Representation, Pragmatism, and Expressivism

Pragmatism: All or Some?[1]

This paper was an invited contribution for an international conference on pragmatism convened in Sydney in August 2007.

This conference promises to be unusual for me. If we think of philosophers who emphasize reference, representation, fact, truth, truth-makers, and ontology as conservatives, and therefore on the Right, and we think of those who talk of expression, discourse, norms and practices as radicals, and therefore on the Left, then I am usually attacked from the right. My quasi-realist, they say, pretends to give us what we want by way of facts and truth, but is really only offering us a sham: fools’ truth, or fools’ facts. But in the present company, I am much more likely to be ambushed from the left. The quasi-realist, it is said, plays along with too much of the stock in trade of the right, retaining local notions of reference and representation, and even attacking iconic figures of the left for their more wholehearted expulsion of any such notion anywhere and everywhere. The quasi-realist is not a card-carrying revolutionary, they say, but an arrant trimmer. In Huw Price’s more sympathetic eyes, I am a valiant but sad Moses figure, who helped to show the way to the promised land, but could never manage to enter it himself. And as any student of politics knows, the temperature when agitators of different shades of Pink air their differences, rises just as high as it does when they rail against those on the Right.

I do not like high temperatures, so I have not come to justify standing in one place or another, or staring at the promised land of pragmatism only from a distance, but to offer a kind of apologia for not knowing where to stand. I find that knowing where to stand requres me knowing where to stand on a lot of other issues, such as Quine versus Carnap on the difference between external and internal questions, or the best way to tell the kinds of genealogical or anthropological stories that are the stock in trade of the left, or even what to think about things like functionalism or the external world. So all I can try to do is to sensitize you to some of my difficulties, and then, in a cooperative and conversational spirit, ask for guidance.

I can best introduce the issues by referring to a discussion Huw Price gives, of a passage from my book on Truth. I had written about Rorty’s substitution of a norm of solidarity for a norm of truth:

To many of us, however, the solution looks worse than the problem: language is not there to represent how things stand—how ridiculous! It is as if Rorty has inferred from there being no innocent eye that there is no eye at all. For after all, a wiring diagram represents how things stand inside our electric bell, our fuel gauge represents the amount of petrol left in the tank, and our physics or history tells how things stand physically or historically.

Price quotes this, alongside a similar passage from Frank Jackson, expressing astonishment at conferences where people attack representational views of language ‘who have in their pockets pieces of paper with writing on them that tell them where the conference dinner is and when the taxis leave for the airport’.[2] Price takes us as examples illustrating how something called ‘anti-representationalism’ often meets with something close to incomprehension, and he goes on to quote as an ally Robert Brandom who also talked of the way a representationalist paradigm is ‘taken for granted’ even in fields outside analytical philosophy

But Brandom had other disreputable branches of philosophy and theory in mind, whereas the opinion voiced in my passage, and I think Frank Jackson’s, was not intended as a philosophical defense of a philosophical position called representationalism. It was intended only as a Wittgensteinian reminder that the term representation and its cousins have perfectly good everyday uses. A historian may represent the court life of James I in a somewhat lurid light. Captain Cook’s charts represented the coast line of New South Wales with astonishing accuracy. The petrol gauge and the wiring diagram and the pieces of paper can do what they are supposed to do, or fail. These are not philosophers’ sayings, but simply parts of the everyday. We mention them in the same spirit as Wittgenstein reminds us of everyday sensation talk, not as something that all by itself demands a particular philosophical approach, but as something like the data that any such approach must end up respecting. In Moorean vein, I would suppose that any philosophy that ends up denying them is less likely to be right than they are. My problem with Rorty was that he was not, in my judgment, respecting them, but at any rate in his persona as cultural agitator and prophet, gleefully bent on trampling on them.

Huw Price and David Macarthur do not present themselves as cultural stormtroopers, bent on excising reference and representation from the everyday. Rather, they say that for the pragmatist the crucial thing is not to answer questions about the function of language in ways that encourage metaphysics.[3] On this I am entirely at one with them, and neither petrol gauges nor timetables, nor in general the Wittgensteinian reminder of the everday that I offered should encourage metaphysics, I hope. On the other hand, I should say that ‘neo-pragmatism’ in general is apt to stray from philosophical theory into the everyday with what to my eye is an alarming nonchalance. A characteristic exhibit is this sentence from Davidson, although it is no worse than many others: ‘“There is no clear meaning to the idea of comparing our beliefs with reality or confronting our hypotheses with observations” .[4] Here what starts life as a deep philosophical objection to correspondence theories of truth alarmingly metamorphoses into the rejection of a perfectly everyday activity, and one absolutely essential to our lives as rational beings. Davidson here falls over a precipice, but he has only himself to blame, since he often skips carelessly along its edge, as here: “The idea of a confrontation of belief with reality is absurd, because we can’t get outside our skins to find out what is causing the internal happenings of which we are aware”.[5] Personally I find I can perfectly well confront my complacent belief that there are plenty of eggs in the fridge with the stark reality of there being few or none, certainly without getting outside my skin, and almost always without being aware of any internal happenings, except when gastric rumbles and gurgles are propelling me to the kitchen in the first place.[6]

Perhaps this casual attitude is explained by a Quinean refusal to distinguish Carnap’s ‘external’ questions, about some kind of thing we say, from ‘internal’ questions that arise within the form of saying itself. The external question is posed, about a piece of language or discourse of some identified kind, when we ask how to explain the fact that we have come to think and talk like that: why do we go in for possible world talk, arithmetical talk, ethical or normative talk, and so on? Carnap himself was fighting ‘metaphysical’ attempts to answer external questions, although the precise interpretation of his own attitude to them is not entirely clear.[7] I suspect that Rorty, Davidson, and perhaps other neo-pragmatists are influenced by Quine’s rejection of an external/internal boundary, supposing that if representation has no proper use, since it introduces metaphysics, in answering the external-sounding question, then it must have no proper use in the internal workings of the discourse itself. I think this is a flat mistake. It certainly has nothing to do with rejection of the analytic-synthetic distinction, which at best bears on Carnap’s own construction of such a question. There is no trace of the analytic-synthetic distinction in Hume’s distinction between the anatomist and the painter, in connection with ethics, for example. Nor is there any metaphysics in his own way of tackling the question; as he himself indignantly insists, if you find metapahysics in his genealogy of ethics, ‘you need only conclude that your turn of mind is not suited to the moral sciences’.[8]

The reason Carnap can maintain his distinction is that simply insisting on the everyday is compatible with offering different interpretations of it, such as those offered by expressivists in their various domains. The propriety of everyday talk offers a datum, but it does not offer a self-extracting philosophical ‘ism’: representationalism, which the propriety of the sayings therefore establishes. It just means that if we set such an ‘ism’ up either as a good thing or as a target, then we ought to be sure what it is. And if the propriety of the everyday talk is a datum, then pragmatists would do well to ensure that what they attack as ‘representationalism’ does not encompass the everyday, so that the ordinary human baby gets thrown out with any undesirable bathwater.

One could, indeed, see Rorty himself as simply offering an interpretation of the everyday use of ‘truth’, ‘description’, or ‘representation’, in spite of his frequently derogatory remarks about them. The interpretation I went on to discuss was that in offering everyday remarks that allow sayings to be true or to say how things stand, or to represent the way things are, we deploy nothing more than a norm of solidarity with others.[9] I argued that this was insufficient for familiar reasons which boil down to this: that justifying ourselves to our peers is different from getting things right, and only offers any pale surrogate for truth provided our peers are fully paid-up members of the community that matters: fellow historians, if we are doing history; fellow legal practitioners if we are interpreting law, fellow scientists if a scientific question is on the table. But to achieve that status, these peers must have mastered techniques and norms of practice that go beyond what is properly comprehended as ‘discursive’ or belonging to discourse. For their opinions to be worth listening to they need to be more than good inference makers, for example. They need to be masters of the sextant or the archive or the laboratory, or at least to be well attuned to the results of those who are masters of these things. They need to be plugged into techniques or practices, and they need to follow the norms that belong to them. It is those that entitle them to a hearing in the après-truth coffee lounge where we try to become of one mind about something. We must not gaze at this coffee lounge where the scientists and historians congregate to chat and try to become like-minded about things, without remembering that it is a small oasis surrounded by the laboratories and instruments and libraries with which they work. One could, indeed, in desperation, try saying that the laboratories and instruments and libraries are in turn simply parts of a normative discursive practice: their use is the way to find yourself successful where it matters, in the coffee lounge. That is like saying that training as a batsman is not done with the purpose of enabling you to cope with the bowling, but in order to garnish applause and solidarity from the team afterwards in the dressing room. It’s an odd opposition to mount, and a false way of looking at the run of cricketers once it is mounted.[10]

I could put this in Sellarsian terms by saying that Captain Cook, for instance, might literally have had an entry rule for an element of his chart. You do not write a figure indicating a depth unless you have dropped a piece of lead to the bottom and measured the number of marks on the line. Had he not followed many such rules meticulously, his charts would not be revered, as they are, for their representational accuracy. There are also exit rules or in other words, ways to use his chart to navigate the waters around the coast. The chart is useful, of course, because there is a harmony between the entry rule, getting the chart to say that there are two fathoms of water in a strait, say, and the exit rule or practice, which gives you success in sailing a boat drawing anything less than two fathoms, but no more, through the strait. But there is no useful contrast here between coping and copying: the chart enables you to cope because it represents correctly the amount of water in the strait. There is no other explanation of the successes that attend sailors who use it.

Huw wonders how, if I stand so close to Wittgenstein as I have claimed, I yet cast aspersions on Rorty, who represents himself—if we may now be permitted the term— as standing equally close. I suppose I might say that Wittgenstein, trained as an engineer, was far more prone to emphasize norms of technique or practice, than purely conversational norms. In fact there is something rather comical about imagining the aristocratic and misanthropic Wittgenstein paying much attention to conversation at all, unless he was conducting it.

A pragmatist, or anyone else, would be perfectly right to insist at this point that the norms governing investigation are our norms. It is we who determine what we want to know, and how to set about finding it out. In one sense this is obviously true, but in another it may be misleading. For it is not simply down to us and our conventions whether any particular investigation is well adapted to give us results about what we want to know. Finding which do and which do not can be a long and sticky and fallible process. We cannot solve it by decision or convention. It is a matter of making ourselves into good instruments for detecting how things stand, and that is no easier than making a good petrol gauge or a good sextant.

I think that the practices of everyday assertion are sufficient as well to help with one problem Huw Price raises for me. Here he contrasts a heterological practice with an autological one, introducting the contrast with two kinds of exam. The one asks whether Aristotle was Belgian, in order to test the pupil’s knowledge of where Aristotle was born. The other asks in order to find out what the pupil thinks. A sincere answer is all that is required in the second practice; the first deploys another more exacting norm or standard. Huw points out, rightly, that for all deflationism tells us about the truth predicate, we could be in either practice. The autological pupil can say ‘it is true that Aristotle was Belgian’ as easily as saying ‘Aristotle was Belgian’ and still get the tick. Hence more remains to be said about norms of assertion than anything deflationism gives us. For in general we are in heterological practices. Sincerity is not enough (I say in general because there are, I think, conversational practices which pretty much approach it. Much vocalization in art galleries, for instance, and especially modern art galleries, is little more than autological. We effuse and compare effusions rather than trying to get something right.) As Huw knows, I have been concerned to defend the heterological parts of ethics, which does not stop with the swapping of responses, but includes a healthy practice of disagreement and doubt and persuasion, at least partly because it is more important for us to be of one mind and to have a tale about why we are minded as we are, when the topic is whether early term abortion is to be banned, than when the topic is whether Jackson Pollock was a disaster. In the empirical sciences, heterologicality is more visibly a part of the practice, since our responsibility to verification procedures is a firm norm for assertion, and falling short in implementing them is a firm reason for criticism and dissent. In Bernard Williams’s terms, we do not merely want the person producing the timetable to be sincere, but to be accurate. With ethics the elusive nature of the ‘right’ verification procedures is one of the problems, and one of the pressure points that starts theory on its road.

So much for the everyday. With it firmly in place—although, as I have already said, potentially ripe for further interpretation—what remains of an ‘ism’ for pragmatism to oppose? Well, Huw gives us a great deal of help here, in the kind things he says about my quasi-realist program as a kind of Trojan horse for introducing pragmatism into the representationalist citadel, or as a shining example for the rest of the movement to follow. He has also said some very useful things about the relation between the kind of expressivism that quasi-realism tries to help, and minimalism in semantics. Putting the two sides together, I think we can identify pragmatism in something like the following terms:

You will be a pragmatist about an area of discourse if you pose a Carnapian external question: how does it come about that we go in for this kind of discourse and thought? What is the explanation of this bit of our language game? And then:

(1) you offer an explanation of what we are up to in going in for this discourse, and

(2) the explanation eschews any use of the referring expressions of the discourse; any appeal to anything that a Quinean would identify as the values of the bound variables if the discourse is regimented; or any semantic or ontological attempt to ‘interpret’ the discourse in a domain, to find referents for its terms, or truth makers for its sentences.[11] Instead:

(3) the explanation proceeds by talking in different terms of what is done by so talking, or by offering a revelatory genealogy or anthropology or even a just-so story about how this mode of talking and thinking and practising came about. and the functions it serves.

I do not offer this as a prescriptive, defining description of neo-pragmatism. Some thinkers who like the label may reject the whole enterprise of answering a Carnapian external question, rather than giving an answer of a certain shape to it. But it will serve for the moment. With this account in front of us we can now put in place Huw’s compelling use of minimalism about truth and other semantic notions, as a useful, or perhaps vital prop for pragmatism:

(4) A pragmatist who has completed his explanation need not worry at finding truth, or other semantic notions, woven into the target discourse. By minimalism, they will be serving the same logical purposes, such as enabling generalization to take place, there, as they do anywhere else.

All this is very much in accord with the approach expressivists such as Gibbard and myself have taken to the ethical, and which can encompass the more general area of the ‘normative’; it shows us standing on the same podium as pragmatists, and possibly with a few campaign decorations showing as well.

What then of the fear, voiced by Wright, Boghossian, and others, that minimalism is inconsistent with expressivism, or at least deeply in tension with it? That would arise only if pragmatism included a zeroth law, along the lines of this:

(0) it is worries about whether ethical terms represent, or ethical sentences can be true, or about what truth makers they have, that alone motivate us to set out on the explanatory story crafted according to (1), (2), (3) and finally making use of (4).

For then there is a threat that the minimalism made use of at the fourth stage, would not itself dismiss and dissolve the worries that set the whole enterprise going. But we can now see that there are two answers to this charge, which eventually coincide. One would be that it is not those worries, or just those worries, that motivate the enterprise. The other would be that it is those worries, but that they can be expressed without the explicitly semantic vocabulary. After all, minimalism itself forces this possibility upon us. If there is a legitimate worry somewhere, put by employing a notion of truth, then by minimalism it ought to be capable of expression without it. If we can skip up or down Ramsey’s ladder without cost or concern, then equally we must be able to frame genuine problems that arise when we do use the vocabulary, without so doing. In a nutshell, it wasn’t facts that were the problem, ready to be dissolved by minimalism, but morality.

Thus, suppose we express a discontent with our understanding of ethics, by saying with John Mackie that we do not see how we can credit ourselves with knowledge of moral facts, when we are conscious that a faultless difference, such as being born in another, equally admirable culture, would have led us to an opposite opinion on what those facts are. And suppose someone tries to soothe us with minimalist thoughts about facts. There is no worry, they say, of this kind, since we no longer theorise in terms of facts: minimalism shows us how to dispense with them as thick or robust elements in any theory. Well and good, we should reply, I now express my worry without mentioning facts: I do not see how to claim that I know that p when I am conscious that a faultless difference, such as being born in another, equally admirable culture, would have led me to think that –p. In general, I continue, I adhere to norms that suggest that I should not maintain knowledge when I also accept that an equally defensible view suggests the negation of what I claim to know. And I can’t see how to exempt myself from the accusation that this is what I am doing in the present case.

I do not say that this ‘argument from relativism’ is particularly compelling—in particular, the admission that the other culture is equally admirable is usually one we do not make, and without it the worry solves itself—but it is just as compelling put without mention of truth as with it.

Or again, suppose Mackie comes out with an argument from queerness, framed in terms of the mysterious magnetic properties of supposed moral facts. Thanks to minimalism we can rephrase this: Mackie fails to see how being convinced that p can by itself involve being motivated to do some related thing, without there being an additional, independent, and contingent component of desire in the agent. Again, we may or may not be impressed, but the new phrasing is on all fours with the old.

In other areas we find the same kind of transformation. If a worry about numbers were put in terms of the difficulty of referring to abstract, non-located, causally inefficacious objects, and deflationism about reference gallops in to help, the worry will relocate itself in the question of how we know about abstract, non-located, causally inefficacious objects. Or, it might tellingly ask why we should be concerned about them. And the philosophy of mathematics again gets a motivation and a foothold. A similar transformation could be offered for puzzles about reference to possible worlds. In each case, the substantive puzzle can be relocated away from the insubstantive notions of representation and reference.

To return to the characterization of pragmatism given above, we should now see not a binary opposition, between pragmatism and some competitor called representationalism, but at least a fourfold division of alternatives. We could hold out for pragmatic stories everywhere. The opposition would be representationalism somewhere. Or, we could hold our for pragmatic stories somewhere, and the opposition would be representationalism everywhere. The last of these is, I suppose, the position manifested by those conservative philosophers with whom I started, who automatically react to any pragmatic story by reaching for notions of truth, truth-condition, truth-makers, and their kin, and proclaiming that these lie beyond the pragmatist’s grasp. Let us call these global representationalists. I stand shoulder to shoulder with Huw and I hope many others here in finding that attitude reprehensible. Still, all that is needed to oppose it are local pragmatisms, for which, of course, I am more than happy to sign up.

On the other hand, I am much less certain about global pragmatism, the overall rout of the representationalists promised by Huw and perhaps Rorty and others. The reason is obvious enough. It is what Robert Kraut, investigating similar themes, calls the No Exit problem. It points out, blandly enough, that even genealogical and anthropological stories have to start somewhere. The Humean genealogy of justice, for example, takes us as human beings with limited capacities, very definite needs, situated in a relatively niggardly environment where it is hard to satisfy those needs, and therefore having to evolve cooperative mechanisms regulating mutually beneficial conduct, restraint, and coordination. A wider Humean genealogy of values in general talks of natural propensities to pain and pleasure, love and hate, and an ability to take up a common point of view with others. It postulates a human nature in which some particle of the dove is kneaded together with the wolf and the serpent, and provides a story of our evaluative practices on that basis. I suppose a Fregean genealogical story of arithmetic and then mathematics more generally would start by placing us in a world of kinds of objects with distinct identity conditions, such as tigers and eggs and warriors, and a capacity to tally them, with there being an advantage to us in being able to rank pluralities of them by magnitude: three tigers are more of a problem than one, five eggs are better than three; eighteen warriors coming our way make for a disaster, although we could probably fight off ten. And so on.

Such genealogical stories start with a common-sense background of us, and a world of physical objects, with distinct locations, changing only according to distinct regularities with a distinct speed limit. In the books in which he provides a genealogy of morals, Hume simply takes all that for granted, just as a Fregean account of arithmetic takes the tigers and eggs and warriors for granted. If we ask the Carnapian external question about all that, then I suppose we face a choice point. It may be that we take an Aristotelian, or perhaps Wittgensteinian, line on the priority of the everyday. There is simply no place for ‘first philosophy’ to stand behind the endoxa, the given in our common-sense situation. This attitude would accord well with Huw Price’s association of pragmatism with quietism, or the rejection altogether of metaphysical questions. If we insist on posing the Carnapian external-sounding question: how come that we go in for descriptions of the world in terms of surrounding middle-sized dry goods?—then the answer is only going to be a stutter or self-pat on the back: it is because we are indeed surrounded by middle sized dry goods. That answer, obviously, draws on the referential resources of the object language, and according to the account in front of us, amounts to a victory for representationalism over pragmatism. A similar fate awaits us, in many peoples’ view, if we pose a Carnapian external-sounding question about at least the coastal waters of science. How come we go in for descriptions of the world in terms of energies and currents? Because we have learned to become sensitive to, measure, predict and control, and describe and refer to, energies and currents. That is science’s own view of how we have got where we are, and there is none better.

We may think our spade is not turned so quickly, and that we can dig below our everyday landscape. Hume thought so when he tackled the external world in Part Four, section two, of Book I of the Treatise, but he never revisited the dig, perhaps because the trench could not be shored up with the materials he had left himself, and collapsed upon him. Berkeley thought our spade was not turned so quickly, and others influenced by Descartes, such as Hobbes, did so too. The aim will be to see reference to everyday objects as an instrument for coping with something else, and the only plausible candidate will be the orderliness of experience, the only ‘given’ that looks capable of distinguishing experience of a real independent world from a mere ‘rhapsody of sensation’. As Peter Strawson so marvellously indicated in Individuals, the possibility of spatial organization of the world requires orderliness, stability and repetition, giving rise to the idea of a revisit to the same place, and the reidentification of the same kind of thing, rather than the substitution of a qualitatively identical but different thing. But whether this is a genuinely distinct and satisfying ‘genealogy’ for the concepts of a public world is, obviously, extremely doubtful, and to most contemporary philosophers it would be complete heresy, facing a battery of objections, from those centred on the impossibility of recognizing orderliness, or effecting reidentifications, in a purely private world (Wittgenstein) to those querying the possibility of even something so basic as awareness of time in such a world (Kant).

It would be very odd if either classical pragmatism in its early American dress, or neo-pragmatism as we have it now, depended on the old Cartesian priority of the Inner against the Outer. And it would be even more odd to see Wittgenstein as any kind of champion of a global pragmatism which is trying to take over the common sense homeland of representationalism by using materials fashioned from the inner life of consciousness. It would be nearly as odd to take Davidson as a similar champion of the Inner. Instead, neo-pragmatism attempts a genealogy by taking certain social facts for granted, including conversation, inference, scorekeeping, and other discursive activities, and constructing its genealogy of reference and everyday ontology on that basis. I see this as an exercise with, perhaps, its own value and its own successes. But I find myself very unclear about the motivation: epistemologically or cognitively I should have thought that what people say is a special case of what things do, and the child’s reidentification of its rattle and bricks and its ability to locate itself, comes at around the same time and requires the same cognitive resources (it may require different neural resources) as its similar reidentification of its mummy and daddy and its discernment of structure, pattern, and repetition, in what they are saying to it. Similarly, as someone who thinks that genealogical stories about norms and values are our best examples of neo-pragmatism in action, I am sceptical about reversals which give the learner’s sensitivity to norms priority over its sensitivity to the recurring elements of its environment. Generally speaking, you learn that you must stop at red lights only after you have learned to recognize red lights.

It has been well said that every explanation must start somewhere, but there is no particular place that every explanation has to start. So one could imagine a kind of rolling global pragmatism. Whenever an area of discourse becomes a target for philosophical theory, and we find ourselves worrying about its ontology or the kind of epistemology or the kind of saying about the world that constitute it, step aside to a place which, at least for the moment, seems not so worrisome, and essay a pragmatic story about the utility of the target way of thought and talk, given an environment composed in the other, less demanding way. A rolling pragmatism would differ from a foundational pragmatism in that there would be no objection to patching it together from piecemeal, and together potentially circular, explanatory projects. You might explain our penchant for ethics and normativity taking middle-sized dry goods, and some facts about human nature and human needs for granted. You might eplain the way we think about the ongoing identity of human beings in terms of our concern with psychological connectedness, and you might explain our talk of psychology in turn in terms of sensitivity to behaviour. You may talk about our sensitivity to powers and dispositions, and talk of that kind of talk as a way of organizing patterns in the Humean mosaic and reactions to them, as Hume’s own theory of causation did. But then thought in terms of a Humean mosaic might in turn be explained as a kind of abstraction out of things presented to us in our lives in the external world. And if the external world is the problem, then rolling pragmatism might equally step aside to construct a genealogy from our exposure to the Humean mosaic. Global pragmatism would be a patchwork of local pragmatisms, living by taking in each others’ washing. There never comes a point at which our spade is turned and explanation can go no further, although as the case of the external world suggested, it may often be open to doubt whether the explanations on offer always deserve the title, or always avoid drafts covertly drawn on the kinds of thing talk about which is allegedly being explained. I am not sure that rolling pragmatism would appeal to pragmatism’s founding fathers—James, for instance, at least in his later empiricist and neutral monist phase, seems much closer to being a closet foundationalist—but it is the best I can do to sympathize with anything worth calling a global program.

All this is indeed highly abstract, and as Huw and I would agree, often the devil lies in the details. Still, I find that thinking in terms of rolling pragmatism gives me some kind of entrée into global ambitions. First, it explains something of pragmatism’s, and neo-pragmatism’s hostility to Kantianism, since for Kantianism there are modes of thought that do not get this kind of genealogical or anthropological makeover. They are foundational and compulsory. They drive the whole bus, rather than arriving, contingently, as passengers. It explains at least a kind of contempt for representationalism. In terms of rolling pragmatism, “explaining” a mode of discourse simply by citing our having cottoned on to an ontology, or the facts, or the truth-makers, would be abandoning the only kind of worthwhile philosophical explanation there could be. It would be announcing that our spade had been turned, and then, amazingly, patting ourselves on the back for this fact.

Although, I think we ought to ask why Rorty, of all people, with his desire to sink philosophy and its explanatory pretensions, should have minded about that. Common sense’s answer to the Carnapian sounding question, from within common sense, and science’s answer from within science, should surely be a model for freedom from philosophy, not a target of contempt. What they model is the vanity of any philosophical ambition to step outside and to do better. It is the rolling global pragmatist who is an addict of new, philosophical, explanatory perspectives! The representationalist, on this account, is the true minimalist, modestly and sometimes admirably shying away from theory. ‘Representationalism’ on this story is what is left when philosophy becomes very, very, boring. But some, such as Wittgenstein, Davidson, and especially Rorty, might say, in at least some areas, none the worse for that.

Let me return to Carnap and the distinction between external and internal theory. One way of vindicating Frank Jackson and me, insisting on perfectly proper everyday talk of representation, is supplied by pragmatists themselves:

Realism manifestly is a theory of very great pragmatic value. In ordinary life we all assume that we live in an “external” world, which is “independent” of us, and peopled by other persons as real and as good, or better, than ourselves. And it would be a great calamity if any philosophy should feel it its duty to upset this assumption. For it works splendidly, and the philosophy which attacked it would only hurt itself. (F.C. Schiller, Studies in Humanism, p. 459).

Contrary to Dewey, perhaps far from burying it, pragmatism should be seen as vindicating realism. This view has a pedigree: it is found in James, and perhaps most famously in Quine. In effect, what is happening here is that Carnap’s external question is allowed. It is not dismissed as ‘metaphysical’, but instead it is given a pragmatic answer. The ‘language’ or mode of thought that embraces external, independent, public, objects earns its living. It works, and nothing else of which we have the faintest conception does so. So we are to embrace it.

Theorists who like their pragmatism, or their realism, global rather than local may scent an opening here. If in this way pragmatism vindicates realism about chairs and tables, why not about possible worlds, numbers, rights and duties, selves, the passage of time, and all the other posits of our everyday speech? These parts of thought or language also earn their keep, so should we not accept the inevitable, and announce ourselves as representationalists and realists about them too?

No, because if we look back at the description of pragmatism that I gave, we find there is a huge asymmetry between the case of common sense and what I called the

coastal waters of science, on the one hand, and cases like possible worlds, numbers and rights and duties or the passage of time on the other.

In embracing the common sense scheme, we embrace not only the tables and chairs it posits, but a distinct view about our relation to them. We must think of ourselves as causally influenced by them, and sensitive to their multitide of properties: their position, creation, destruction, properties and changes. To say that we mirror their doings now becomes a way of summarizing a whole host of facts about our sensitivities that come along with first positing them: that if my chair collapses, I will notice it, that if the table dances around or bursts into flames, I will register that, that were it to grow in size it would have all kinds of other consequences that I could also register and so on and so on. A mirror is quick to reflect the surrounding scene; I am not quite so quick, but I do such a good job that comparing myself to a mirror becomes almost irresistible.

Furthermore nature itself has imprinted its demands upon us. Our visual systems, for example, are hard-wired, and modular in the sense that their output lies outside our control and outside the influence of other cognitive functions. We might know that the conjurer is not producing an egg out of thin air, but we cannot stop seeing the act as if that is exactly what he is doing. Other areas lack this fixity: ethics, for instance, attracts attention partly because while its demands seem so absolute to those of us who were well brought up, we also know that they are interpreted differently, or even invisible to those who were not.

Finally, the doings of the items of common sense are directly witnessed, reflected in experience or what Kant called intuition. Their whole life, as it were, consists in their role as systematizers and explainers of experience. There is therefore no option of embracing the scheme, while holding back on its own explanations of why we do so. Whereas in the other cases, there is every prospect of bracketing the existence of possible worlds and the rest, and coming to understand why we go in for the mode of thought in question in other terms. In other words, there is every prospect of giving an anthropology or genealogy which is itself free of the commitments in question.

As already touched upon, there is the traditional empiricist option of wrestling the common sense example into the same shape as the others, by going fundamentally private: indeed one might argue that this option is already foreshadowed by Quine with the very idea of a ‘posit’, since the model is one of a theoretical entity posited in order to help with some independently known phenomenon. But as I have said, this seems not to be the neo-pragmatist intention, taking us back, as it does, to the dark days before Wittgenstein and Sellars. From this point of view, Quine’s cheerful assimilation of common sense to basic science was a throwback to the bad old days in Vienna.

A more promising, or at any rate a more up-to-date strategy for a global theory would be to urge that more is involved with the common sense scheme than meets the eye. It is only to a superficial glance, it might be said, that chairs and tables form part of a scheme that can be separated from modality, arithmetic, or normativity. It is here that various arguments against the possibility of ‘disentangling’ the one part of discourse from the other come into play. I believe that they all fail, and that the natural presumption of difference remains. The natural presumption is that we can know about the things around us without having the tools to think of them in connection with numbers (except perhaps adjectivally) or possible worlds, let alone rights, duties or values. There is a hierarchy of modes of thought, some at the bottom available to quite simple creatures, and others further up available only to very complex, self-conscious creatures who have developed specific tools for dealing with the world—dealing with what is just there anyway. And it is those modes of thought that form the precise, local, topics on which pragmatist modes of explanation get a grip.

III

Let us pause to take stock. With minimalism in place, there is a legitimate, harmless, and unilluminating place for terms like truth and representation. But for an ‘ism’ to be born, we need more than that. I have been suggesting that it is sufficient for representationalism if we have an ineliminable use of the referring expressions of the vocabulary in providing our best explanation of why we use it—this is substantially what Price refers to as the Eleatic Criterion.[12] We talk of chairs and tables because we are in a world of chairs and tables. We talk of the moons of Jupiter and forces and electrons because we are sensitive to the moons and the forces and electrons. But we do not have to give these flat-footed answers everywhere, and it is where we do not that pragmatism blooms.

Price and Macarthur represent pragmatism as the combination of a starting point in the explanation of some tract of language, and adds to that ‘a rejection of the semantic or ‘representationalist’ presuppositions which otherwise lead our theoretical gaze from language to the world—which turn an anthropological concern into a metaphysical concern, in effect’.[13] What I am querying, in these cases, is first whether the anthropological concern can be pursued without our gaze including the world, and second whether this makes that gaze specificially metaphysical. I urge that we talk of chairs and tables because we are surrounded by them and often have our attention fixed on them, and that saying this is no more metaphysical than saying that we sit on chairs and eat at tables because it is comfortable to do so. In other words, there is nothing metaphysical, to my ear, about the inshore waters of science and common-sense, and it is those that generate these answers. We only stray from common-sense to metaphysics when we start giving the same form of answer in other cases: we talk of possible worlds because the actual world is surrounded by shells of possible worlds; we talk of numbers because there are numbers, and talk of rights because people have rights. Minimalism allows us to say that it is true that there are possible worlds and numbers and rights, and true that there are facts about them, but it does not force us to regard these as ineliminably featuring in the best explanations of why we think and talk in such terms.

It does not force us, but perhaps it does not forbid us either. In some contexts there is no problem about offering explanations of this and that, using these vocabularies. Why did John take immodium on holiday? Because there is a real possibility of a stomach upset when you travel in the third world. Why do some cicadas only breed every thirteen or seventeen years? Because these are prime numbers. The peasants revolted because their rights were infringed.[14] Explanation is notoriously contextual and interest-relative, so why should it be censored in philosophical contexts?

I take it that the simple answer to this is that we do not want our philosophy to remain flat-footed and disappointing. If an external question is worth asking, it must be because the area in question has generated some kind of puzzle, and a flat-footed explanation will be one that fails to address it. It is ill-adapted to engage with whichever motivations that prompted the question in the first place. Here, clearly, there is room for differences of taste, and the soothing voice of the quietist will be heard, reassuring us that there was no need to be puzzled from the outset.

But quietism in turn is harder to believe in some cases than others. Consider for instance David Lewis’s complaint that his modal realism was apt to be met by an ‘incredulous stare’, and let us ask why that was so. Lewis himself presented the realism simply as a consequence of things we all believe and say about what might have been the case, or what would have been the case had other things also been the case. Why should philosophers have found themselves incredulous when presented with a credible paraphrase or systematization of things they believed all along?[15] Was it the geographical imagery—but what harm does that do? Some may mutter darkly about desert landscapes and profligate ontologies, but again, nobody has ever shown the benefits of the former nor the dangers of the latter. So why couldn’t we all join in a relaxed realism? My answer is that it does not help with the puzzles that prompt us to want some explanation of our temptation to modal thought and talk in the first place. Mired in actuality, how is it that we are interested in mere possibility? If we know that something is universally the case, why should it bother us whether it is necessarily the case? In other words, what motivates us to think in modal terms, and what confidence do the results of that thought inspire? Modal realism, conceived as simply a systematic rephrasal of the kinds of ways we do think, gives us no answers to such questions. A neo-pragmatism, starting with the utility of everyday thought about what would happen if… or what would have happened if… alone offers any prospect of illumination.

Price and Macarthur hint, as moral realists sometimes do, that there is something underhand about using the Eleatic Criterion, making explanatory presence into a criterion for realism. It is, as it were, tailor-made to privilege common-sense and the inshore waters of science, and tailor-made to exclude vocabularies that have different rationales and roles. I think this is right, but it should not worry us. It is precisely where we find those different rationales and roles that the space for a different kind of anthropology or genealogy opens up. We have to remember that the pragmatist’s genealogical and anthropological stories are themselves advanced as explanations. And to repeat, science and common sense tell us that the best explanations of our belief that Jupiter has more than four moons, and our coming to think in terms of chairs and tables, are respectively that Jupiter has more than four moons, and we are surrounded by chairs and tables.

They also say that quasi-realism should not be too quick to accept use of the criterion, since it takes its emulation of realism to be a ‘matter of entitlement to the semantic trimmings’, whereas if we adopt the Eleatic Criterion, it would need ‘quasi-causation, not quasi-truth’.[16] But this objection seems to rest on a misunderstanding: the goal of the quasi-realist was not to have us end up saying everything that a realist about ethics, or modality, might say. It was to have us end up saying everything that the folk say, or that is essential to the working use of the vocabulary in our thought and talk. If the realist adds to that working use a story about explanations and causation, then there is no ambition to imitate him, but to reject him. The point is to save the phenomena, not to save any old misguided philosopher.

I should like to conclude by thinking a little further about the problem of disentangling. I shall do this by considering in a little more detail the prospects for an expressivist, or neo-pragmatist genealogy for the concept of causation. Here the disentangling problem is clearly acute. If, as in the case of the ethical, we start by placing a common-sense subject in the external environment as we understand it, then there is the problem that this environment is only populated by solid, impenetrable, massive, and cohering objects: in other words, objects identified by a plenitude of causal powers and properties. So, whether or not we explicitly mention causation in giving our genealogy, it will be implied by the terms we do use, and by the standards we set down, this is enough to capsize neo-pragmatism. Hume, of course, avoids the problem by starting with an ontology only of ideas and impressions, and although these have causal powers, they are not in the same way individuated by them. But this, then, seems to reintroduce the spectre of the priority of the inner or the mental, and runs foul of the same ideology of the priority of the public that we have already highlighted.

However, the threat can be averted by a third way. We do not have to deny at any point that we are indeed in an environment of causally powerful objects. But the genealogical story gets going without retreating to an ontology of the inner, by distinguishing between the phenomenal properties of those objects, and the causal powers with which we judge them to be replete. In other words, we can ‘bracket’ the causal powers as we talk of the succession of phenomenal changes, and go on to share Hume’s account of how that succession gives rise to causal talk and thought. And we can do that without drawing upon an ontology of the inner.

It is clear that this tactic has to be legitimate in some cases. The question of what we observe is distinguishable from the question of what is causing what in many everyday contexts. Otherwise the interpretation of, say, medical epidemiology in terms of cause and effect would not be the difficult and fraught business that it is. Again, in simple perceptual cases, psychologists can and do investigate the aspects of the phenomena that give rise to causal interpretation, homing in, as Hume predicted on such things as spatial and temporal proximity, intervals, direction of motion, speed of motion and repetition.[17] We can think in this way of the phenomenal properties of screen displays in which dots appear to buffet and nudge other dots, and readily see that the interpretation of the events we see as involving causation is an artefact of our own propensities, not a given in the scene.[18] It is a matter of getting from kinematics to dynamics, and of course there is an immense psychological literature detailing aspects of this process, and the variations in phenomenal features that aid or disrupt it. It is true that a Humean story is generalizing this, seeing what is obvious in these cases as unobvious but present in all steps from phenomena to causal interpretation. But there can be no general argument that this is illegitimate, nor that the mechanisms he isolates are not universally operative.

Of course, this talk of mechanisms mmight excite a different objection, asking whether free use of causal terminology in presenting the Humean view does not somehow compromise the pragmatist promise, all by itself. It is not that the genealogy starts with a perception of a causal relation, but it does depend upon associative and other mechanisms of the mind, which bring about the functional change in the subject who now offers a causal interpretation of whatever was witnessed. The objection is not really serious, however. It would be if either more is required of this relation than is present in other cases to which Hume’s view applies, or if there was some kind of obstacle to applying Hume’s own view to the very causal mechanisms that the view specifies. But neither difficulty obtains. Hume can and will say of the mechanisms in the perceiver, that they are regularities which we interpret as causal by another application of the very same habits that have us speaking of causation in any other case. We are ourselves parts of nature, and as Hume insists when talking about the will, no more transparent to ourselves, or infallibly acquainted with thicker causal connections in our own case than are revealed in other cases. So the theory can comfortably embrace its own use of causal vocabulary in its own presentation. Thinking otherwise, is, I believe, simply a relic of the old and discredited idea that Hume, being a “sceptic about causation” should not allow himself to use the concept in any of his own writing. But that is a completely distorted account of anything he is doing, and even more certainly a distorted account of anything the neo-pragmatist would wish to take from his work.

[1] This paper is intended to be self-standing, but it owes its existence to the generous, yet critical, work of Huw Price and others. See Huw Price and David Macarthur, ‘Pragmatism, Quasi-Realism, and the Global Challenge’ in New Pragmatists, ed. Cheryl Misak, Oxford: Oxford University press, 2007, Huw Price, ‘One Cheer for Representationalism’, forthcoming in the Richard Rorty volume of the Library of Living Philosophers series; Huw Price, ‘Blackburn and the War on Error’ in Australasian Journal of Philosophy, v. 84, December 2006.

[2] Frank Jackson ‘Pragmatism and the Fate of the M-worlds’, Proceedings of the Aristotelian Society Supplementary Volume, 1997.

[5] ‘A Coherence theory of Truth and Knowledge’

[6] I have the premonition of a clatter of feet as people rush to tell me that Davidson didn’t mean to deny that, to which I can only reply that if so, he should have learned to write what he did mean.

[7] I am indebted to Robert Kraut for alerting me to some of the ambiguities in Carnap’s own view. Kraut raises the possibility of what to me would be a very congenial interpretation according to which Carnap himself allows an expressive function to ‘metaphysical’ sayings.

[8] David Hume, Enquiry Concerning the Principles of Morals, 2^nd edn. ed. Selby-Bigge, Oxford: Oxford University Press 1902, Appendix 1, p. 289.

[9] Simon Blackburn, Truth: A Guide for the Perplexed, London: Penguin Books, 2005, p. 160ff.

[10] False because vulnerable to the same kinds of argument that Bishop Butler advanced against the similar relocation of human motives in psychological egoism.

[14] Nicholas Sturgeon…See also my ‘Just Causes’ in Essays in Quasi-Realism, pp.

[15] Perhaps some Quineans had trained themselves not to believe them, but the incredulous stares were more widespread than that.

[17] Michotte, A.E. The Perception of Causation, New York: Basic Books, 1946/1963. There is a vast psychological literature following Michotte’s pioneering work.

[18] Writers influenced by Elizabeth Anscombe may resist this, claiming that causation is often a given in experience, or in other words that the brute phenomenology is infused and shaped by causal interpretations. This may be so, but does not affect the point that this in turn is explained by the dynamic properties of other elements in the scene, and the habits and sensitivities of the experienced subject. This is what enables the experimental psychologist to control the appearances and chart the variation in interpretation.