How many grains make a heap?

Richard Rorty
Philosophical Analysis in the 20th Century. Vol. I: The Dawn of Analysis byScott Soames Princeton, 432 pp, 15.95, February 2005, ISBN 0 691 12244 X Philosophical Analysis in the 20th Century. Vol. II: The Age of Meaning byScott Soames Princeton, 504 pp, 15.95, March 2005, ISBN 0 691 12312 8

I had hoped my department would hire somebody in the history of philosophy, my friend lamented, but my colleagues decided that we needed somebody who was contributing to the literature on vagueness. The literature on what? I asked. Dick, he replied, exasperated, youre really out of it. You dont realise: vagueness ishuge. My friends judgment is confirmed by Scott Soamess 900-page history of analytic philosophy. In an epilogue titled The Era of Specialisation, Soames cites the investigation of vague predicates as an area of philosophical inquiry that has exploded in the last thirty years. The intensity with which such specialised inquiries are being pursued is, he says, indicative of the fact that the discipline itself philosophy as a whole has become an aggregate of related but semi-independent investigations, very much like other academic disciplines.
Soames welcomes this change. He ends his book by saying that what seems to be the fragmentation in philosophy found at the end of the 20th century may be due to more than the institutional imperatives of specialisation and professionalisation. It may be inherent in the subject itself. Philosophers used to think that the point of their discipline was to attain a synoptic vision to see how everything hangs together. But, Soames seems to suggest, they may finally be disabusing themselves of this millennia-long misunderstanding of their own enterprise. To see what philosophy may look like in the future, consider the problem that gave rise to the huge literature on vagueness: the paradox of the heap. Soames formulates it as follows: If one has something that is not a heap of sand, and one adds a single grain of sand to it, the result is still not a heap of sand . . . if n grains of sand are not sufficient to make a heap then n+1 grains arent either. So it seems that no matter how many grains of sand may be gathered together, they are not sufficient to make a heap of sand. Some philosophers, such as Crispin Wright, respond to this paradox in the spirit of Wittgenstein. They argue that (as Soames puts it) the rules governing ordinary vague predicates

simply do not allow for sharp and precise lines dividing objects to which the predicates apply from objects of any other sort. Others, such as Timothy Williamson, hold (again in Soamess words) that vague predicates are in fact perfectly precise in the sense that there are sharp and precise lines dividing objects to which they truly apply from objects to which they truly do not but it is impossible for us ever to know where these lines are. An educational administrator (a dean in the US, a pro-vice-chancellor in Britain), asked to ratify the appointment of someone who has produced a brilliant new theory of heaps, might be tempted to ask whether this sort of thing is really philosophy. Most analytic philosophers would think this a dumb question as silly as whether inquiry into the neural processes of squids is really biology. Fruitful work in an academic discipline is whatever those trained in that discipline find it important to do. Outsiders do not get to kibitz. But suppose the dean remains obdurate. I know that biology has not reached the stage of decadent scholasticism, she might say, if only because biological research links up with medical progress. The biology department, she continues, had no trouble explaining to me why the work of their squid-neurone specialist might eventually culminate in a cure for Parkinsons Disease. I expect something similar from the philosophy department. I am told that many philosophy professors in France and Germany think that Anglophone philosophy has lost its way: that it no longer has any relevance to anything else in the intellectual world, and that its hyper-professionalism is a symptom of senescence rather than of robustness. I also hear that the undergraduates keep complaining that your department never gives courses on the philosophers whom they want to hear discussed. Before I ratify the proposed appointment, I need to be told why I should disregard such rumours. Faced with such obduracy, the head of the philosophy department might look around for a book that will straighten the dean out, one that will help her understand why analytic philosophy is (whatever disgruntled foreigners and disappointed undergraduates may say) a very good thing, deserving not only of autonomy, but of enthusiastic encouragement. Soamess book will not serve this purpose. It is a book for insiders. If you do not know before reading it why the philosophers whose work Soames treats at length Moore, Russell, Wittgenstein, Ayer, Stevenson, Ross, Quine, Ryle, Strawson, Hare, Malcolm, Austin, Grice, Davidson and Kripke are thought important, you may still be baffled after finishing the second volume. People who are already convinced that the questions Russell asked about the relation between language and reality were good ones will get a lot out of Soamess careful and perspicuous accounts of what is living and what is dead in the work of Russells successors. But those who first need to be persuaded that Russell steered philosophy onto the right track, and that Continental philosophers remain on the wrong one, will get no help.

Soames does little to tell us what counts as philosophical analysis what makes analytic an appropriate word to describe the movement whose course he traces. He is content to define that movement as a certain historical tradition in which the early work of G.E. Moore, Bertrand Russell and Ludwig Wittgenstein set the agenda for later philosophers.

He goes on to say that philosophy done in the analytic tradition aims at truth and knowledge, as opposed to moral or spiritual improvement . . . the goal in philosophy is to discover what is true, not to provide a useful recipe for living ones life. But attempting to discover what is true is not a helpful way to specify what analytic philosophers do; even Heideggerians would say the same. Soamess book is not a dramatic narrative, but rather a staccato series of summaries of a philosophers view, followed by accounts of what he got right and where he went wrong. Thus Russell and Ayer were right in thinking it important to follow Leibniz, Hume and Kant in distinguishing between necessary and contingent truths, but they were wrong to insist that analytic truths (All bachelors are unmarried or, in times past, Marriage is between a man and a woman) are the only necessary truths there are that all necessity is a matter of how people have decided, at least for the moment, to use words. They were right in thinking that facts about the meanings of our words and the information semantically encoded by our sentences are . . . real and important. But they thought, wrongly, that we have epistemically privileged access to what our words mean and what our sentences say. Others, notably Quine, realised that we had no such access, but wrongly inferred from this that there are no such things as meanings to be isolated and analysed. Quine thought that we should set aside Russells distinction between analytic truths, which are true by virtue of the meanings of terms, and synthetic truths, which signify contingent matters of empirical fact. He urged that we substitute a distinction of degree, between relatively uncontroversial assertions and relatively controversial ones, for the previous distinction of kind. Quines repudiation of Russells central doctrine was of a piece with the line of thought Dont look for the meaning; look for the use pursued in WittgensteinsPhilosophical Investigations. That book was sceptical about the analytic approach to philosophy that had appealed both to Russell and to the younger Wittgenstein. The confluence of Quinean and Wittgensteinian lines of thought found, for example, in the work of Donald Davidson created a philosophical climate in which the very idea of necessary truth was viewed with scepticism. But the climate suddenly changed, thanks to Kripke. Celebrating the entrance of his hero on the philosophical scene, Soames for once allows himself a bit of drama: With the publication in 1951 of his celebrated article Two Dogmas of Empiricism, Quine became the dominant philosopher in America, which he remained until January of 1970, when

Saul Kripke . . . gave the three lectures at Princeton that became Naming and Necessity. Soames regards the deposition of Quine and the enthronement of Kripke as a great intellectual advance. What he calls Kripkes discovery of the necessary a posteriori the observation that the truth of sentences like Whales are mammals and Water is H2O is neither contingent nor knowable by examining the uses of words is one of the great philosophical achievements of the 20th century. No single insight, he continues, has been more important in gaining the perspective needed to understand and critically evaluate the philosophical tradition stretching from Moore, Russell and Wittgenstein, through logical positivism and the ordinary language school, to Quine, Davidson and Kripke himself. Without Kripkes discovery, the history told in these pages would have been very different; indeed, these volumes would scarcely have been possible. Soames gives Kripke most of the credit for the two most important achievements that have emerged from the analytic tradition from 1900 to 1975: namely, (i) the recognition that philosophical speculation must be grounded in pre-philosophical thought, and (ii) the success achieved in understanding, and separating from one another, the fundamental methodological notions of logical consequence, logical truth, necessary truth, and a priori truth. To understand what was so daring and startling about Kripkes lectures of 1970, one has to appreciate that Quine thought only an old-fashioned Aristotelian essentialist would insist on distinguishing between necessary and contingent truths. Essentialists believe that some properties of things are intrinsic to them, that they would not be the same things if they lacked those properties. Quine agreed with Russell that this notion of intrinsic nature is a relic of pre-scientific thought (though, admittedly, one that is preserved in the speech of the vulgar). Ever since Newton replaced the Aristotelian thing-nature model of scientific explanation with a law-event model, this notion has been obsolete. Once Aristotles spell had been broken, Quine thought, everybody should have been willing to concede that lines between essence and accident were arbitrary. If drawn at all, they should be drawn with an eye for pragmatic convenience. Kripke was the first important analytic philosopher to insist that the plain man was quite right in remaining an essentialist, and that it was high time that philosophers showed proper respect for intuitions that are, in Soamess phrase, grounded in pre-philosophical thought. Water is H2O is a necessary truth, for if the chemical constitution of a given fluid were not H2O it would not be water. It is not just that we might not call it water; it would not be water. It has, in Lockes terminology, a real, not just a nominal, essence. We had not always known that this particular chemical

composition was intrinsic to the nature of water, but now we do. So our knowledge of a necessary truth is a posteriori, a result of empirical inquiry. Before Kripke, most analytic philosophers would have said that all essences were merely nominal. That is, they thought that the question of whether water was essentially H2O, or whether something with much the same properties but a different chemical composition might also be water, was uninteresting, because merely verbal. (This is also the view of most non-analytic philosophers: Heideggerians treat talk of real essences as part of the discredited onto-theological tradition, and Derrideans as a distressing symptom of phallogocentrism.) On a pre-Kripkean view, it may indeed be found convenient to find a word other than water for the strange new substance, but there are no deeper reasons nothing like what Kripke had dubbed metaphysical necessity. Darwin was generally thought to have struck a blow against Aristotelian essentialism by showing that the lines between biological species had not been drawn by God, and that species kept mutating into different species. But Kripke argued that one could accept Darwins story but still say that Whales are not fish is a necessary a posteriori truth. For whales would not be whales if they did not have a certain DNA sequence, just as water would not be water if it were not made of hydrogen and oxygen. Microstructure is a tip-off to intrinsic nature, not just a pragmatically useful redescription of things that were originally identified by their macrostructural properties. Kripke thought that their refusal to take natural kinds seriously showed that everybody from Russell to Quine had been arrogantly turning their backs on what Soames calls the great mass of ordinary, pre-philosophical convictions arising from common sense, science and other areas of inquiry convictions that philosophy cannot overturn wholesale. A typical result of this arrogance was Quines claim that everything we talk about water, electrons, numbers, mountains, you, me, the Olympian deities is just a pragmatically convenient posit. This looked to Kripke, as it does to Soames, like frivolous paradox-mongering. The popularity of such frivolity in the winter of 1970 was, Soames thinks, a sign that analytic philosophy was in dire need of reform. To succeed in demonstrating Kripkes importance, however, Soames would have to tell us more about how the discovery of the necessary a posteriori has changed things. He frequently remarks, at the end of a chapter, that the pre-Kripkean philosopher he has just polished off has confused necessary truth with analytic truth. But he often fails to make clear what revisions that philosopher would have been forced to make if

he had written after the discovery of the necessary a posteriori. For example, would Ryles anti-Cartesian polemic have been seriously weakened if he had accepted the Kripkean point that the connection between a mental state and a neural state might be necessary without being knowable a priori? It would also have been useful if Soames had been more specific about how the Kripkean revolution has changed the agenda of philosophical inquiry. He does not try to relate the necessary a posteriori to the literature on vagueness, or to another of his examples of the vigour of contemporary analytic philosophy, the explosion of philosophical and logical work on the Liar done in the last thirty years. (The Liar is shorthand for the various paradoxes created by assertions that comment on their own truth: e.g., This sentence is false, which is true if false and false if true.) It is hard to see how to reconcile the thesis that Kripke made a decisive difference to analytic philosophy with Soamess claim that it is a mistake to look for one big, unified picture of analytic philosophy in this [post-Kripkean] era. My own sense of the matter is that the discovery of the necessary a posteriori has not made the big difference that Soames attributes to it, and that it is unlikely to do so. I do not see much evidence that analytic philosophers are using different fundamental methodological notions from those they were using before 1970. My impression is that now, 35 years after the Kripkean revolution, a lot of philosophers would say to Kripke: We see your point about the Aristotelian intuitions of the plain man, and so we are willing to go along with you in calling Water is H2O and Whales are not fish necessary truths. We hereby abjure the claim that necessity is, if anything, analyticity. But so what? What follows? What have you done except alter our use of the term necessity so that we now sound a bit less paradoxical? As I see it, Kripkes lectures in 1970 aroused the interest they did not because people cared all that much about which truths should be called necessary and why, but because they cared a lot about whether truth is correspondence to reality. Philosophers are still, just as they were in Russells day, very worried about whether there is any clear sense in which our beliefs about the world are like maps whether they are somehow isomorphic to the pre-existing contours of reality (whether, in Platos metaphor, they cut nature at the joints, like a good butcher). Common sense takes for granted that there is such isomorphism that just as bits of a reliable map can be paired off with bits of a landscape, so the terms of a true scientific theory can be correlated with features of the

way things really are. But those who think that all essences are nominal point out that such pairing was easier when, as in Aristotles time, the objects of scientific inquiry were observable things such as stars and animals. It got harder when Newton began talking about unobservables such as force, mass, and acceleration, and harder still when Planck began talking about quanta. Is force the name of a natural kind? Is Hilbert space? Do these terms cut nature at the joints? Who knows? How could it matter? The more unobservables science posits, the less relevant the notions of mapping and corresponding to reality seem. However, many philosophers fear that if we cannot specify some sense in which our scientific theories map onto reality in the same way as do perceptual reports (the cat is on the mat), we are in danger of losing touch with the world. We may be tempted to become instrumentalists, people who think that we should accept scientific theories simply because they give us what we want (roughly, prediction and control of the environment), rather than because we think they accurately represent the real. For those who had such fears, Kripkes neo-Aristotelian outlook had great appeal. Quines holistic claim that the unit of empirical significance is the whole of science (rather than individual words or sentences) had done a lot of damage to the notion of correspondence. So had Kuhns denial that scientific progress is a matter of getting closer and closer to the true nature of things. Kripkes willingness staunchly to oppose the drift toward pragmatism that characterised analytic philosophy during the 1960s won him an enthusiastic audience. The controversy between realists, who think that the notion of truth as correspondence to reality can be saved, and pragmatists, who regard it as hopeless, seems to me much more fruitful than the question of whether Water is H2O is a necessary truth. The debate about the utility of the map metaphor has been going on for a long time now, and shows no signs of abating. It seethes beneath the surface of discussions of many seemingly unrelated questions. One such question is the nature of vague predicates. Timothy Williamson ends his much discussed book Vagueness with arguments against the nominalist suggestion that properties, relations and states of affairs are mere projections onto the world of our forms of speech, and concludes that our contact with the world is as direct in vague thought as it is in any thought. Crispin Wright takes up the topic of vagueness not because he cares deeply about how many grains it takes to make a heap but because doing so helps him formulate a view about the extent to which mastering a language can be treated as a matter of obedience to semantical rules rules about how to line words up with things. It is an underlying concern with the question of

whether and how language gets in touch with the world that has made vagueness a hot topic. Perceived relevance to such larger questions enables philosophers who specialise in heaps to shrug off the suggestion that they are trivialising a discipline that once had considerable cultural importance (and, in some countries, still does). To my mind, the story of 20th-century analytic philosophy (including the role of Kripke in that story) is best told by highlighting questions about whether truth is a matter of correspondence, about what is and is not out there to be corresponded to, and about whether there is any sense in which thought makes direct contact with reality. So I regret that Soamess history shoves these issues into the background. But perhaps correspondence is just my hobbyhorse, as necessity is his. However that may be, his book embodies an energetic, sustained, and praiseworthy effort to tell a story that has not previously been told in the same detail. Other versions of that story will doubtless appear before long. There will probably be one which attempts to rebut Soamess criticisms of Wittgenstein, Quine and Davidson, and which treats Kripkes revival of essentialism as a short-lived, reactionary fad. Another will tell a story that revolves around philosophy of mind rather than philosophy of language, and so focuses on figures whom Soames does not discuss (Sellars, Putnam, Chomsky, Dennett and Fodor, for example). We can also expect, sooner or later, a history of the analytic movement that deplores the situation Soames describes (accurately) when he writes: Gone are the days of large, central figures, whose work is accessible and relevant to, as well as read by, all analytic philosophers. Philosophy has become a highly organised discipline, done by specialists primarily for other specialists. As these diverse narratives appear, we shall become better able to evaluate Soamess suggestion that such specialisation is inherent in the subject itself, and that the day of synoptic philosophical visions is over. We shall also be in a better position to foresee which 20th-century Anglophone philosophers will continue to be read. Nobody in 1904 would have predicted that Frege and Nietzsche were the only two philosophers of the late 19th century whose writings would still be studied intensively in 2005. We are in no position to tell which, if any, of the figures whom Soames discusses will still seem important in 2105. It is anybodys guess whether analytic philosophy will burst the boundaries of the Englishspeaking world and become dominant in universities around the globe, or whether the work of some synoptic visionary will persuade young philosophers in Britain and the US to turn their backs on the movement that Russell initiated. Someday but not for quite a while intellectual historians will be in a position to render judgment on the question of

whether that movement succeeded in bringing quasi-scientific rigour to philosophy, or instead ran out into the sands.

Letters
Vol. 27 No. 5 3 March 2005 From Scott Soames Richard Rorty begins his provocative discussion of my book on analytic philosophy with a fable about a department chair facing the daunting task of persuading the dean to hire a philosopher working on vagueness (LRB, 20 January). As Rorty himself implicitly acknowledges, the candidate is not trying to become an expert on the composition of heaps of sand in order to help solve the practical problems facing civil engineers, but so as to illuminate the rules of logic and language, and their application to the world. This enterprise is one of several in which analytic philosophers are advancing by replacing Rortys metaphorical question Are the sentences we use to describe the world maps of an independent reality? with more specific, nonmetaphorical questions. Rortys failure to register this is connected to a broader failure of perspective arising from his disengagement from analytic philosophy in the last twenty-five years. A case in point is Rortys query as to whether Ryles anti-Cartesian polemic is seriously weakened by his ignorance of the necessary a posteriori. Rorty could answer his own question by rereading my discussion of Ryles argument that perception is not a brain process, and that belief and desire cannot be seen as causally efficacious internal states unless one is willing to embrace the idea of a Cartesian ghost in the machine. I show how these arguments depend on the faulty assumption that since our knowledge of the connection between the mental (seeing something red) and the physical (the brain processes by which this takes place) is a posteriori, or empirical, the two are not necessarily connected, and so cannot be identified. What explains Rortys blindness to such points? As I see it, his weary scepticism is not so much the result of Olympian detachment as the disappointment of a true believer. The analytic philosophy that both captures and limits his imagination is that of its middle period from the early 1920s to the late 1960s when different versions of the dream of starting philosophy anew, by identifying it with the rigorous analysis of language, competed for the honour of making the subject respectable to the modern mind. Having succumbed to this dream, and then rejected it for the wrong Quinean and Wittgensteinian reasons, Rorty cannot understand Kripkes achievement in expanding the philosophical conception of necessity beyond the purely linguistic. Finally, Rortys distrust of specialisation and his desire for a grand synthesis ignore the quality and quantity of what there is to synthesise. The value of specialisation is that it increases the chances of getting things right in each of the areas to be synthesised something that great philosophers from Plato to Descartes, Hume and Kant have always recognised. In earlier eras, when it was not obvious that the scope of human knowledge far exceeded what could be encompassed by a single mind, the challenge of explaining how everything hung together was not so clearly unmanageable. Today, it is, and the solution is not to do badly what cannot be done, but to do well what can be: to construct a series of limited, but accurate and overlapping syntheses that together illuminate reality as we know it. This is what we should ask of analytic philosophy. Scott Soames University of Southern California

From Matilde Macagno As a mathematician, I would say that one needs at least four grains of sand to make a heap. Matilde Macagno University of Iowa