Philosophy is a bitchy subject. That is not to say that philosophers are nastier to each other in print than people in other subjects are, but that in philosophy the distinction between academic discussion and personal attack is more subtle: it is harder in philosophy to tell when the attack is on a writer’s motives rather than his work. Michael Dummett, in these two books on Frege (and there is more to come), evaluates the views of a fair number of other writers, many of whose errors he considers to border on the perverse, and in so doing finds himself on the boundary between comment and invective. Sometimes he succeeds admirably in showing why he thinks someone thoroughly mistaken without becoming at all splenetic; sometimes he fails.
In the course of some thirteen hundred pages Dummett evaluates the central beliefs of most other English-writing philosophers of language. And in so doing he presents his own views on linguistic meaning, abstract objects, identity and, above all, the structure of the skills and knowledge one brings to speaking a language. Much of it is fascinating and masterly, and the first edition of the first of these books has become a standard text. Both books are shaped around an exposition of the views of Gottlob Frege, who worked in almost total obscurity between about 1870 and about 1920 and published in his lifetime roughly a thousand pages, all concerned in one way another with the foundations of mathematics.
There has been a tendency in recent British philosophy for people, typically at a point their work when one would expect a major original statement, to come up instead with study of an earlier writer, from behind whose mass they can emerge periodically to throw out a sharp non-interpretative idea: Berna Williams on Descartes, Crispin Wright on Wittgenstein; before either of them, P.F. Strawson on Kant. Dummett is not the first to see Frege’s suitability for this role: twenty years ago, in Three Philosophers, Elizabeth Anscombe and Peter Geach treated Frege alongside Aristotle and St Thomas, two philosophers whose usefulness has often been their magisterial quality and the ways it can be appropriated. There are good reasons for assigning this kind of role to Frege. He writes in a wonderfully clear, untentative way, which generates the impression that there is a deep and systematic doctrine behind what he says, which is not always obvious only because of the weakness of language and the reader’s obtuseness. And when one works on Frege one finds this impression confirmed: one sees some justification for the condescending tone which, no doubt, was one reason he had so few readers in his lifetime.
Dummett’s claim is that, in what Frege had to say to sort out the half-truths, confusion and idiocies in the philosophy of mathematics of his day, we can see clearly what an analytical philosophy of language must be. That is, we can extract from his writings ‘basic tenet of analytical philosophy’: that the only way to study thought is to describe it in terms of the structures and properties of language. Dummett therefore has two task. He must describe those views of Frege’s which allow one to formulate a version of this principle. And then he must use the principle as a means of sizing up the positions of other philosophers. Frege: Philosophy of Language is directed largely to the second task; The Interpretation of Frege’s Philosophy to the first. One would therefore expect that the latter book could be read independently of the former. But it cannot be. The obstacle is Dummett’s inclusion in The Interpretation Frege’s Philosophy of a series of emendations, additions and self-justifications pertaining to Frege: Philosophy of Language. The most remarkable instance of this is in the 11th chapter, ostensibly on Geach’s interpretation of Frege, which begins with a seemingly interminable apology for the way Geach’s views were handled in the earlier book. If a second edition of Frege: Philosophy of Language had to appear now – I would myself have preferred an earlier chance of seeing the promised book on Frege’s philosophy of mathematics – then much of the content of the second book should have been worked into it. And some of the rest is not best suited to book form at all. It’s an editorial mistake, and Duckworth’s must bear some of the blame. The result is an oversized and overpriced book to add to the burdens of students and university libraries.
Dummett’s central claim is that there is a Fregean theory of meaning which provides a kind of foundation for philosophy, just as Frege’s work was intended to provide a foundation for mathematics. This theory of meaning is to be a theory of linguistic knowledge: it is to centre on an explicit statement of the knowledge that any competent speaker implicitly has in being able to use such a language as ours. There are two kinds of linguistic knowledge involved: knowledge of the meaning of particular words and constructions, which consists in having a grasp of what Frege called the sense of the expression concerned, and a knowledge of the more general capacities and uses of language. It is not clear quite what range of philosophical questions Dummett thinks can be dealt with just by appeal to either of these sources of knowledge, but it seems to be quite a wide one. For example, in Chapter 11 of Frege: Philosophy of Language he takes it as evident that issues about fatalism and the determinateness of the future can be settled by reference to the workings of tenses and other time-indicators in language. As if physics could not come into it too.
Dummett works with what is in effect a two-tier articulation of philosophical questions. On the bottom tier there are questions which can be settled by reference to implicit knowledge of language alone: With these one proceeds by direct argument, and false conclusions can be traced to bad reasoning or mis-articulation of the obvious. Superimposed on this is a layer of questions about language and thought which cannot be settled in the same way, where one has to proceed by making alternative hypotheses and seeing which ones best account for the facts. Here false conclusions have no uniform origin, and a steady mind may not suffice to discover the truth.
The first tier constrains one to interpreting language and thought along Fregean lines, in terms of senses of words which allow one to use them to refer to things. Dummett’s greatest ire is reserved for those philosophers, mainly American, whose construal of language would deny the authority of the knowledge encapsulated in senses. The text is dotted here with ‘Quine is guilty of confusing ...’, ‘Quine’s sloppy use’, ‘Putnam makes some attempt to pretend ...’, and so on. These sections are unusually irritating in their condescension to the dominant writers in the field, and whole chapters are made difficult to read by the annoyance that this tone produces.
Quine’s views entail that understanding the meaning of a word largely consists in believing a theory (or saying yes to one, at any rate): a theory which need not be true. Understanding a word need not therefore give any knowledge: you cannot settle disputes just by reference to your skill in making explicit what you mean. To understand a language, the way a community speaks and thinks, one will first of all, on this model, have to understand the claims it expresses, so that to understand our thought one has to understand the theories, for us to a large extent scientific theories, which inform what we say. The source of semantical-philosophical authority thus lies, not in implicit knowledge of language, but in a reflective understanding of theory. It is not a big step from here to the ‘realism’ of Saul Kripke and Hilary Putnam, according to which the factors which determine a sentence’s truth (notably the determination of what its terms refer to) may to a large extent consist in objective facts quite independent of the knowledge of most speakers. On this view, to understand some features of some languages one will need, not just to be able to speak the language and describe what one does and knows in speaking, but also to describe the causal workings of the subject-matter of the language and of the connections between this subject-matter and what speakers say.
Dummett’s real fire is reserved for these latter views, which are rivals to a Fregean account, while a view such as Quine’s, rejecting the purposes as much as the claims of a Fregean approach, is treated with an icy disdain. Quine’s model in not allowing an inquiry into senses is ‘in effect, though not in intention, anti-intellectual’. Quine does indeed work with a very different picture of the intellectual enterprise, and with a very different picture of thought for that matter, from the ones Dummett uses, and I don’t doubt that Dummett is right in seeing a more direct rivalry with views such as those of Kripke and Putnam. These do allow for a specifically linguistic knowledge, but give it a radically different, un-Fregean content. For Frege, the sense of a word is that specification of its reference which determines what it is and allows one to know what one is talking about. (I put it in these vague terms to avoid committing myself on questions I’m not sure Frege even formulated to himself.) Kripke attacks just this idea in Naming and Necessity, a text which is itself treated with religious devotion in some quarters. He argues that one can understand a proper name without having enough direct information about the thing it refers to to distinguish it from others. The knowledge one has in understanding a word may instead link one’s usage to that of others (perhaps particular others – the information can be quite complex), some of whom will have direct distinguishing knowledge of the object. Putnam describes further possibilities of linkage between other kinds of words and other aspects of the speech of a community. There is a double threat to Frege’s programme here: words seem not to have the right kinds of senses, and the sum of a standard speaker’s information about his language does not add up to an understanding of it. The first of these threats is the more serious to Frege’s own programme, the second to Dummett’s use of it.
Dummett is confident that both threats can be met. In Frege: Philosophy of Language, he says, discussing Kripke’s views, that ‘no shadow of reason has yet emerged for rejecting even the strongest conceivable version’ of Frege’s account of how reference depends on sense. Kripke’s views are quite without force, then. In The Interpretation of Frege’s Philosophy the conclusion is weaker, though nothing has been retracted. It seems that Dummett’s later view is that Kripke’s arguments have considerable intuitive appeal, but must in the end be rejected. Dummett’s strategy, in both cases, is to argue that there are alternative explanations, in terms of Fregean senses, of the phenomena that one might explain in Kripke’s or Putnam’s way. However, to show that there is an alternative to a position is not to show that that position is wrong: it is to create a situation in which the advantages of each view have to be carefully considered.
The contrast between the position of the earlier book, that the necessity of a Fregean theory of sense can be conclusively established, and that of the later one, that such a theory can be defended against its rivals, is crucial in terms of Dummett’s use of his articulation of Frege’s theory as an authority by which to judge other philosophers. The later position does not seem to support such a use of Frege’s philosophy. In the first book he seemed to be presenting a comparison of the explanatory merits of competing theories as if it were a deduction from first principles, and here, in the later work, he seems to be presenting his arguments as such a comparison, and then treating the results as a weighing-up of the rival views against a fixed standard. One sad result of this is that the reader has to piece together for himself what the favoured Fregean theory is that works better than all the others. The best construal of the sense of a word is left in doubt.
Much of The Interpretation of Frege’s Philosophy is spent commenting on the views of other interpreters of Frege, and one might expect this discussion to shed some light on the true Fregean philosophy. Unfortunately, many of the rival interpretations of Frege get hung up on various ‘isms’. Was Frege a Rationalist, an Idealist, a Realist? And the easiest conclusion to draw is that the histories of these movements are unlikely to help classify the work of a vastly original and rather naive thinker. There are two exceptions to this. The first is the debate between Dummett and Hans Sluga over Frege’s realism. Sluga has argued, against Dummett, that instead of being the first philosopher clearly to separate questions of how we refer to objects from questions of what our ideas of them are, Frege conceived of objects less ‘objectively’, more as being given by our ways of thinking of them. Sluga points out the undeniably strong influence of Kant on Frege, and the fact that although Frege expresses disapproval of the ‘idealism’ in Kant’s system, he never says how he would replace it. Dummett replies essentially by pointing to Frege’s disapproval, and to the ways he has suggested of disambiguating Frege into a full coherent realism. The discussion here is much more relaxed than in many other parts of the book, under a superficially grumpy manner, and very helpful.
The general issue of realism lurks behind another interpretative dispute, this time spread throughout both books: this is the dispute with P.T. Geach about identity. In the course of a long scattered and tangled discussion the following becomes clear: if one is making a Fregean analysis of language, there will be close connections between one’s interpretation of identity (‘x is the same thing as y’), of quantification (‘there is an x such that ...’), and of the sense of a proper name. Only a few combinations of views on these three topics are coherent, and Dummett has one; moreover, his combination emphasises a ‘realism’, or at any rate an independence of objects from language, that some others do not. ‘Realism’ is for Dummett the central undetermined question in the philosophy of language, the one that is not settled by appeal to what makes sense. It amounts to the question of whether all indicative sentences have a determinate truth-value: whether if a sentence is unambiguous and is clearly related to its context, it must, as a matter of objective fact, be either true or false. There is a subtle and pervasive irony in Dummett’s insistence on taking the reading which makes Frege most a realist: Dummett’s own inclinations are anti-realist, and he thinks that only by working out the most coherent realist account of language can we see what a persuasive anti-realist variation might look like. In Chapter 20 of the second book Dummett gives a long, candid, often fascinating discussion of the difficulties in seeing what the issue of realism really amounts to. It reinforces my suspicion that there is no single issue here, but a large number, about meaning, reference, negation and quantification in particular, all of which can be settled only in the tentative way that Dummett reserves for adjudicating the realism/anti-realism dispute. (Of course, for an anti-realist there is no reason to be sure that the dispute can be adjudicated; and for an anti-realist answers to questions about the subtler intentions of a thinker of the past, separated by both time and subjectivity from any easy assignment of a truth-value, need have no definite answer, even when the thinker is Frege.)
Any coherent view of language will reveal something about the relations between other possible views, if it is sufficiently comprehensive and well-articulated. More than coherence will be needed, though, if the theory is to be used as a basis for dismissing rival views as inherently misconceived, if it is to impose on philosophy the kind of two-tier structure I mentioned earlier. But thinking that one’s own theory has imposed such a structure will easily lead to the kind of grumpiness that I have described as permeating the subject. There is no easy dividing line in philosophy between those questions which can be settled just by appeal to the force of arguments and those which require the slow evaluation of rival hypotheses. There are questions of both kinds, but it is never a mechanical matter to tell which is which. And as a result we will always draw the dividing line in different places, thinking our opponents’ views to be not just wrong but perverse. One symptom of this will be a general testiness, and another will be the use of philosophers of the past as models or standards, for it is sometimes easier to see in retrospect where the line once lay, how argument and conjecture once co-operated. Surely the hope that we may get from a study of a not-too-distant thinker some idea of how this co-operation may be organised today is not completely unfounded.
There is another kind of philosophical scholarship, which consists in simply trying to recapture what it was that a past philosopher said. One needs a very good knowledge of Frege’s text and of contemporary philosophical issues to read Dummett in this way. Readers without this preparation, who simply want a straightforward account of Frege’s doctrines, will find Gregory Currie’s Introduction to his Philosophy very useful.
Send Letters To:
London Review of Books,
28 Little Russell Street
London, WC1A 2HN
Please include name, address, and a telephone number.