Sometimes I wonder why nobody reads philosophy. It requires, to be sure, a degree of hyperbole to wonder this. Academics like me, who eke out their sustenance by writing and teaching the stuff, still browse in the journals; it’s mainly the laity that seems to have lost interest. And it’s mostly Anglophone analytic philosophy that it has lost interest in. As far as I can tell, ‘Continental’ philosophers (Derrida, Foucault, Habermas, Heidegger, Husserl, Kierkegaard, Sartre and the rest) continue to hold their market. Even Hegel has a vogue from time to time, though he is famous for being impossible to read. All this strikes me anew whenever I visit a bookstore. The place on the shelf where my stuff would be if they had it (but they don’t) is just to the left of Foucault, of which there is always yards and yards. I’m huffy about that; I wish I had his royalties.
Royalties aside, what have they got that we haven’t? It’s not the texture of their prose I shouldn’t think, since most of us write better than most of them. (I don’t include Kierkegaard. He was a master and way out of the league that the rest of us play in.) Similarly, though many of the questions that Continental philosophy discusses are recognisably continuous with ones that philosophers have always cared about, so too, by and large, are many of the questions that we work on. For example, Kripke’s metaphysical essentialism (of which more presently) has striking affinities with the metaphysical Realism of Aristotle and Augustine. True, we sometimes presuppose more logic than you’re likely to come across on the omnibus to Clapham. But I’m told that an intelligent reading of Heidegger requires knowing more about Kant, Hegel and the Pre-Socratics than I, for one, am eager to learn. Anyhow, our arguments are better than theirs. So sometimes I wonder why nobody (except philosophers) reads (Anglophone, analytic) philosophy these days.
But, having just worked through Christopher Hughes’s Kripke: Names, Necessity and Identity, I am no longer puzzled. That may sound as though I’m intending to dispraise the book, but to the contrary; I think it’s a fine piece of work in lots of ways. To begin with, the topic is well chosen. By pretty general consent, Kripke’s writings (including, especially, Naming and Necessity) have had more influence on philosophy in the US and the UK than any others since the death of Wittgenstein. Ask an expert whether there have been any philosophical geniuses in the last while, and you’ll find that Kripke and Wittgenstein are the only candidates. Again, as far as I can tell, Hughes’s exposition is accurate and sophisticated, and his coverage is more than adequate. Unless you are yourself a practitioner, this is all the Kripke that you need to know about. The book has, as Hughes admits, a certain unevenness of tone: some chapters are clearly introductory, others are polemical. But there’s nothing that a competent graduate student won’t be able to cope with; and skimming will give a sense of the geography for those who don’t care about details.
And yet I can’t shake off the sense that something has gone awfully wrong. Not so much with Hughes’s book (though I’ll presently have bones to pick with some of his main theses) as with the kind of philosophy that has recently taken shelter under Kripke’s wing. There seems to be, to put it bluntly, a lot of earnest discussion of questions that strike my ear as frivolous. For example: ‘I have never crossed the Himalayas, though I might have done. So there is a non-actual (or, if you prefer, a non-actualised) possible world (or possible state of the world) in which someone crosses some mountains. Is that person me, and are those mountains the Himalayas? Or are they (non-actual) individuals different from me and from the Himalayas?’ Or: ‘Water is the stuff that is in the Thames and comes out of the taps. The stuff that is in the Thames and comes out of the taps undeniably contains impurities (bits that are neither hydrogen nor oxygen nor constituents thereof). So how can water be H2O?’ But how could it not? Is it that, chemistry having discovered the nature of water, philosophy proposes to undiscover it? In any case, could that really be the sort of thing that philosophy is about? Is that a way for grown-ups to spend their time? A brief sketch of how we got into this, and of Kripke’s role in getting us here, is the burden of what follows. I offer a very condensed account of changes, over the last fifty years or so, in the way that analytic philosophers have explained to one another what it is that they’re up to. It is, however, less historical than mythopoetic. The details aren’t awfully reliable, but maybe the moral will edify.
Stage one: conceptual analysis. A revisionist account of the philosophical enterprise came into fashion just after World War Two. Whereas it used to be said that philosophy is about, for example, Goodness or Existence or Reality or How the Mind Works, or whether there is a Cat on the Mat, it appears, in retrospect, that that was just a loose way of talking. Strictly speaking, philosophy consists (or consists largely, or ought to consist largely) of the analysis of our concepts and/or of the analysis of the ‘ordinary language’ locutions that we use to express them. It’s not the Good, the True or the Beautiful that a philosopher tries to understand, it’s the corresponding concepts of ‘good’ ‘beautiful’ and ‘true’.
This way of seeing things has tactical advantages. Being good is hard; few achieve it. But practically everybody has some grasp of the concept ‘good’, so practically everybody knows as much as he needs to start on its analysis. Scientists, historians and the like need to muck around in libraries and laboratories to achieve their results, but concepts can be analysed in the armchair. Better still, the conceptual truths philosophy delivers are ‘a priori’ because grasp of a concept is all that’s required for their recognition. Better still, whereas the findings of historians and scientists are always revisable in principle, it’s plausible that the truths conceptual analysis reveals are necessary. If you want to know how long the reign of George V lasted, you will probably need to look it up, and you’re always in jeopardy of your sources being unreliable. (I’m told he reigned from 1910-36, but I wouldn’t bet the farm.) But the philosopher’s proposition that a reign must last some amount of time or other would seem to be a conceptual truth; being extended in time belongs to the concept of a reign. Historians might conceivably find out that George V reigned from, say, 1910-37. That would no doubt surprise them, but evidence might turn up that can’t be gainsaid. Philosophy, however, knows beyond the possibility of doubt – beyond, indeed, the possibility of coherent denial – that if George V reigned at all, then he reigned for a while. The truths that conceptual analysis arrives at are thus apodictic, rather like the truths of geometry. Such a comfort. Ever since Plato, philosophers have envied geometers their certitudes. So it’s not surprising that the story about philosophy being conceptual analysis was well received all the way from Oxford to Berkeley, with many intermediate stops.
Still, there was felt to be trouble pretty early on. For one thing, no concepts ever actually did get analysed, however hard philosophers tried. (Early in the century there was detectable optimism about the prospects for analysing ‘the’, but it faded). Worse, the arguments that analytic philosophers produced were often inadvertently hilarious. Examples are legion and some of them are legendary. Here are just two that will, I hope, suffice to give the feel of the thing. (Truly, I didn’t make up either of them. The second comes from Hughes, and I’ve heard the first attributed to an otherwise perfectly respectable philosopher whose name charity forbids me to disclose.) First argument: the issue is whether there is survival after death, and the argument purports to show that there can’t be. ‘Suppose an airplane carrying ten passengers crashes and that seven of the ten die. Then what we would say is that three passengers survived, not that ten passengers survived. QED.’ Second argument: the issue is whether people are identical with their bodies. ‘Suppose you live with Bob . . . who went into a coma on Wednesday . . . Suppose that a friend calls on Thursday and says: "I need to talk to Bob: is he still in England?” You might naturally answer: "Yes, but he’s in a coma.” Now fill in the story as before, but suppose that Bob had died. When the friend says "I need to talk to Bob: is he still in England?” would you really answer, "Yes, but he’s dead,” even if you knew that Bob’s (dead) body still exists and is still in England?’ Presumably not, so QED once again. Now, I don’t myself believe that there is survival after death; nor do I believe that persons are identical with their bodies. But, either way, these arguments strike me as risible; dialectics dissolves in giggles. If, as would appear, the view that philosophy is conceptual analysis sanctions this sort of carrying on, there must surely be something wrong with the view. So much for stage one.
Stage two: Quine. In 1953, W.V. Quine published an article called ‘Two Dogmas of Empiricism’. Easily the most influential paper of the generation, its reverberations continue to be felt whenever philosophers discuss the nature of their enterprise. In a nutshell, Quine argued that there is no (intelligible, unquestion-begging) distinction between ‘analytic’ (linguistic/conceptual) truth and truth about matters of fact (synthetic/contingent truth). In particular, there are no a priori, necessary propositions (except, perhaps, for those of logic and mathematics). Quine’s target was mainly the empiricist tradition in epistemology, but his conclusions were patently germane to the agenda of analytical philosophy. If there are no conceptual truths, there are no conceptual analyses either. If there are no conceptual analyses, analytic philosophers are in jeopardy of methodological unemployment.
Whether Quine was right remains the bone of vigorous philosophical contention to this day. In fact, despite their extensive influence, there isn’t any robust consensus as to what, exactly, the persuasive arguments in ‘Two Dogmas’ are or were supposed to be. (Philosophy is like that.) Suffice it that, since Quine, the practice of conceptual analysis has lacked a fully credible rationale. That’s not to say that anybody much stopped doing it. To the contrary, it’s often suggested that Quine must have been wrong because conceptual analysis is what analytic philosophers do, and there must be something that they’re doing when they do it. That put a brave face on it, but there were guilty consciences wherever you looked. And so things stood for several decades.
Stage three: Kripke. The way to save analytic philosophy from the embarrassments Quine raised is to construct some proprietary notion of necessity that doesn’t presuppose the notion of conceptual truth. It would be nice if there were some close connection between necessity, so construed, and a priority. That would explain how such necessities could be accessible with the epistemic equipment professional philosophers have available: some common sense and three or four years of postgraduate study. We arrive at the crux. Kripke has made major contributions to several areas of philosophy, including the interpretation of modal logic. (Don’t ask.) He has radically revised once standard views about the semantics of proper names. He has sponsored a revival of an essentialist programme in metaphysics that traces back to Aristotle and the Scholastics. There’s more, but it’s all fairly technical stuff; if you’re interested, you’ll find readable expositions in Hughes.
The point for present purposes is that Kripke can be read as having provided the very notion of necessity that the vindication of analytical practice required, thereby saving analytic philosophers from Quine. That is, in fact, pretty much the way that Hughes reads him.
For much of the first half of the 20th century, modality [i.e. necessity] had a somewhat marginal place in analytic philosophy. Kripke contributed more to its ‘demarginalisation’ than any other analytic philosopher. He did this by . . . vigorously and effectively addressing Quinean worries . . . and by bringing modal issues into various central debates in philosophy . . . The ‘remodalisation’ of metaphysics and the philosophy of language may retrospectively come to be thought of as Kripke’s most important contribution to 20th-century philosophy. Those of us who, as undergraduates, learned philosophy from Quineans think of Kripke as a philosopher who (almost single-handedly) transformed the philosophical landscape.
I think that’s right, but with a caveat. It’s not that pre-Kripkean analytic philosophy marginalised modality. Rather, it took for granted that necessary propositions arise from the analysis of concepts (or words, or both). That was the view that Quine seemed to have undermined, thus leaving analytic philosophers with two unsatisfactory choices: give up on analysis, or proceed without a credible account of their methodology. Kripke seemed to relieve them of this dilemma. No wonder analytic philosophy fell in love with him.
Here’s the basic idea. One drops the traditional thesis that necessary propositions are linguistic or conceptual, and one substitutes a metaphysical account of necessity. Philosophy is to recognise not just the actual world that we live in but also a plethora of ‘possible worlds’. The actual world is itself possible, of course; but so, too, is the world that’s just like this one except that Mr James (a domestic feline who’s currently having a nap) is awake and chasing mice. Similarly, there are worlds that are just like ours except that there’s nobody in them, and worlds just like ours except that everybody is in them except President Bush. Likewise there are (brave, new) worlds in which I get Foucault’s royalties and he gets mine. And so on. Notice, however, that there is no (possible) world in which 2+2=5; and none in which bachelors are married; and none in which George V reigned, but for less than a while. So, given this new ontology, we can identify necessarily true propositions with the ones that are true in every possible world, necessarily false propositions with the ones that are false in every possible world, and contingent propositions with the ones that are true in some possible worlds but not in all. Here we seem to have a nonconceptual notion of necessity. Whereas analytic philosophy used to be seen as tracing relations among concepts, it is now seen as tracing relations among possible worlds.
A quick example will show how this is supposed to work. Some years ago, Hilary Putnam raised the following question, which analytic philosophy has been gnawing at ever since. Suppose somebody discovered a sort of stuff that is, to casual inspection, just like water (it’s wet, it’s clear and potable, it freezes at zero centigrade, has specific gravity 1, dissolves sugar, puts out fires and so forth) but the molecules of which have some chemical structure other than H2O (‘XYZ’ by convention). You are now invited to consult your intuitions: is XYZ water? If not, why not? The canonical intuition is that XYZ isn’t water because being made of H2O is an essential property of water; whatever is a sample of water is ipso facto a sample of H2O, and nothing else could be. (It’s an epistemological worry for essentialists that not everybody has the canonical intuition; in fact, some people don’t have it quite vociferously, and perhaps they’re right not to. But it would ease the exposition if you will kindly agree to ignore that. You can always change your mind about it later.)
Interesting things follow if the intuition is granted; including, in particular, interesting modal things. For example, if it’s right that nothing but H2O would count as water, then water is H2O in every possible world (more precisely, in every possible world where there is any). That is, given the modal intuitions, it’s necessary that all and only water is H2O according to the metaphysical construal of necessity. Note further that this necessary truth is available a priori; at no point in the course of its discovery did philosophy stir from the armchair in which we found it. A little caution is, however, required here. What’s a priori is the hypothetical proposition: ‘If samples of water are samples of H2O, and nothing else is, then it’s necessary that water is H2O.’ By contrast, it isn’t a priori that samples of water are samples of H2O; to the contrary, that’s just the sort of grimy empirical generalisation that chemists discover inductively in their laboratories, to the accompaniment of bangs and stinks. A gratifying division of labour is thus perceptible: the chemists do the heavy lifting and the philosophers do the heavy thinking. It’s clear from the empirical research that water is H2O in every possible world that is compatible with chemistry. What remains for philosophers to determine is whether water is H2O in every possible world tout court. Presumably it’s our modal intuitions that decide this if anything does; they would seem to be all there is that’s left unaccounted for by the time the chemists finish their investigations. It’s therefore unsurprising that, in practice, analytic philosophers take it for granted that modal intuitions aren’t fallible.
This story ramifies in all sorts of directions; Hughes will fill you in. Once again, suffice it for our purposes to consider just the methodological implications. The situation pre-Kripke was that philosophers were supposed to disclose necessary, a priori truths that they arrived at by analysing words or concepts. Quine’s attack seemed to put this project in jeopardy. If there are no conceptual truths, then, a fortiori, there are no conceptual truths for philosophy to deliver. But now it appears that Kripke has saved the bacon since there are, in any case, plenty of metaphysical necessities. And, as we’ve seen, metaphysical necessities can be discovered a priori by examining philosophically relevant intuitions. These are not, however, intuitions about relations among concepts: they’re modal intuitions about what’s possible and what isn’t. In effect, analytic philosophy was doing the right sort of thing (viz, analysis) but for the wrong sort of reasons. That being straightened out, the pangs of conscience can now be soothed and everybody can go back to doing what he learned to do in graduate school. General rejoicing in the philosophical community. Plus or minus a bit, this is how Hughes sees the current methodological situation. I think that it’s probably the majority view.
But I doubt that it can be sustained. In this respect, the significance of Kripke’s work has, I think, been much overestimated. If analytic philosophy had methodological problems pre-Kripke, it continues to have the very same problems, and for the very same reasons. Something about that to conclude.
A kind of question that doesn’t get asked often enough is: what are modal intuitions intuitions of? Consider, for example, the intuition that water is necessarily H2O. How do things have to be for it to be right? Or wrong? What’s its ‘truth maker’, to use the philosophical jargon? An answer springs to mind in light of the previous discussion, but it doesn’t survive reflection: ‘For water to be necessarily H2O is just for water to be H2O in every possible world. For water not to be necessarily H2O is just for there to be possible worlds in which there’s H2O but no water (or water but no H2O). That all follows from Kripke’s account of necessity and is unproblematic. So there’s nothing to worry about.’ I guess that’s alright as far as it goes; it is, as remarked, just a consequence of defining ‘necessarily true’ as ‘true in all possible worlds’.
But the question I was trying to raise wasn’t: ‘What about possible worlds makes it necessary that water is H2O?’ My question was: ‘What about water makes it necessary that water is H2O’? There must be something about water that does because, notice, there are plenty of kinds of stuff for which the corresponding modal claim would be false. For example, there’s Coca Cola; Coke behaves quite differently from water in modal contexts. Suppose XYZ is the formula for Coke (I’m told they keep one in a vault in Atlanta). So, every (actual) sample of Coke is a sample of XYZ and vice versa. It doesn’t follow that ‘Coke is XYZ’ is true in every possible world. To the contrary, the Coke people could change the recipe tomorrow if they wished to and, no doubt, there are possible worlds in which they do. The new stuff will still be Coke if they say it is. Likewise, mutatis mutandis, for smog. Every sample of smog is a sample of CO2 and god knows what else; but that’s only contingently true. Perhaps tomorrow they’ll find a way to pollute the air by using XYZ. Then, ceteris paribus (according to my modal intuitions), the right story would be that they’ve found a new way to make smog, not that they’ve found a way to make something that seems just like smog but isn’t.
So then, what’s the actual difference between water, on the one hand, and Coke and smog, on the other, that accounts for these modal differences? I can only think of one answer: if water is actually H2O, then ‘water is necessarily H2O’ is some kind of conceptual truth. The idea (endorsed in one form or other by many analytic philosophers) is that ‘water’ is the concept of a ‘material kind’. What’s special about material kinds is that what possible things of that kind there are depends on what actual things of that kind there are. In effect, the kind is defined by reference to its actual instances. So, water is a material kind because every sample is ipso facto required to have the same microstructure that actual samples do. It follows that, if water is H2O in this world, it’s H2O in every possible world. It also follows that samples of XYZ couldn’t be water samples even if they seemed to be. Compare smog. What possible samples of smog have in common with actual samples isn’t what they are (would be) made of but rather the way they (would) affect your eyes, nose, throat and view. In short, if K is the concept of a material kind, and if every actual thing that K applies to is made of n-stuff, then it’s necessary that every thing that K (would) apply to is made of n-stuff. As far as I can make out, this is more or less the view that Hughes himself holds. He says: ‘If it should turn out that only philosophers baulk at classifying XYZ as water, I am ready to defer in my usage to the non-philosophical majority and say that "water", like "glue", is not the name of a kind with a chemical essence.’ I guess what’s going on is that, because he thinks Kripke refuted Quine, Hughes feels free to treat the modal status of ‘water is H2O’ as linguistically (or conceptually) determined. So it is, after all, our grasp of concepts (or our mastery of language) that underwrites the modal intuition that ‘water is H2O’ is necessary. It’s just like the old days, really.
It’s past time to draw the moral, which I take to be that a plethora of claims to the contrary notwithstanding, you can’t escape Quine’s web just by opting for a metaphysical notion of necessity. Not, anyhow, if the latter is grounded in intuitions about what possible worlds there are. That’s because some story is needed about what makes such intuitions true (or false) and, as far as I can see, the only candidates are facts about concepts. It’s ‘water’ being a material kind concept that vindicates the intuition that water is necessarily H2O. Well, but if Quine is right and there aren’t any such facts about concepts, then there is nothing to vindicate modal intuitions. Accordingly, if the methodology of analytic philosophy lacked a rationale pre-Kripke, it continues to do so.