Carnap himself may not have provided the materials for a defence of this doctrine, but others have done so. Provided we add something to Carnap's view, then, even its stronger doctrine remains a viable position. The issue turns out to rest on a first-order scientific issue about linguistic behaviour which Quine, of all people, is in no position to rule on a priori.

[8] Indeed, there turns out to be rather satisfying two-way relationship between these two doctrines in Carnap and my use of his views in defence of functional pluralism. On the one hand, my use of his views seems to need only the weaker doctrine, which Quine accepts -- so no problems for me there. On the other hand, the key to a defence of the stronger doctrine seems to be descriptive pluralism itself. The first-order scientific issue in question just is the issue as to whether descriptive language is functionally homogeneous, in the way that Quine implicitly affirms and Carnap implicitly denies. In other words, it is just the issue I try to make visible in that paper.

1. Against Metaphysics

[9] The rejection of traditional metaphysics was one of the key projects of logical positivism, indeed of the positivist movement in general. Positivists hold that meaningful enquiries are of one of two kinds: empirical enquiries, answerable to observation, and logical enquiries, founded on relations of meaning. In so far as traditional metaphysics falls under neither heading, its enquiries are devoid of either kind of content. Metaphysics becomes an elaborate system of idle cogs -- a self-sustaining game, which connects with nothing in reality. In broad terms, the view goes back at least to Hume.

[10] For the logical positivists the point takes a particularly sharp form. Metaphysical statements are unverifiable, and are hence meaningless, or at least without "cognitive meaning". This was the prevailing view among the positivists of the Vienna Circle in the 1920s and 1930s. In this form, however, the position is vulnerable to criticism of the logical positivists' verificationist criterion for meaningfulness. Hence it is important that Carnap's critique of metaphysics, though continuous in spirit with earlier positivist views, relies on a different point about language. Carnap's point is not simply that metaphysical claims are unverifiable, but that the activity of metaphysics relies on a confusion about what can be done with language.

[11] Carnap argues that there is no theory-independent ontological viewpoint available to metaphysics. Ontological questions about the entities mentioned in a particular theory or linguistic framework can properly be raised as what Carnap calls 'internal questions' -- questions posed within the framework or theory in question -- but not as 'external questions', posed from a stance outside that framework. One of Carnap's examples concerns what he calls "the thing world": "the spatio-temporally ordered system of observable things and events" (1952: 210).

[12] Once we have accepted this thing-language and thereby the framework of things, we can raise and answer internal questions, e.g., "Is there a white piece of paper on my desk", "Are unicorns... real or merely imaginary", and the like. These questions are answered by empirical investigations.

The concept of reality occurring in these internal questions is an empirical, scientific, non-metaphysical concept. To recognize something as a real thing or event means to succeed in incorporating it into the framework of things at a particular space-time position so that it fits together with the other things recognized as real, according to the rules of the framework. (1952: 210)

[13] Carnap goes on to say that "from these questions we must distinguish the external question of the reality of the thing world itself". He notes that this is the kind of question metaphysicians take themselves to be addressing, but argues that

it cannot be solved because it is framed in the wrong way. To be real in the scientific sense means to be an element of the framework; hence this concept cannot meaningfully be applied to the framework itself.... The thesis of the reality of the thing world... cannot be formulated in the thing language or, it seems, in any other theoretical language. (1952: 210-11)
He suggests that what the metaphysicians perhaps had in mind is "not a theoretical question as their formulation seems to suggest, but rather a practical question, a matter of a practical decision concerning the structure of our language. We have to make the choice whether or not to accept and use the forms of expression for the framework in question" (1952: 210).

[14] Thus Carnap allows that there is a legitimate pragmatic issue which may be raised from the external standpoint -- roughly, an issue concerning the utility of the framework concerned. What is disallowed, Carnap tells us, is the external ontological question as to whether the framework is true, or corresponds to reality. Ontological questions are legitimate only if internal, for they presuppose the adoption of rules specifying a domain of discourse. Carnap says that traditional metaphysics confuses formal issues about the structure and utility of linguistic frameworks, for ontological issues about the nature of reality.

[15] There seem to be a number of ingredients to Carnap's position, which need to be distinguished:

1. Linguistic internalism. We can't speak from outside language, but only from within the constraints of those linguistic frameworks to whose rules we already conform.

2. Framework pluralism. There are different frameworks in play in language, doing different jobs; and philosophical confusion may result if we fail to notice this fact.

3. Frameworks are ontologically committing. The adoption of a framework involves ontological commitment to the entities over which quantifiers of the framework range. This commitment isn't a separate thing, of course. It is not that we say "We adopt these quantifiers, and we also accept that such things exist", but rather that there isn't anything more to accepting the existence of a category of entities than being prepared to quantify over the things in question.

4. Ontology is framework-dependent. This is a corollary of (1). We can't speak from outside linguistic frameworks altogether, and hence there is no standpoint for ontology other than that supplied by the frameworks to which we subscribe.

5. Ontological pluralism. A corollary of (2) and (4). To the extent that different frameworks are independent, and doing different jobs, their existential quantifiers also seem to be doing different jobs -- each framework seems to bring with it its own notion of reality.

6. Pragmatism about the adoption of frameworks. There are legitimate pragmatic issues which may raised about a framework, from a perspective external to the framework in question (though not, of course, to all frameworks). At least in principle, the practical decision to adopt a framework turns on issues of this kind.
With these ingredients in mind, let's turn to Quine's criticism of Carnap's view. With which ingredients does Quine disagree? And -- bearing in mind that this is our main concern -- to what extent does this disagreement touch Carnap's anti-metaphysical conclusions?

[16] Quine criticises Carnap on two main grounds. His first objection is to what I have called Carnap's pluralism -- Carnap's idea that adoption of a new framework involves adoption of new quantifiers, whose range is restricted to the new entities countenanced by the framework in question. Against this, Quine argues that there is no principled reason to think that we need more than one existential quantifier, capable of ranging over anything at all. Quine appears to hold that the alternative is to grant that the notion of existence is ambiguous, or non-univocal, a view he regards as unacceptable. This theme is one to which Quine returns elsewhere. In a well-known passage of Word and Object, section 27, he takes Ryle to task for, as Quine seems to see it, a similar doctrine about the plurality of notions of existence.

[17] Quine thus rejects ingredients (2) and (5): Carnap's framework pluralism and ontological pluralism. Later in the paper I want to argue, against Quine, that the sort of pluralism required by Carnap's view does not conflict with the intuition that the existential quantifier is in an important sense univocal. (Indeed, I think that Ryle can be invoked in Carnap's defence at this point.) Before that, however, I want to consider Quine's second criticism, which he himself regards as the more fundamental -- the "basic point of contention" (1966: 133), as he puts it. It is that Carnap's internal-external distinction hangs on an illegitimate appeal to the analytic-synthetic distinction.

2. The Relevance of the Analytic-Synthetic Distinction

[18] Recall that Carnap says that in so far as external issues are legitimate, they are pragmatic in nature. The issue is whether a framework is useful, what it would do for us, not whether it is true. Legitimate non-pragmatic issues are necessarily internal, on Carnap's view, and hence of no use in vindicating traditional metaphysics. These internal questions may be analytic or empirical. In case of mathematics, for example, all are analytic, though with different degrees of immediacy: the existence of numbers is immediate, but that of prime numbers between 500 and 510 presumably less so.

[19] However, Quine argues that in virtue of the failure of the analytic-synthetic distinction, even internal question are ultimately pragmatic (and not sharply distinguished into analytic and synthetic groups). Referring to Carnap's view that, as Quine puts it, "philosophical questions are only apparently about sorts of objects, and are really pragmatic questions of language policy" (1960: 271). Quine asks: "But why should this be true of the philosophical questions and not of theoretical questions generally? Such a distinction of status is of a piece with the notion of analyticity, and as little to be trusted" (1960: 271). In other words, Quine's claim is that there are no purely internal issues, in Carnap's sense. Our commitment to a framework is never absolute, and no issue is entirely insulated from pragmatic concerns about the possible effects of revisions of the framework itself. Pragmatic issues of this kind are always on the agenda, at least implicitly. In the last analysis, all judgements are pragmatic in nature.

[20] Let us grant that this is true. What effect does it have on Carnap's anti-metaphysical conclusions? Carnap's internal issues were of no use to traditional metaphysics, and metaphysics does not lose if they are disallowed. But does it gain? Science and mathematics certainly lose, in the sense that they become less pure, more pragmatic, but this is not a gain for metaphysics. And Quine's move certainly does not restore the non-pragmatic external perspective required by metaphysics. In effect, the traditional metaphysician wants to be able to say, "I agree it is useful to say this, but is it true?" Carnap rules out this question, and Quine does not rule it back in.

[21] Quine sometimes invites confusion on this point, I think. He says that "if there is no proper distinction between analytic and synthetic, then no basis at all remains for the contrast which Carnap urges between ontological statements [i.e., the metaphysical statements that Carnap wants to disallow] and empirical statements of existence. Ontological questions then end up on a par with the questions of natural science" (1966: 134). This sounds like good news for ontology, but actually it isn't. Quine's criticism of Carnap is in no sense a vindication of metaphysics. Of course not, for if all issues are ultimately pragmatic, there can't be the more than-pragmatic issue of the kind the metaphysician requires. The main effect of abandoning the analytic-synthetic distinction is that Carnap's distinctions are no longer sharp -- there are no purely internal (non-pragmatic) issues, because, as Quine has shown, linguistic rules are never absolute, and pragmatic restructuring is never entirely off the agenda. But a metaphysician who takes this as a vindication of his position -- someone who announces triumphantly that Quine has shown us that metaphysics is in the same boat as natural science, that "ontological questions [are] on a par with the questions of natural science" -- is someone who, in Brecht's words, has not been told the terrible news. Quine himself has sunk the ontologists' traditional boat, and left all of us, scientists and ontologists, clinging to Neurath's raft.

[22] As Quine himself puts it elsewhere,

Carnap maintains that ontological questions... are questions not of fact but of choosing a convenient scheme or framework for science; and with this I agree only if the same be conceded for every scientific hypothesis. (1966: 134)
Indeed -- leaving aside Carnap's nascent pluralism -- Carnap's point is very close to that developed explicitly by Quine in his classic essay "Ontological Relativity": We cannot address ontological issues from outside language, but only from the standpoint of some background theory. The viewpoint of ontology is always internal in this sense.

[23] In other words, if we delete the pluralist ingredients from the list above, Quine appears to agree with what remains. What Quine has attacked is something which wasn't on the list, namely the view that there are, in practice, some purely internal issues, in Carnap's sense. It is not clear that Carnap would have disagreed with Quine on this point. As Stathis Psillos (1997) notes, Carnap had affirmed much earlier that linguistic rules are not rigid: "All rules are laid down with the reservation that they may be altered as soon as it seems expedient to do so" (Carnap 1937: 318). This doesn't sound like the view of someone who thinks that pragmatic issues can be quarantined in science. But whether Carnap disagrees with Quine or not, the point isn't essential to his anti-metaphysical view.

[24] I want to turn to the issue which does seem to separate Quine from Carnap, that of pluralism. First, however, it may be helpful to revisit Carnap's main thesis. Carnap claims that ontological issues cannot be addressed from an external standpoint, except in pragmatic terms. But why is this so, and what underlies the internal-external distinction? To the extent that Carnap's paper answers these questions, it may seem to rely on Carnap's rather formal conception of the structure of language -- a fact which perhaps explains its apparent vulnerability to Quine's much more pragmatic conception of language. However, I think the core of Carnap's thesis is actually independent of a formal conception of language, and relies on a distinction which no Quinean should fail to heed.

3. Reassessing Carnap's Internal-External Distinction

[25] Consider a system of concepts -- again, those of the theory of number, for example. The traditional metaphysician wants to know whether anything answers to these concepts -- whether there are numbers, to put it more simply. Carnap's thesis comes down to two basic points, I think: first, that we cannot ask such questions without using the concepts concerned, and second, that their use already commits us to the existence of numbers. We cannot use the concept unless we use it in accordance with the rules which govern its use, and yet these rules make it analytic that there are numbers.

[26] Thus Carnap distinguishes internal questions, which may be raised within a linguistic framework or system of concepts, from external questions, which are addressed from some other standpoint. As I say, I think the most helpful way to understand this distinction is in terms of the notion of use. Internal questions simply are those that can only be posed by using the concepts concerned. Carnap points out that although some ontological questions are properly internal -- "Are there prime numbers greater than 100?", for example -- the traditional concerns of metaphysics do not lie at this level. And nor do they properly lie at the external level, for one cannot ask whether numbers exist without using the concept number, at which point one's use is internal, not external. One cannot "subtract" the commitments which are part and parcel of proper use of the concept.

[27] We saw that Carnap does not rule out the external perspective altogether. He allows that from outside a framework -- without using its concepts -- we can consider the issue as to whether we should adopt it. But this standpoint is purely pragmatic, according to Carnap. We can consider the consequences of adopting the framework, but we can't ask whether anything answers to its concepts, or whether its claims are true or false. This external stance is like that of an anthropologist, who studies a pattern of apparently linguistic behaviour from the outside, without ascribing interpretations to its ingredients. The anthropologist is not in a position to pass judgement on the truth and falsity of assertions made within the practice in question. Indeed, it is controversial whether from this perspective there can be good grounds for saying that the practice involves assertions at all -- Davidson denies it, for example. However, it is uncontroversial that the anthropologist can assess the pragmatic significance of the practice in the lives of the people in question, just as she can that of practices which are not linguistic utterances at all, in themselves (practices such as marriage, football and afternoon tea, for example). This anthropological assessment is the role that Carnap allows for metaphysics, pragmatically construed.

[28] It is true that in practice, insiders may often be better placed to answer these pragmatic questions than the anthropologist herself. After all, insiders are in a position to enjoy the pragmatic advantages which the anthropologist can only hypothesize about at second-hand. (This is true of non-linguistic cases, too, of course.) This doesn't alter the character of the anthropological stance, however. It remains a sideways-on reflection on practice, and to extent that this sideways-on perspective on one's own practice is difficult to achieve, the advantages of detachment may outweigh those of first-hand experience.

[29] I think it is important not to confuse the externality of the anthropological perspective with the pseudo-externality provided by what Quine calls semantic ascent. Semantic ascent is available to us from within a linguistic framework. It allows us to pose what are really internal questions by talking about the framework itself. Instead of asking "Is there a prime number greater than 100" we can ask "Is the sentence 'There is a prime number greater than 100' true?". Questions of this kind are only sensible if the sentences in question retain their interpretation. Otherwise, it would be like asking 'Is the sentence "!@#$%^&*" true?' -- without the framework, the marks on the page are simply not words, in any determinate sense. So Carnap's external standpoint must be more remote. In principle, the issue as to whether to adopt a framework is like the issue as to whether to interpret in the first place -- though again, the issues concerned may often be best addressed after the fact, in reflection on the value of existing practices.

[30] We might say that on Carnap's view, the metaphysician's mistake is to think that he can stand in two places at the same time: both within the circle, so as to claim entitlement to use the terms that have their home there; and also outside the circle, so as to challenge what membership of the circle entails -- to ask whether what it presupposes is actually true. The result is a bit like trying to ask whether the rules of a game are true. For example, someone might say, "Chess players take it for granted that the chess board has sixty four squares, but is this really the case?" -- meaning not simply to challenge the counting ability of those who claim to find sixty four squares on actual chess boards, but the very assumption that this is the right size for the board to be. Clearly, something goes wrong here. If we're not talking about a game played on a eight by eight board, we're not talking about chess. There is a legitimate pragmatic issue as to whether such a game is better or worse than various alternatives, but this is not the issue as to whether the rules correspond with the way things actually are.

[31] The effect of Quine's attack on the analytic-synthetic distinction is to make linguistic examples of this kind of thing a little less sharp. No linguistic rule is sacred, or immune from challenge. But if such a challenge is to be analogous to asking whether chess is rightly played on an eight by eight board, the latter question must be taken in its pragmatic sense: "Would an alternative be more useful?", not "Is this version true?"

[32] There do seem to be some linguistic presuppositions which not even Quine can allow us coherently to challenge. Consider a sheet of paper appearing to carry the following inscription, for example:

The ink marks on this page should not taken to be letters
It is impossible both to treat this as a warning sign, and to pay heed it. To heed the warning would be to entertain the possibility that there is no warning to heed. I think this gives a sense of the flavour of the mistake that Carnap attributes to traditional metaphysics, in trying to stand both inside and outside the circle of particular linguistic frameworks. (Here, the tension comes from trying to stand both inside and outside the circle of language itself.)

4. When is the use of Concepts Ontologically Committed?

[33] I have suggested that Carnap's objection to traditional metaphysics turns on the following claims: (i) that we cannot address the question of the existence of something that answers to a concept without using that concept; (ii) that we cannot use a concept without taking on certain commitments -- roughly, those that locate the concept in our conceptual web; and (iii) that these commitments typically include a commitment to the existence of entities of the kind concerned -- in other words, a willingness to quantify over a domain of such objects. But is (iii) at all plausible? After all, it might be said, we often understand a concept without feeling committed to the existence of anything which answers to that concept. Take the concept dragon, for example. We know pretty much what would have to be the case for dragons to exist -- we understand the concept, in other words -- but we're certainly not committed to the view that there are such things. So when, if ever, does the use of a concept imply ontological commitment in the sense that Carnap requires?

[34] Carnap would say that we don't need to introduce a new quantifier to discuss the existence of dragons. We simply use an existing quantifier, ranging over creatures, say, or spatio-temporally located things. The existence of dragons is an internal issue, within such a framework. (Recall that Carnap says exactly this about unicorns.) The contrast is supposed to be with the kind of ontological commitment which comes with adopting the framework as a whole -- to spatio temporal objects in general, for example.

[35] Here's a homely example which seems to give the flavour of Carnap's view. Consider the coordinate framework of latitude and longitude, which we use to refer to positions on the Earth's surface. Working within this framework, we can ask such questions as, "Is there land or sea at position 154°E, 31°S?" This is a contingent matter, which we settle by empirical means. We can also ask "Is there such a position as 154°E, 31°S?", or "Is there an equator?", but -- taken internally -- these are analytic matters, settled by the rules of the framework. But Carnap wants to say that if we try to address such questions from an external point of view they no longer make sense, for the external perspective takes us outside the framework which gives content to expressions such as "154°E, 31°S" and "the equator". (From the external perspective we can ask about the pragmatic value of adopting the framework, but that's a different matter.)

[36] So when is the use of concepts ontologically committing, according to Carnap? Simply when the commitment flows from the rules of the framework, as a commitment to coordinate positions and lines of longitude and latitude does in this example. If we want to know what this amounts to in any particular case, we need to ask ourselves, in effect, what commitments we need take on, in order to be players in the linguistic game concerned. (There seems to be no hard and fast general principles here. The dispute with traditional metaphysics needs to be fought case by case.)

[37] Thus Carnap thinks that the external ontological stance is disallowed because if we step back too far, we step outside the relevant game altogether, and can no longer use the notions that have their home there. But how do we count linguistic games? In particular, what is to stop us treating all ontological issues as internal questions within a single grand framework? Why shouldn't we introduce a single existential quantifier, allowed to range over anything at all, and treat the question of the existence of numbers as on a par with that of the existence of dragons?

[38] These are exactly Quine's objections to Carnap's pluralism. Quine himself characterises Carnap's views as follows:

It begins to appear, then, that Carnap's dichotomy of questions of existence is a dichotomy between questions of the form "Are there so-and-so's?" where the so-and-so's purport to exhaust the range of a particular style of bound variables, and questions of the form "Are there so-and-so's?" where the so-and-so's do not purport to exhaust the range of a particular style of bound variables. Let me call the former questions category questions, and the latter ones subclass questions. I need this new terminology because Carnap's terms 'external' and 'internal' draw a somewhat different distinction which is derivative from the distinction between category questions and subclass questions. The external questions are the category questions conceived as propounded before the adoption of a given language; and they are, Carnap holds, properly to be construed as questions of the desirability of a given language form. The internal questions comprise the subclass questions and, in addition, the category questions when these are construed as treated within an adopted language as questions having trivially analytic or contradictory answers. (1966: 130)
Quine goes on to argue that
the question whether there are numbers will be a category question only with respect to languages which appropriate a separate style of variables for the exclusive purpose of referring to numbers. If our language refers to numbers through variables which also take classes other than numbers as values, then the question whether there are numbers becomes a subclass question, on a par with the question whether there are primes over a hundred....

Even the question whether there are classes, or whether there are physical objects, becomes a subclass question if our language uses a single style of variables to range over both sorts of entities. Whether the statement that there are physical objects and the statement that there are black swans should be put on the same side of the dichotomy, or on opposite sides, comes to depend on the rather trivial consideration of whether we use one style of variables or two for physical objects and classes. (1966: 131)

[39] In effect, Quine is arguing that there is no principled basis for Carnap's distinction of language into frameworks, where this is to be understood in terms of the introduction of new quantifiers, ranging over distinct domains of entities. If there is only one existential quantifier, ranging over entities of any kind, then there would appear to be nothing to whose existence we are necessarily committed by virtue of using a particular system of concepts. We can always step back, consider the broader range of entities, and ask ourselves whether anything within this range answers to the concepts in question.

[40] If Quine is right, then supposedly metaphysical issues -- "Are there numbers?", for example -- are indeed on a par with the ontological issues that Carnap wants to regard as internal. It is true that all ontological questions have a pragmatic ingredient, by Quine's lights, but this is no longer the comfort that it was before. At that stage, the point was that Quine's attack on the analytic-synthetic distinction seemed to worsen things for science, without improving things for metaphysics -- it didn't challenge the idea that metaphysics involves a linguistic mistake. But it now looks as though Carnap's main objection to metaphysics rests on an unsupported premise, namely the assumption that there is some sort of principled plurality in language which blocks Quine's move to homogenize the existential quantifier.

[41] Quine's view does not legitimize metaphysics entirely, of course. Quine agrees with Carnap that we can't stand outside language and consider ontological issues from an extra-linguistic standpoint. No doubt certain metaphysical views -- call them global externalist views -- are guilty of imagining that we can do that. However, Carnap seems to have thought that metaphysics makes this kind of mistake locally, with respect to different metaphysical topics -- numbers, physical objects, etc. -- and this is what Quine denies. Without Carnap's implicit pluralism, in other words, there is much more scope for interpreting actual metaphysical debates in a manner which renders them respectable, by Carnap's own lights -- for there is no such mistake as local externalism.

[42] Thus, as I noted at the beginning, there seem to be two strands to Carnap's attack on metaphysics. What the strands have in common is a principle we might call simply "internalism", which amounts to an aspect of what Quine later dubs "ontological relativity" -- simply the point that we can't address ontological issues except from the standpoint of some theory or other. Where the strands differ is in how they propose to count theories. For Quine there is only one, in effect, namely the current configuration of our web of belief as a whole. For Carnap there are several, however, because different parts of the web do different jobs, in some philosophically interesting sense. In other words, Carnap's weaker thesis is internalism simpliciter, and with this Quine agrees. His stronger thesis is the local internalism which flows from pluralism about theoretical functions, and with this Quine disagrees. The disagreement turns on the doctrine of functional pluralism.

[43] So far as I can see, however, Carnap himself does not have a satisfactory defence of this doctrine. In Quine's terms, he does not have any principled way to distinguish between category questions and subclass questions. What he needs, in effect, is an argument that there is some sort of category mistake involved in assimilating issues of the existence of numbers (say) and of the existence of physical objects. He takes for granted that this is so, and his model for the construction of languages reflects this assumption: roughly, speaking, the model requires that we mark the category boundaries in our choice of syntax -- a different quantifier for each category, for example. But he does little to defend the assumption that the boundaries are there to be marked, prior to our syntactical choices -- and this is what Quine denies.

5. Quine and Ryle

[44] The notion of a category mistake was familiar to the logical positivists of the 1920s and 1930s. In the Aufbau of 1928, Carnap himself uses the term "mixing of spheres" ("Sphärenvermengung") for, as he puts it later, "the neglect of distinctions in the logical types of various kinds of concepts" (Schilpp 1963: 45). For contemporary audiences, however, the term and the notion are particularly associated with Ryle. Ryle is quite clear that the notion has implications for ontological issues, and in a famous passage in The Concept of Mind he touches on the question as to whether existence is a univocal notion:

It is perfectly proper to say, in one logical tone of voice, that there exist minds, and to say, in another logical tone of voice, that there exist bodies. But these expressions do not indicate two different species of existence, for "existence" is not a generic word like "coloured" or "sexed". They indicate two different senses of "exist", somewhat as "rising" has different senses in "the tide is rising", "hopes are rising" and "the average age of death is rising". A man would be thought to be making a poor joke who said that three things are now rising, namely the tide, hopes and the average age of death. It would be just as good or bad a joke to say that there exist prime numbers and Wednesdays and public opinions and navies; or that there exist both minds and bodies. (1949: 23)
Given Quine's response to Carnap, it isn't surprising that he has little sympathy for Ryle's apparent ontological pluralism. In a section of Word and Object on the topic of ambiguity, Quine takes the opportunity to put on record his objection to Ryle's view:
There are philosophers who stoutly maintain that "true" said of logical or mathematical laws and "true" said of weather predictions or suspects' confessions are two uses of an ambiguous term "true". There are philosophers who stoutly maintain that "exists" said of numbers, classes and the like and "exists" said of material objects are two uses of an ambiguous term "exists". What mainly baffles me is the stoutness of their maintenance. What can they possibly count as evidence? Why not view "true" as unambiguous but very general, and recognize the difference between true logical laws and true confessions as a difference merely between logical laws and confessions? And correspondingly for existence? (1960: 131)
(The above passage from The Concept of Mind is one of two places to which Quine refers readers for "examples of what I am protesting." The other is a section of Russell's Problems of Philosophy in which Russell discusses the mode of existence peculiar to universals, distinguishing their "world of being" from "the world of existence" of spatio-temporal objects, and affirming that "both are real, and both are important to the metaphysician" (Russell 1966: 57). Nothing in what follows should be taken to be defending Russell's view!)

[45] But what is the disagreement between Quine and Ryle? For Quine, matters of ontology reduce to matters of quantification, and presumably Ryle would not deny that we should quantify over prime numbers, days of the week and dispositions. Indeed, Ryle might reinforce his own denial that there are "two species of existence" by agreeing with Quine that what is essential to the single species of existence is its link with quantification. Ryle simply needs to say that what we are doing in saying that beliefs exist is not what we are doing in saying that tables exist -- but that this difference rests on a difference in talk about tables and talk about beliefs, rather than on any difference in the notions of existence involved. So far this is exactly what Quine would have us say. The difference is that whereas Quine's formulation might lead us to focus on the issue of the difference between tables and beliefs per se, Ryle's functional orientation -- his attention to the question as to what a linguistic category does -- will instead lead us to focus on the difference between the function of the talk of beliefs and the function of the talk of tables.

[46] The right way to read Ryle seems to be something like this. Terms such as "exists" and "true" are not ambiguous, for they serve a single core purpose in their various different applications. In that sense, they are univocal but very general terms, as Quine himself suggests. In virtue of the pre-existing functional differences between the concepts with which they associate, however, the different applications of these terms are incommensurable, in an important sense. Many terms in language seem to fit this pattern, in having a single core meaning or function, with application in several quite distinct cases. (We might call the phenomenon "linguistic multifunctionalism".) A good example is the term Ryle himself offers by way of comparison with "exists", namely "rising". "Rising" certainly has a core meaning. It refers to the increase in some quantity over time. But in virtue of the incommensurability of different kinds of quantities, different risings may themselves be incommensurable. It doesn't make sense to ask whether the average age of death is rising faster than the cost of living, for example.

[47] Similarly with existence, Ryle appears to want to say. The term has a single core meaning or function, tied to that of the existential quantifier. But because the notions of mind and body "belong to different logical categories" -- in other words, as I would put it, have importantly different functions in language -- it doesn't make sense to think of the existence of minds as on a par with the existence of bodies. Ryle himself glosses this incommensurability in terms of the oddity of conjunctions such as "There are beliefs and there are rocks", but this doesn't seem to get to the heart of the matter. The crucial point is that attempts to make ontological comparisons between entities in the two domains go wrong in just the way that attempts to compare different kinds of risings go wrong. (In both cases it is arguable whether we should say that the comparisons are senseless, or merely false. I suspect that it makes little difference, as long as we recognise that even if we call it falsity, it involves a different kind of error from that involved in mistaken intra-category comparisons.)

[48] Of course, more needs to be said about the relevant notion of linguistic function. In some sense, talk of chairs serves a different function from talk of tables, simply because chairs and tables are different kinds of furniture. Yet Ryle (and I) don't want to say that "chair" and "table" belong to different logical categories. So we need a story about which functional differences are the important ones. What marks the joints in language between the logical categories?

[49] This is a crucial question, though not one which I can begin to address in this paper. For anyone attracted to this functionalist account, however, a good first bet is that some of the major joints in language correspond to the "hard problems" in contemporary metaphysics: morality, modality, meaning and the mental, for example -- what elsewhere (Price 1997) I call the "M worlds". This was Ryle's view, and I think that to a limited extent it was also Carnap's. (It was also Wittgenstein's, of course.) What is striking, from the point of view we imagined at the beginning of the paper, is how invisible this kind of view has become at the end of the century. Almost all the apparent positions in contemporary debates about these topics take for granted that the crucial issues are ontological: Are there really entities or facts of the kinds in question? If so, what are they? True, some people go on to ask a question which is about linguistic functions: If there are no such entities, what account can we give of the language which seems to refer to such things? But even here the linguistic point is subsidiary to the ontological point. It isn't Carnap's point, or Ryle's point, namely that the ontological question itself rests on a philosopher's confusion about language -- on a failure to notice the joints.

[50] Quine seems poorly placed to reject the suggestion that there might be important functional differences of this kind in language. The issue is one for science. It is the anthropologist, or perhaps the biologist, who asks, "What does this linguistic construction do for these people?" Quine can hardly argue that the results of such investigations may be known a priori.

[51] True, Quine himself often seems to take for granted that language has a well-defined core descriptive function, common to all well-founded assertoric discourse. This assumption underpins his claim that some apparently assertoric discourses -- those of intentional psychology or morality, for example -- do not serve this function, being rather expressive or instrumental. But as Chris Hookway (1988: 68-69) notes, it is far from clear that this assumption is defensible, in Quine's own terms. For example, given Quine's own minimalism about truth, it is no use his saying that descriptive discourse aims at truth. Why shouldn't a minimal notion of truth be useful in an expressive or instrumental discourse? In other words, why shouldn't a minimalist allow that truth itself is a multifunctional notion, in our earlier sense? And why shouldn't the notion of description be as minimal as that of truth -- thus undermining the assumption that description itself comprises a significant functional category? These are difficult matters, but that fact in itself supports the rather weak conclusion I want to draw. Quine's criticism of Carnap and Ryle's ontological pluralism is inconclusive, to say the least, because the issue depends on substantial issues about language on which the jury is still out.

[52] Perhaps it would be better to say that the jury has been disbanded, for contemporary philosophy seems to have forgotten the case. My conclusion is that there is no justification for this amnesia in Quine's response to Carnap and Ryle. We have seen that Quine agrees with Carnap in rejecting global externalism in metaphysics (and that Quine's appeal to the failure of the analytic-synthetic distinction is largely a red herring at this point). Carnap's claim that traditional metaphysics is also guilty of local externalism turns out to rest on foundations which Carnap himself does not supply -- in effect, functional foundations for Ryle's notion of a category mistake. Nothing in Quine's criticism of Carnap and Ryle's pluralism seems to count against the existence of such foundations, and so the verdict on the Carnap-Ryle view must await excavations -- first-order scientific enquiries into the underlying functions of language in human life. This importance of this kind of investigation is much less appreciated in contemporary philosophy than it was in the 1950s, and Quine perhaps deserves some of the blame.

Huw Price


I am most grateful for comments and suggestions from Chris Daly, Richard Holton, Rae Langton, Bernie Linsky, Stathis Psillos, Jack Smart and Ed Zalta, and an audience at ANU, Canberra in April, 1997.


Beth, E. (1963) "Carnap on Constructed Systems." In Schilpp (1963): 469-502.

Carnap, R. (1937) The Logical Syntax of Language. London: RKP.

----- (1952) "Empiricism, Semantics and Ontology." In L. Linsky (ed.), Semantics and the Philosophy of Language. Urbana: University of Illinois Press: 208-228. (Originally published in Revue Internationale de Philosophie 4 (1950): 20-40.)

Hookway, C. (1988) Quine. Stanford: Stanford University Press.

Quine, W. V. (1960) Word & Object. Cambridge, MA: MIT Press.

----- (1966) "On Carnap's Views on Ontology." In The Ways of Paradox and Other Essays. New York: Random House. (Originally published in Philosophical Studies 2 (1951).)

Price, H. (1997) "Naturalism and the Fate of the M-Worlds." Proceedings of the Aristotelian Society, Supp. Vol. LXXI.

Psillos, S. (1997 [unpublished]) "Carnap's Adventures in Theory-Land." Unpublished typescript.

Ryle, G. (1949) The Concept of Mind. London: Hutchinson.

Russell, B. (1967) The Problems of Philosophy. Oxford: Oxford University Press.

Schilpp, P. (ed.) (1963) The Philosophy of Rudolf Carnap. Library of Living Philosophers, Vol. XI. La Salle, IL: Open Court).

1997 Huw Price

EJAP is a non-profit entity produced at Indiana University. EJAP may not be printed, forwarded, or otherwise distributed for any reasons other than personal use. Do not post EJAP on public bulletin boards of any kind; use pointers and links instead. EJAP retains the right to continuously print all the articles it accepts, but all other rights return to the author upon publication.

EJAP, Philosophy Department, Sycamore Hall 026, Indiana University, Bloomington, IN 47405.