On the weightlessness of reality
Gary Kemp interviewed by Richard Marshall.
Gary Kemp finds Quine and Davidson awesome and has edgy thoughts about them all the time. He thinks Frege is more Newton than Einstein and refines him. Aesthetics isn’t his primary thing but he’s always interested. He keeps reading Proust and doesn’t think Beckett is a window-dresser. He thinks Quine thinks there’s no issue about realism – which is neither a realist nor an anti-realist position. He is thus sensationally groovacious.
3:AM: What made you become a philosopher? Has it been rewarding so far?
Gary Kemp: I think like many people I became what I am before I had any real sense of choosing. At university – the University of Oregon – I was fleeing from my dad (he was an eminent astrophysicist), playing the guitar and being interested in English literature, when I had my first philosophy class, and was hooked. Or second, the first was as a freshman and I had no sense of anything at first (if ever). I don’t know if it’s been rewarding; what would I compare it with? I do feel very fortunate to have a job as an academic philosopher.
3:AM: You’ve written about Frege and his conception of truth. Before we go into details of what you argue, could you say a little about Frege for us here at 3:AM to give us the context? Frege is a hugely important figure but isn’t as well known as, say, Einstein in science, but what he achieved was arguably as important and impressive. So could you say what is at stake in Frege’s work?
GK: As many people have pointed out, Frege was the first to develop rigorously what has since been called a theory of meaning, and truth sits at the centre of his model. And why is a theory of meaning so important? Ultimately I think it isn’t so important, but in certain moods it is not hard to accept that the idea, as has been made very explicit by Michael Dummett, is more fundamental than epistemology or metaphysics. I would say that Frege is more properly compared with Newton than with Einstein, if you had to choose.
3:AM: I think given that introduction it’ll be possible to see why you are interested in making sure we have his notion of the function of truth, for instance, correct. It can seem that philosophers are just being unduly picky and interested in technical details, but getting these details right is very important isn’t it? So, what do you say about Frege?
GK: Well, I don’t think it so important that we have his notion of truth correct. I don’t think there is such a thing as Frege’s exact theory of truth. People were writing about truth in Frege’s time, of course, but there is little evidence that he read many of them, and not with much patience, partly because these thinkers weren’t aware of the kind of role that it played in a precise logic such as his. And logic is, of course, the laws of truth. He was thinking about certain problems pretty much on his own. What he wrote pulls in different directions, but he did have the idea of what is now called the ‘redundancy’ theory of truth – that ascribing truth to a proposition or thought accomplishes no more than asserting the proposition or thought.
It is a diabolically tempting idea, but it’s hard to square with the notion that truth is fundamental to logic and a theory of meaning. But this idea is not worked out at all and it isn’t clear that he actually accepted it; he just was aware of it. It was for people after him – starting with Ramsey and Ayer – to begin to explore the ramifications. Nowadays a subtle variant of it remains popular under the name of truth-deflationism; but on the other hand Davidson’s theory, which has truth as primitive, also has its roots in Frege (sense or meaning being in some sense equated with truth-conditions).
3:AM: Do you consider yourself to have been refining Frege, or was your approach back then in the mid-90s more a revisionary project? Was Dummett, for example, perhaps the top Frege man in the world (and sadly now recently no longer with us) on board with what you had to say?
GK: Not revisionary; I’m not as ambitious as that. Probably a case of refinement (but Frege is much too clear to admit of fantastic new readings); more just thinking further with the ideas he left behind. Off the top of my head, I could name Peter Sullivan, Thomas Ricketts and Peter Carruthers as having explored the angles I explored. My contribution, if contribution it was, was to probe further Frege’s hints of the connection of truth with assertoric force, of truth with Frege’s characterisation of sentential meaning as the truth-conditions of a sentence, and of implications of this for Frege’s estimate of the roles and statuses of logic and semantics.
Dummett of course rejected Frege’s (and Davidson’s) use of truth in favour of a sophisticated form of verification. But to speak Dummett-speak, and very roughly, I suppose the idea that the assertion of a proposition or thought is the advancement of it not as true, but as capable of being verified (in principle), is not far off Dummett’s actual view.
3:AM: I think that one of the interests was in how we were to understand claims of meaning and truth and often this discussion has been cast as one done within the context of science. But you have been interested in raising the issue not just there but also in the realm of aesthetics. So before looking specifically at your work on Quine and Davidson, can you say something about your arguments in this realm? You’ve written about the ‘aesthetic attitude’. What is that, and how does it relate to you broader philosophical interests?
GK: The ‘Aesthetic Attitude’ was just a self-contained piece; I simply thought that George Dickie, famous though he is, was just wrong in his attempt to show the notion of an aesthetic attitude to be empty. Aesthetics generally has always been near the heart of my intellectual interests; in fact the second philosophy class I had at the University of Oregon was one on aesthetics, with Garry Hagberg – a fine aesthetician and jazz guitarist – not teaching at that time but a teaching assistant. It’s never been my primary thing but I’ve I maintained a keen interest in it. Like many I can’t get over the feeling that most modern aesthetics does not get at the heart of our interest in art, but I cannot articulate what is missing.
3:AM: You discuss Croce and Collingwood who are best known for their philosophy of art. What do you take from them that is valuable and do you think they still relevant? Does what they say connect with what you have to say about the relationship between beauty and language?
GK: Croce and Collingwood make the bracing claim that art just is expression. I think that that project ultimately fails, but I do think they have answers to the simple objections that might be raised, and are much closer to something like a viable theory of art than is commonly recognised (not of the concept of art, but of art); and I like the fact that they would not have any time for conceptual art.
I don’t want to take the time to explain, but the point at which they fail is that even with an extended conception of expression, it’s hard to equate expressive value with aesthetic value. This does connect with what I say about beauty: ‘beauty’ is just a word of our language, for which there is no simple definition, no necessary and sufficient conditions. It’s a word whose rhetorical power outleaps its logic; there is no essence of beauty. The same is true of aesthetic value. This isn’t obscurantism, it’s just the rather pedestrian point that these words are family resemblance terms, a point from Wittgenstein.
3:AM: You ask the cool question, ‘Are there philosophical tasks that can be served only by fiction, or which are best served by it? Many philosophers, in various ways, have thought so (Proust himself puzzled over this in what eventually became Remembrance of Things past which began as mere illustrations of a theory of literature).’ In admiring Landy’s book on Proust you called out much writing on Proust as ‘mere rhetoric or window-dressing’. This was bold given that no less than Beckett has written on Proust. You also claim that Landy’s book is a model of how fiction might be explored philosophically. So could you say what Landy did there that was so admirable and useful, perhaps with an example or two. And what’s the answer you give to the question?
GK: Naturally I exclude Beckett from the window dressers. Landy presents a compelling theory of the self that is implicitly at work in Proust, a theory that is breathtakingly lucid and goes some way towards cracking the central riddle of the big book, namely of whether and how the madeleine works, and of the importance it has. I do think that much of what we take ourselves to be is of the same status as a work of literature, and to make that really and theoretically plausible – it goes against a lot of philosophy as well as common sense – is I take it the chief achievement of Landy’s book. It’s not a new idea by any means, but it is extraordinary to show how a worked-out theory applies to Proust’s book – a theory that the book exemplifies self-consciously and reflexively: the book serves as empirical support for its own thesis. He’s got a new one that I’m anxious to read.
3:AM: In the last six or seven years you’ve turned back to Frege, Quine and Davidson. (I’m sure you never turned away, but your writings have been more specifically about these guys than in the previous five or so years.) One is a great excuse for me to sound deep, obscure and dead smart all at once (without being any of these things) so for fun I’m going to go continental and ask: ‘Why is Axiom V important to Julius Caesar, and is its difference from HP dialectical?’
GK: I don’t know if I can answer briefly but informatively, and without boring you, but the line I took in the paper to which you refer boils down to Frege’s having distinguished axioms sharply from definitions. The goal is to show that arithmetic is ultimately reducible to logic, the laws of truth.
Axiom V – the principle that set identity is equivalent to having the same members – is much stronger (it entails much more, in fact too much!) than the more modest HP – Hume’s principle, that the number of Fs is the number of Gs if and only if the Fs admit of one to one correlation to the Gs. HP is strong enough for arithmetic (see Richard Heck or George Boolos), so the question arises of why Frege stuck with Axiom V, especially as it turned out to be inconsistent (unlike HP).
Well, Frege held definitions to be explicit definitions, like ‘bachelors are unmarried men'; they are practical devices, in principle if not in practice superfluous. (He may have came to this view while writing the Foundations of Arithmetic). Frege ultimately did not recognise any other sort of definitions, and HP is not of the right form to serve as an explicit definition. Axioms, meanwhile, are simply not in the business of showing for instance that Julius Caesar is not a number, no more so than it’s up to the principles of thermodynamics to entail that the number two has no heat. Principles or axioms don’t entail all truths regarding the concepts they govern. That’s why Frege could rest with Axiom V, and with not explaining why Julius Caesar is not a set.
Why then didn’t Frege simply count HP as an axiom? As far as I can see, the Julius Caesar problem is, in the end, a red herring. The reason that HP couldn’t be an axiom is just that HP is narrower in scope than Axiom V; therefore less general, less fundamental, and less likely to persuade one that arithmetic is really logic (for example, what sorts of things are these numbers of which you speak?; whereas – remember this is a pre-Russell world – one would have been less likely to ask the same about sets, that is extensions of concepts).
And with Axiom V, HP is derivable (setting aside the inconsistency of the former). Another is that he was going to need Axiom V anyway to develop a system of real numbers. Crispin Wright and Bob Hale have done a lot to show that such considerations can be got round.
3:AM: I was being a bit daft in that last question, but there was a serious underbelly to it. You and Christopher Belshaw edited a book in 2009 which looked at twelve modern philosophers: Quine, Rawls, Davidson, Williams, Rorty, Fodor, Nagel, Kripke, Nozick, Parfit, McDowell and Singer. You don’t actually say it, but it’s a kind of greatest hits list of the modern philosophical scene. We might quibble about one or two entries – I might wonder about Rorty and ask where is Dummett, say – but the list could be extended and different people will have slight differences. But on the whole it seems a pretty good go. What strikes me is that these are all philosophers working or who were working in Anglo-American universities. The so-called continentals are nowhere. No Ranciere. No Badiou, no Levinas. Does this reflect your view that these philosophers are just not as important as many make out. And anyway, is the idea of ‘continental philosophy’ useful anymore?
GK: Well it isn’t the greatest hits of recent philosophy! More guardedly, it is a representative sample of some of the greatest hits for a semi-popular audience; for example I agree that Dummett was a better philosopher than Rorty, but Rorty gets in because we needed some variety; David Lewis didn’t get in either – he’s great but a philosopher’s philosopher. But I totally agree that such things cannot escape arbitrariness and whim.
About continentals: it didn’t used to be case, and it perhaps won’t be very shortly, but from say the outbreak of WWII to about now the best philosophy was done in English – with certain big exceptions, notably Wittgenstein’s Philosophical Investigations. There are various causes of this but the biggest cause was that most philosophers on the continent fled the Nazis to the United States, if they were not killed by them. As for the philosophers you mention – along with figures such as Derrida, Foucault – Belshaw and I find some of what do very interesting, but not anything like of the same kidney as the philosophers in the book; either it’s not what we were taught to call philosophy, or it’s philosophy not of the highest quality. Jurgen Habermas is another you might’ve mentioned, who comes closer to making it into the book.
3:AM: You have written admiringly of Quine who you call ‘the leading exponent of naturalism of the twentieth century’. Before discussing some of your views about Quine in detail, can you briefly introduce him to a non-specialist audience?
GK: Burton Dreben said something like the great lesson of Quine is that there is no understanding him without understanding all of him. Well, despite that I’ll try to give a few words that might not be too misleading. The main point is that for Quine there is no ‘first philosophy’ – knowledge as a whole including philosophy is the product of a team (not its owners, and definitely not of the one-percent) – it is not hierarchical or foundational but is structured holistically, with philosophy attending to the more abstract, systematic or problematic bits, but incapable of branching off on its own. A favourite image Quine got from Otto Neurath is as perfect as such things can be: we are all sailing the same ship, which we must repair without docking, without sinking. Not unconnected with this is Quine’s naturalism: knowledge is what is delivered in the first instance by natural science, with its various norms, strictures and presuppositions. Which is not, I hasten to add, something known by obscure a priori means; it’s a fallible claim derived from reflection on knowledge as we have it. And did I say that Quine is an empiricist? He is.
3:AM: You point out that Quine changed his mind about what he said in ‘Two Dogmas of Empiricism’ where he said that a sentence may be held to be true come what may. The original claim depended on a radical holism about meaning. So is it wrong to think of Quine as being a radical holist?
GK: Later on, Quine said that radical holism is true but only legalistically – in fact there is no appreciable chance that simple sentences of arithmetic might be revoked; the conceptual entanglement of the concepts they involve is too great. By the way, I don’t think Quine was a holist about meaning; only about epistemology, confirmation. Not because he was an atomist or whatever about meaning – he was an eliminativist about meaning; language is to be theorised about without using such a notion. I’m one of the few who thinks he was right on this.
3:AM: You say that he also back-tracked on his anti-analyticity view too. I thought he was the guy who broke the analytic/synthetic distinction, but you say he ended up thinking the distinction was a useful one. Is that right?
GK: Not quite. He saw how something can be defined in terms that are acceptable to him that accounts for people’s sense that sentences like ‘All and only mares are female horses’ is trivial, empty: a sentence has this character if one acquires mastery of a word by learning to assent to such a sentence. So one learns ‘Mare’ by learning the truth of the sentence. That is an extremely rough notion, but then so is ‘True by virtue of meaning’. The point is that such ‘analyticity’ is not of much epistemological significance; whereas Carnap, Ayer etc. conceived of analyticity as the glue that ties evidence to theory, as fundamental to knowledge, not as a rag-bag of inessential though inevitable spandrels. So analyticity in this sense is real but not a supporting member of any general theory. True, he also said that ‘obvious, or reachable by obvious steps of reasoning’ might serve for analyticity as applied to logic, but that surely doesn’t get one very far. He’s just trying account for intuition of analyticity.
3:AM: You say that it was his developed views about naturalism that pulled him away from his original iconoclastic positions don’t you? Can you say something about this? Can you explain how his Darwinism changed his position? I think you say it becomes stark when considering his definition of ‘observation sentences’ in his Sixties work Word and Object but he came to adjust his position thirty years later in ‘Darwins solvent of metaphysics’. You summarise the situation ‘… the last significant manoeuver of the leading exponent of naturalism of the twentieth century involves a straightforward appeal to considerations made standard by the leading naturalist of the nineteenth, Charles Darwin.’ Can you say what this involved?
GK: Not sure they pulled him towards an embrace of the icons. In the 1950s, as he came to know more about science generally – and wrote about it – he simply became more self-consciously confident, I think, that his view really added up to something in the general philosophy of science. I think that few people quite grasp the message of Word and Object, that one can account for human knowledge without all this baggage that philosophers insist on. The centrepiece is of course (his version of) naturalism as you appreciate. Darwin entered into his thought in a big way in his answer to Goodman in ‘Natural Kinds‘: the reason that past evidence is evidence that all emeralds are green and not grue is that we have been conditioned over the eons by natural selection to expect ‘green’ rather than ‘grue’, despite the two expectations’ being logically on a par with respect to the evidence. It’s the same answer he gave to Hume’s problem of induction; it just an explanation of why we reason as we do, and a rather shallow one at that. What neither is, is anything like an a priori or god forbid a transcendental justification. He thinks that that is hopeless; the principle of induction simply has to be assumed at every turn in science. But with holism, that isn’t a problem; in fact the P of I is shown to have the strongest support, and in that sense a (circular) a posteriori justification.
Your question about observation sentences is important, but it’s hard to answer non-tediously. The problem was how to reconcile the fact that observation sentences are neurologically private – in that there no reason to suppose that people’s neurological receptors are similar – with the demand that observation sentences be intersubjective (for that is the key to communication and objectivity). If we are trying to make sense of this through strictly physical means – no ‘noticing’ or ‘perceiving’ is to be relied on – it looks like a real problem.
As I say I’m not sure of how much detail is appropriate here, so I’ll just say that Quine succeeds, or thinks he does, in solving the problem by showing that an explanation in terms of natural selection is possible of why it is that creatures arrive at more less the same dispositions – including dispositions with respect to observation sentences – despite potentially having very different sensory makeups. The key is that each of us acquires our individual dispositions in response to the common environment.
3:AM: Your recent book is called Quine vs Davidson: Truth Reference, and Meaning. As we can note from the top twelve list of your earlier book, this pitches two of the greats against each other. The key idea is to show that Davidson’s philosophy of language presupposes errors from a Quinean naturalistic perspective, and therefore many of his commitments are unsustainable. Can you outline the basic thoughts here?
GK: Boy. Well, I claim that the two clash over truth and reference. Over truth, because Quine is a Tarski-inspired deflationist, whereas Davidson has to hold that truth is primitive, in order to make his interpretative program run (Davidson of course claimed to be a champion of Tarski as well, but I think that is ultimately incorrect; he used a Tarskian framework, of course). Various people have had the idea that Davidson could embrace deflationism; I spend a lot time and probably too much effort to arguing that that cannot be.
But then Davidson is stuck with the problem that Tarski warned us about, that any use of truth to cover all of language with as much mathematical power as English is bound to be inconsistent; it will contain Cretan parodoxes. Of course people have tried to save Davidsonian semantics from this problem, but to me it’s a pretty good sign that something is rotten in Denmark.
Whereas Quine does not use the notion of truth in his account of language, and is happy to say that Tarski showed that although you can always have a better truth-predicate than a given one, there is no best one, no ultimate one. It’s a fact of life, get over it! Over reference, their difference boils down to their reactions to Quine’s argument for the inscrutability of reference. Davidson accepts it, and tries to argue that the implied relativity of reference to schemes of reference can be dismissed as harmless, comparing the situation to the relativity of temperature; a statement that it’s 28 degrees is relative to a scale, Fahrenheit or Celsius. I don’t think that works. Yet he needs the notion of reference – a full-bodied one, despite his attempts to say it isn’t – for his interpretative project. Quine takes inscrutability as showing that reference cannot play a foundational role in a theory of language.
3:AM: You say that once the errors of Davidsonian approaches are understood and resisted philosophers will be well on the way to have removed the big obstacles to a stringent naturalism. Can you say why?
GK: Quine, I think, can be understood as trying to show that the explanation for many of the things we want to say about language is that they are not themselves explanatory in the way that many philosophers take them to be, but that they have a basis either in individual biology, or in the pragmatics of language use. It’s exactly like Hume and causation: Hume could see nothing in the objects that answered to the ordinary or philosophical ideas of power or necessity, but he saw a way to explain our tendency to ascribe those notions to objects; we expect constant conjunctions to continue, like Pavlov’s dog, and lazily ascribe a queer power to the objects.
Quine did the same, although in a somewhat more complicated way, with analyticity (as we saw), and more importantly with reference. It’s another case where it would take too long to explain it properly, but basically the explanation for why we so naturally think there is a queer relation between word and object, like magic rays connecting ‘rabbit’ to rabbits (one that rules out cosmic complements of rabbits, undetached rabbit parts,or more wildly numbers), is simply that ‘rabbit’ is the word we’re accustomed to, that we’ve grown up with.
A translation of you as talking about those other things would do as much justice to the facts as the normal translation, but it would be abnormal, unfamiliar. So no wonder a straight-laced Davidsonian scheme will feel welcome, like home, but it isn’t because it is somehow better at getting the facts right. Quine can describe the facts without varnishing them, in terms of the nervous system. That’s just one piece of the picture, of course.
3:AM: Do approaches to intentionality such as found in McDowell, Anscombe avoid the pitfalls of Davidson or does your thesis damage these and other philosophical approaches. So this is a question about how this thesis affects contemporary philosophical projects. And are its issues pertinent to how you answer the question of your forthcoming new book What is this Thing called Philosophy of Language?
GK: I hadn’t thought about McDowell, Anscombe etc., so I’d better not say. And various pieces of Quine’s view that I’ve not mentioned have to be in place before his thoughts about the contemporary scene will properly make sense. But I’ll soldier on. Very generally, many of the great themes of the past 50 years or longer have centred on ‘what sort of theory of meaning is correct?'; that settled, it has been thought that important metaphysical and epistemological conclusions more or less fall out. Dummett is perhaps the clearest case of this.
Quine’s view is that, closely considered, there is no issue about realism; one has to begin philosophy with the background that reality is as we think it is – not that our thinking is sacrosanct, but that there is nowhere else to start. He is not an anti-realist, but he isn’t a metaphysical realist either. He is more of an Austin-style realist, maybe even a G.E. Moore style realist. And then Quine proceeds to explain why semantical notions – reference, truth – cannot bear the philosophical weight that is put upon them.
So I would say that the view is inconsistent with certain contemporary philosophical projects; they try for depth, but they are unchained, floating free (I don’t want to say they are spinning in the void!). I fear I’m basically saying that the sort of depth that philosophy sometimes aspires to is but a mirage (we Irishmen do not attain to such truths). But strangely perhaps the view is not inconsistent with semantics of the MIT variety – one can regard what they do as ultimately structural (subject to Benacerraf-style or Resnik-style provisos), but one can largely forget about it in the actual work.
My textbook What is this thing called Philosophy of Language? has been an exercise in Buddhist self-abnegation (maybe quite unsuccessful); rather than saying what I think, and rather than give my own very clever interpretations, I try to present the basics of the subject as it has existed for the past one hundred years, getting out the way as much as possible. And I try to proceed very slowly; there is only time for Frege, Russell, Kripke and Putnam, Kaplan, Austin, Grice, Salmon, Quine, Davidson and Wittgenstein.
3:AM: Finally, for the philosophically curious readers here at 3:AM, are there five books you could recommend (other than your own of course, which we’ll be streaming away to read straight after this) that would enlighten us further?
GK: Recent books, you mean? Gosh; will it be bad if I pass? I’m ashamed to say that I tend to read, re-read, and re-re-read a few great books, rather than keep up with what’s happening now. I’ve read Proust five times; it’s so good that when you finish you just want to start over, and so long that when you do you’ll have forgotten the first part! Likewise with philosophy; terrible, I know.
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
First published in 3:AM Magazine: Friday, October 12th, 2012.