SYNTHESE LIBRARY
STUDIES IN EPISTEMOLOGY, LOGIC, METHODOLOGY, AND PHILOSOPHY OF SCIENCE
Managing Editor: JAAKKO HINTTKKA, Boston University
Editors:
DIRK VAN DALEN, University of Utrecht, The Netherlands
DONALD DAVIDSON, University of California, Berkeley
THEO A.F. KUIPERS, University of Groningen, The Netherlands
PATRICK SUPPES, Stanford University, California
JAN WOLEŇSKI, Jagiellonian University, Krakow, Poland
VOLUME 239
THE PHILOSOPHY OF MICHAEL DUMMETT
Edited by
BRIAN McGUINNESS
Department of Philosophy and Social Science, University of Siena, Italy
and
GIANLUIGI OLIVER!
Department of Philosophy, University of Leeds, U.K.
w
KLUWER ACADEMIC PUBLISHERS
DORDRECHT / BOSTON / LONDON
THE SOCIAL ASPECT  OF LANGUAGE
Donald Davidson
1.
Which is conceptually primary, the idiolect or the language? If the former, the apparent absence of a social norm makes it hard to account for success in communication; if the latter, the danger is that the norm has no clear relation to practice. Michael Dummett thinks that by promoting the primacy of the idiolect I run afoul of Wittgenstein's ban on private languages; in my view Dummett, by making language primary, has misplaced the essential social element in linguistic behavior. In this paper I want to try to sort out and clarify the issues involved.
"There is no such a thing as a language", I wrote in a piece called A Nice Derangement of Epitaphs.1 This is the sort of remark for which one can expect to be pilloried, and Michael did not spare me. I must think, he teases, that when Bretons, Catalans, Basques and Kurds declare that language is the soul of their culture, or dictators attempt to suppress minority languages, that Bretons, Catalans, Basques, Kurds and dictators are all suffering from the illusion that there are such things as languages to cherish or suppress. Michael realizes, of course, that what I actually said was, "There is no such a thing as a language, not if a language is anything like what many philosophers and linguists have supposed." But he won't let me get away with this, for he contends that I have offered no alternative account of what a language is.2This is a little unfair; I did delineate with some care the concept of a language to which I object. If I were right in saying no actual language is like that, it would not invalidate my argument, even if I offered no alternative view. So when Michael says "The occurrence of the phenomena that interest Davidson is incontrovertible: but how can an investigation of them lead to the conclusion that there is no such a thing as a language?" I can only
1 See [35] in [74]. Reprinted in [96]. (Page numbers will be to this reprinting.) 2See [52], in [96].
1
B. McGuinness and G. Oliveri (eds.), The Philosophy of Michael Dummett, 1-16. © 1994 D. Davidson. Printed in the Netherlands.
2
DONALD DAVIDSON
agree; it can't lead to this conclusion. But it does lead to the conclusion that there is no such thing as what some philosophers (including me) have called a language.
In fact, I also did offer an alternative; of that, more later. But first, let's look at the concept of a language I opposed. It was this: in learning a language, a person acquires the ability to operate in accord with a precise and specifiable set of syntactic and semantic rules; verbal communication depends on speaker and hearer sharing such an ability, and it requires no more than this. I argued that sharing such a previously mastered ability was neither necessary nor sufficient for successful linguistic communication. I held (and hold) that the linguistic skills people typically bring to conversational occasions can and do differ considerably, but mutual understanding is achieved through the exercise of imagination, appeal to general knowledge of the world, and awareness of human interests and attitudes. Of course I did not deny that in practice people usually depend on a supply of words and syntactic devices which they have learned to employ in similar ways. What I denied was that such sharing is sufficient to explain our actual communicative achievements, and more important, I denied that even such limited sharing is necessary.
It is clear that there are two theses here which must be kept separate. The first thesis is that theje is a Platonic concept of a language which is neither instantiated in practice nor (therefore) what we normally mean by the word 'language'. The second is that neither the usual concept nor the philosophical concept is very important in understanding what is essential to verbal communication. The ultimate persuasiveness, if not the correctness, of this second claim depends on presenting an alternative account of what is essential to verbal communication.
Now let me try to clarify, still in a preliminary way, where I think Michael and I agree and where we don't. With respect to the first thesis, that there is no rigid set of rules to which those who share a language must conform, I think we have no serious argument; I have the impression that Michael holds, as I do, that actual linguistic practice is only loosely related to any fully and precisely specified language, with phonetics, semantics and syntax made explicit. What I say about proper names in this regard, for example, is close to what Michael says;3 he accepts that there is a good deal of flexibility in what we count as two people speaking the same language and he realizes that in understanding
3See[51],p. 189 ff..
THE SOCIAL ASPECT OF LANGUAGE
3
others we must sometimes draw on more than our previously mastered linguistic skills. Our differences here are matters of degree and emphasis. Nor do I think my failure to produce an alternative account of language is really what bothers Michael. I am happy to say speakers share a language if and only if they tend to use the same words to mean the same thing, and once this idea is properly tidied up it is only a short, uninteresting step to defining the predicate 'is a language' in a way that corresponds, as nearly as may be, with ordinary usage. What bothers Michael is not my failure to take this step (somewhere I do take it), but my failure to appreciate that the concept of a speaker meaning something by what he says depends on the notion of a shared language and not the other way around. My mistake, in his eyes, is that I take defining a language as the philosophically rather unimportant task of grouping idiolects, whereas he thinks I have no non-circular way of characterising idiolects. I shall come to this crux presently; but first I want to try to remove, or defuse, some differences that seem to me to be mainly verbal.
Michael chides me for extending the usual use of the word 'interpret' and its cognates to those ordinary situations in which we understand others without conscious effort or reflection, and he hints that this reveals an underlying error or confusion on my part. I do not think I have ever conflated the (empirical) question how we actually go about understanding a speaker with the (philosophical) question what is necessary and sufficient for such understanding. I have focused on the latter question, not because I think it brings us close to the psychology of language learning and use, but because I think it brings out the philosophically important aspects of communication while the former tempts us to speculate about arcane empirical matters that neither philosophers nor psychologists know much about. So let me say (not for the first time): I do not think we normally understand what others say by consciously reflecting on the question what they mean, by appealing to some theory of interpretation, or by summoning up what we take to be the relevant evidence. We do it, much of the time, effortlessly, even automatically. We can do this because we have learned to talk pretty much as others do, and this explains why we generally understand without effort much that they say.
It is significant, though, that Michael tries to saddle me with the extremely restricted meaning given the word 'interpretation' by the translators of Wittgenstein. According to this meaning, an interpretation of a word or expression is always another word or expression. This is
4
DONALD DAVIDSON
quite definitely, and I should have thought clearly, not the meaning I have in mind, though confusion is possible. If I ask how someone interpreted an utterance of the sentence 'Snow is white', and am told that she interpreted it as meaning that snow is white (or as being true if and only if snow is white), my question was not, as the answer shows, what other words the hearer might have substituted for the sentence "Snow is white." I am asking how the person understood the utterance of those words. Of course I must use words to say how she understood those words, since I must use words to say anything, but my words are not offered as the interpretation; they merely help describe it. The confusion results from conflating the use of words (to describe, in this case a mental act or state), and the mention of those words (to specify the words that constitute an interpretation). I agree with Michael that "one who ... understands a sentence need not be able to say how he understands it. He does not have to be able to say it even to himself.. ."4
It would obviously have been absurd of me to have claimed, as Dum-mett implies I have claimed, that whenever we understand a speaker we translate his words into our own. Translation is no part of the transaction between speaker and hearer that I call interpretation. Where translation of a sort may be involved is in the description the philosopher gives in his language of what the hearer makes of the speaker's utterances.
There is, I think, a related confusion about my use of the word 'theory'. I do, in A Nice Derangement of Epitaphs and elsewhere, allow myself to speak of the theory a hearer has when he understands a speaker. But like Humpty Dumpty after he has told Alice what he means by "There's glory for you", I explained first that this was a mere f agon de parier; here is what I said:5
To say that an explicit theory for interpreting a speaker is a model of the interpreter's linguistic competence is not to suggest that the interpreter knows any such theory. It is possible, of course, that most interpreters could be brought to acknowledge that they know some of the axioms of a theory of truth; for example, that a conjunction is true if and only if each of the conjuncts is true .. .In any case, claims about what would constitute a satisfactory theory are not ... claims about the
4 See [52], p. 464. 5See [35], p. 438.
THE SOCIAL ASPECT OF LANGUAGE
5
prepositional knowledge of the interpreter ... They are rather claims about what must be said to give a satisfactory description of the competence of the interpreter. We cannot describe what an interpreter can do except by appeal to a recursive theory ...
So Dummett is agreeing with me when he says, "We shall go astray ... if we make a literal equation of the mastery of a practice with the possession of theoretical knowledge of what the practice is" .6 You will notice that I do not speak of implicit knowledge here or elsewhere: the point is not that speaker or hearer has a theory, but that they speak and understand in accord with a theory — a theory that is needed only when we want to describe their abilities and performance.
On a further important issue Michael and I again see eye to eye: we both insist that verbal behavior is necessarily social. In my view, and I think in his, this is not just a matter of how we use the word 'language': there couldn't be anything like a language without more than one person. Perhaps we even agree on the underlying reason, namely Wittgenstein's, that without a social environment nothing could count as misapplying words in speech. Where we part company is in how we think the social environment makes its essential contribution.
Hilary Putnam has made much of 'the linguistic division of labor', and Michael has made clear that he too thinks the phenomenon is an important example of the way human communication depends on the society in which it is embedded.7 I do not doubt the existence of the phenomenon, or even its importance. But what does it show? Like Dummett, I don't think it shows, as Putnam insists, that "meanings ain't in the head"; for we can take it to be part of the meaning of an expression that its reference is to be determined by expert opinion. This would demonstrate that a speaker must believe there are experts, but not that there must be. So for the words 'elm' and 'beech' to pick out the appropriate trees there would have to be experts, but we cannot conclude that the meaningful use of these words demands a social setting. Dummett makes a similar point against Kripke's causal theory of names. More significantly from my point of .view, it is obvious that the linguistic division of labor is a device that can come into play only after the basic linguistic skills that tie words directly to things are already in
6See [52], p. 476.
7See [52], p. 475, and [44] in [50], p. 424 if..
6
DONALD DAVIDSON
place. So no matter how universal the linguistic division of labor is in practice, it cannot constitute the essential social element in language. We could get along without it.
Dummett writes "Davidson would like us to believe that our whole understanding of another's speech is effected without our having to know anything" and in support of this attribution he quotes me as saying "there is no such a thing as a language to be learned or mastered". Of course even if there were no such thing to be learned it wouldn't follow that we could understand speech without knowing anything; we would have to know much more. And it is in fact a major contention of my paper that we do know, and use, much more, even in grasping just the literal meanings of a speaker's words, than our mastery of any fixed set of rules would allow us to grasp. But this is not the central misunderstanding; it springs once more from the fact that Dummett does not want to notice that what I said was that there is no such thing as what some philosophers have described as a language to be learned. We all do learn languages (in the ordinary, vague sense of language Dummett and I and everyone else have in mind). As a practical matter one can't make too much of this. I did my best to sketch how I think this works. However, my interest at this point was not to describe actual practice, but to decide what is necessary to linguistic communication. And here I thought I saw (and see) clear reasons to doubt that language, if language is taken to imply shared ways of speaking, is essential. The same doubts apply to the notion of following a rule, engaging in a practice, or conforming to conventions, if these are taken to imply such sharing. (Please note the proviso.)
What is the source of these doubts? Well, starting at the small end, there is the simple fact that almost no two people share all words. Even during a conversation, each is apt to use words the other did not know before the conversation began, and so cannot belong to a practice the speakers shared in detail; here I think particularly of names and of words new to the vocabulary of one'or the other speaker. Then there are malapropisms which are nevertheless understood, slips of the tongue, and all the 'errors', as we think of them, that we would not normally commit ourselves (perhaps), but that as hearers we take in our stride: "The plane will be landing momentarily", "The phenomena is ...The data is .. .The octopi are ..." These are often part of the practice of one speaker but not of another, but communication does not suffer, though affection or admiration may wither. We have no trouble following the
THE SOCIAL ASPECT OF LANGUAGE
7
conversation of the child who says "He wented to the store" and who generally forms the past tense according to a rule which is not part of 'the language'. Actual cases grow rarer as they grow more extreme, but more extreme cases certainly exist. People who speak dialects of what we call the same language may not at first be able to make anything of what the other says; after they learn to understand each other, each may continue to speak in his own way, just as I have learned to answer letters in German, Spanish and French in English. Someone with a unique and serious speech defect may be understood by those around him.
Now to make a leap. There seems to me to be no reason, in theory at least, why speakers who understand each other ever need to speak, or to have spoken, as anyone else speaks, much less as each other speaks. Of course, the concept of 'same' (as in 'speak in the same way', or 'speak the same language') that we are depending on so heavily is already that philosophically teasing notion of similarity. I assume that two speakers couldn't understand each other if each couldn't (pretty well) say in his way what the other says in his. If we employ the translation manual relating the two ways of speaking to define what we mean by speaking in the same way, we can after all salvage something of the claim that communication requires a shared practice. But this is not what anyone would call sharing a language, nor what anyone has meant by a common practice or a shared set of rules or conventions. It is a question how Dummett might specify in a non circular way how speakers of 'the same language' must resemble one another. As Warren Goldfarb emphasizes (in discussing Kripke's 'sceptical' solution to Wittgenstein's problem), "... any problem we find in rule following will arise even with respect to what counts as the same", and he quotes Wittgenstein,8
If you have to have an intuition in order to develop the series 1,2,3,4,... then you must also have one to develop the series 2,2,2,... (Philosophical Investigations §214.) I can think of three strategies for dealing with my doubts: one can claim that I have ignored the fact that speakers of a language are responsible to a social norm even if they do not hold to it; one can concede that communication without shared practices may be theoretically possible, but argue that this is pointless speculation given that it never occurs in a pure form and probably couldn't; and, finally, it may be urged that no alternative answer to Wittgenstein's query has been offered, the query
"See [72], p. 485.
8
DONALD DAVIDSON
being: what is the difference between using words correctly and merely thinking that one is using them correctly? I will take up these three responses in turn.
According to Dummett.9
Figures of speech and other deliberately non-standard uses apart, a speaker holds himself responsible to the accepted meanings of words and expressions in the language or dialect he purports to be speaking; his willingness to withdraw or correct what he has said when made aware of a mistake about the meaning of the word in the common language therefore distinguishes erroneous uses from intentionally deviant ones.
Of course it is easy to agree that people speak as they think others do except when they don't. And if dialects can be divided as finely as need be, I can have no objections to much of the spirit of Michael's claim. The blacks in Brooklyn don't want to speak as whites do, and some individuals (James Joyce), though they want to be understood, don't want to talk as anyone else does. But the crux is the idea of obligation to the norm constituted by the 'accepted' meanings of words, for it is in omitting this idea that I have apparently left out something essential to characterising the kind of meaning involved in verbal communication. I don't see how. Suppose someone learns to talk as others do, but feels no obligation whatever to do so. For this speaker obligation doesn't enter into it. We ask why she talks as others do. "I don't do it because I think I should", she replies, "I just do talk that way. I don't think I have an obligation to walk upright, it just comes naturally." If what she says is true would she not be speaking a language, or would she cease to be intelligible? In other words, what magic ingredient does holding oneself responsible to the usual way of speaking add to the usual way of speaking?
Perhaps the answer will be that the sense of obligation only reveals itself when one is made aware of'a mistake about the meaning of a word in the common language, and one willingly corrects oneself. Of course if one thinks she is wrong about what a word means to others, she will change her mind, just as she would about anything else; will and obligation have nothing to do with it. So it must be the public gesture that counts. And no doubt most of us make such gestures willingly under appropriate circumstances.   My wife is embarrassed because I have in
9 See [52], p. 462.
THE SOCIAL ASPECT OF LANGUAGE
9
my vocabulary the word (non-word?) 'as-cer'tainable.' I'm embarrassed, too, to learn that my word is not part of the English language. I'll try, probably unsuccessfully, to change my ways. But why? Well, 1 don't want people to think 1 don't know that others say 'as-cer-tain'able' where I say 'as-cer'tain-able'. Who wants to label himself as ignorant? I'm too old to be embarrassed much by not being able to spell, and it amuses my students; but I'd spell things right if I could. These pressures are social and they are very real. They do not, however, as far as I can divine, have anything to do with meaning or communication. Using a word in a non-standard way out of ignorance may be a faux pas in the same way that using the wrong fork at a dinner party is, and it has as little to do with communication as using the wrong fork has to do with nourishing oneself, given that the word is understood and the fork works.
Of course, I don't mean that there is no reason why we are taught, and why we learn, to speak more or less as others around us do. Nothing could be more obvious: we want to be understood and others have an interest in understanding us; ease of communication is vastly promoted by such sharing. Most of us do not have the time or ability to learn very many different languages. In the case of our children, or certain poets and writers, we must or do make exceptions, but in general our tolerance of strongly deviant idiolects is limited by clear practical considerations. None of this creates a free-standing obligation, however. Any obligation we owe to conformity is contingent on the desire to be understood. If we can make ourselves understood while deviating from the social norm, any further obligation has nothing to do with meaning or successful communication. As Aristotle says, "It would be absurd to wish good for wine; if one wishes it at all, it is that the wine may keep, so that we may have it for ourselves".10 It is absurd to be obligated to a language; so far as the point of language is concerned, our only Obligation, if that is the word, is to speak in such a way as to accomplish our purpose by being understood as we expect and intend. It is an accident, though a likely one, if this requires that we speak as others in our community do.
"In employing words of the English language", writes Dummett, "we have to be held responsible to their socially accepted use, on pain of failing to communicate".11 But if the threat of failure to communicate is the reason for conforming, responsibility is irrelevant: Michael might less
10See [1], 1155 B 29-31. uSee[44],p. 429.
lü
DONALD DAVIDSON
tendentiously have written, "If we want to communicate, we should use words in their socially accepted way". The residual problem with this is that it is false in all those cases when we will be better understood if we deviate from the 'socially accepted' use. If we want to be understood, all we need to worry about is how our actual audience will take our words. The correct advice is Lord Chesterfield's: "Speak the language of the company you are in; and speak it purely, and unlarded with any other". What, after all, is the point of speaking in accord with 'accepted usage' to a company that we know will understand us only if we depart from accepted usage? I don't say there couldn't be a point in doing this, but what would it have to do with communication?
Now to address the contention that it is pointless to speculate on the remote possibility of there being speakers who, though they express themselves in distinct idiolects, understand one another. I have agreed that the possibility is in practice restricted to special cases, and I have stressed both the obvious utility of the large degrees of overlap in verbal performance we find in groups that live and talk together, and the inevitability that conformity will be learned and encouraged. The theoretical possibility of communication without shared practices remains philosophically important because it shows that such sharing cannot be an essential constituent in meaning and communication. If I am right, then important claims by Tyler Bürge, Saul Kripke and perhaps Wittgenstein and Dummett must be false, for certainly the first two have insisted that speaking in the 'socially accepted' way is essential to verbal communication, and if this is not Dummett's view it is obscure what argument he thinks he has with me. I'd better leave Wittgenstein out of this; I'll just say Kripkenstein. It also seems to me important to emphasize that much successful communing goes on that does not depend on previously learned common practices, for recognizing this helps us appreciate the extent to which understanding, even of the literal meaning of a speaker's utterances, depends on shared general information and familiarity with non-linguistic institutions (a 'way of life').
I now turn to the third challenge the idiolect must face. The challenge is to draw the distinction Wittgenstein has made central to the study of meaning, the distinction between using words correctly and merely thinking one is using them correctly, without appeal to the test of common usage. This is the hardest, and the most important, challenge, and I agree with Michael if he believes the challenge can be met only by
■:-r.-,ri<K~-
THE SOCIAL ASPECT OF LANGUAGE                     U
appeal to a social setting.12
What is needed is a norm, something that provides a speaker with a way of telling (not necessarily always) that he has gone wrong, a norm the failure to satisfy which he or she will count as having gone wrong. (There is a further condition on a satisfactory description of the norm that I shall come to later.) Speaking in accord with socially accepted usage is such a norm, but one which, 1 have argued, is irrelevant to communication unless the audience of the speaker happens to speak as he does, in which case the norm is relevant not because it is a shared practice or convention, but because conforming to it results in understanding. My proposal takes off from this observation: what matters, the point of language or speech or whatever you want to call it, is communication, getting across to someone else what you have in mind by means of words that they interpret (understand) as you want them to. Speech has endless other purposes, but none underlies this one: it is not an ultimate or universal purpose of speech to say what one thinks is true, nor to speak as one thinks others do.
The intention to be taken to mean what one wants to be taken to mean is, it seems to me, so clearly the only aim that is common to all verbal behavior that it is hard for me to see how anyone can deny it. But I can easily understand why this observation can seem too true to be interesting, given that it assumes the notion of meaning. Still, if it is true, it is important, for it provides a purpose which any speaker must have in speaking, and thus constitutes a norm against which speakers and others can measure the success of verbal behavior.
Success in communicating propositional contents — not just accidental or sporadic success, but more or less reliable success, achieved by employing devices capable of a wide range of expression — such success is what we need to understand before we ask about the nature of meaning or of language, for the concepts of a language or of meaning, like those of a sentence or a name or of reference or of truth, are concepts we can grasp and employ only when the communication of propositional contents is established. Meaning, in the special sense in which we are interested when we talk of what an utterance literally means, gets its life from those situations in which someone intends (or assumes or expects) that his words will be understood in a certain way, and they are. In such
121 am not impressed with the self-testing procedures suggested, e.g., by Simon Blackburn, nor with David Pears' similar claim.
12
DONALD DAVIDSON
cases we can say without hesitation: how he intended to be understood, and was understood, is what he, and his words, literally meant on that occasion. There are many other interpretations we give to the notion of (literal, verbal) meaning, but the rest are parasitic on this.13 Thus for me the concept of 'the meaning' of a word or sentence gives way to the concepts of how a speaker intends his words to be understood, and of how a hearer understands them. Where understanding matches intent we can, if we please, speak of'the' meaning; but it is understanding that gives life to meaning, not the other way around.
This explains why I am not impressed by Michael's or Bürge's or Putnam's insistence that words may have a meaning of which both speaker and hearer are ignorant. I don't doubt that we sometimes say this, and it's fairly clear what we have in mind: speaker and hearer are ignorant of what would be found in some dictionary, or of how people with a better or different education or a higher income use the words. This is still meaning based on successful communication, but it imports into the theory of meaning an elitist norm by implying that people not in the right social swim don't really know what they mean.14
What should we say of the many cases in which a speaker expects, or hopes, to be understood in a certain way but isn't? I can't see that it matters. If we bear in mind that the notion of meaning is a theoretical concept which can't explain communication but depends on it, we can harmlessly relate it to successful communication in whatever ways we find convenient. So, if a speaker reasonably believes he will be interpreted in a certain way, and speaks with the intention of being so understood, we may choose to say he means what (in the primary sense) he would have meant if he had been understood as he expected and intended. Reasonable belief is itself such a flexible concept that we
This formulation of the notion of meaning is not, it should be clear, Gricean, for where the present formulation rests on the (at this point unanalysed) concept of understanding, Grice aimed at defining linguistic meaning, as well as non-natural meaning generally, in terms of intentions that do not involve meaning at all. The Gricean element in my formulation is the dependence of meaning on intention.
In 'the normal1 case, Dummett writes, "speaker and hearer treat words as having the meanings they do in the language ... The view I am urging against Davidson is an adaptation of Alice's picture, according to which wordB have meanings in themselves, independently of speakers". Not independently of all speakers, he adds, since the meanings do depend on a social practice. So he must mean independently of whether the speaker or his audience happen to know what the social practice is on a particular occasion. See [52], pp. 472-473.
THE SOCIAL ASPECT OF LANGUAGE                      13
may want to add that there must be people who would understand the speaker as he intends, and the speaker reasonably believes he is speaking to such a person. Further refinements suggest themselves. But the point remains; the concept of meaning would have no application if there were not endless cases of successful communication, and any further use we give to the notion of meaning depends on the existence of such cases. These remarks should make plain why Dummett's accusation, that I endorse a variety of Humpty Dumpty's theory that meaning depends only on intention, does not find its target.
Michael objects to making understanding depend on the intentions of a speaker, especially intentions that depend on beliefs about how an audience will interpret his utterances. He says that in the 'normal' case speaker and hearer "treat the words as having the meanings they do in the language. Their so treating them does not consist in their having any beliefs about the other person". What they are going on15
are their beliefs (if they can be called beliefs) about what the words mean, not about what the other takes or intends them to mean ... No speaker needs to form any express intention, or to hold any particular theory about his audience, or, indeed, about the language, in order to mean by a word what it means in the language.
I agree that the speaker does not usually 'form an express intention', and he does not 'hold a theory', but I do say that even when a speaker is speaking in accord with a socially acceptable theory he speaks with the intention of being understood in a certain way, and this intention depends on his beliefs about his audience, in particular how he believes or assumes they will understand him. It may be that once again Michael and I are using words in somewhat different ways, in this case the words 'intention' and 'belief. I think someone acts intentionally when there is an answer to the question what his reasons in acting were, and one can often tell what an agent's reasons were by asking whether he would have acted as he did if he had not had those reasons. I don't think of consciously rehearsed beliefs or deliberately reasoned intentions as the only beliefs and intentions we have. Suppose 1 put one foot in front of the other in the course of walking to the kitchen to get myself a drink. I give the motion of my foot no thought whatever, I don't ask if it is an appropriate means for achieving my purpose. I am just walking as
15 See [52], p. 472 ff..
14
DONALD DAVIDSON
I habitually do. But if I were to decide I didn't want the drink after all, or that the door I was approaching was locked, I wouldn't take that step. I bad reasons for taking the step, and would not take it without the reasons. Similarly, it seems to me obvious that I would not speak the words I do if I thought they would not be understood. In speaking, I intend to accomplish something, perhaps to warn someone of a bear trap he is about to fall into, and I intend to accomplish this through his understanding of my words. I may take for granted how he will understand my words, but taking for granted is a form of belief. If I didn't think he would understand me I would say something else, or warn him in a non-verbal way.
2.
When misunderstandings are cleared away, what remains in this apparent dispute? We end up with me claiming that neither the ordinary, nor a certain philosophic, concept of a language is basic to the understanding of verbal communication; Michael thinks at least the former, and probably the latter, is basic. In the papers on which I have been concentrating, Michael avails himself of a notion of meaning that he does not explain, while I avail myself of a concept of understanding I don't explain. Neither here nor elsewhere, so far as I know, has Michael given an argument to show that a shared way of speaking, a practice or convention, is essential to meaning something by what one says. We know there is an argument, however, and it is possible that Michael has it in mind: it is that only a shared practice can supply an answer to Wittgenstein's question what distinguishes following a rule from merely thinking one is following a rule. I accept the fundamental importance of the question: an adequate account of meaning must provide a test of what it is to go on in the same way, that is, to continue to speak as one has previously spoken. At this point a crucial gap opens between my claims and Michael's: he has available an argument that purports to show that a shared practice is required in order to answer Wittgenstein's question, while I have only contended that a common practice isn't necessary for communication if each speaker goes on more or less as before. I have given no answer to the question what it is to go on as before. As a corollary, neither have I given any reason to think meaning is an essentially social phenomenon.
Here I will try briefly to summarise how I have tried to answer these
THE SOCIAL ASPECT OF LANGUAGE                     15
questions. Agreeing with Dummett and Kripke, and perhaps with Wittgenstein, I hold that the answer to the question what it is to go on as before demands reference to social interaction. Where I disagree is on how this demand can be met.
Suppose that each time I point to my nose you say 'nose'. Then you have it right; you have gone on as before. Why do your verbal reactions count as 'the same', i.e., relevantly similar? Well, I count them as relevantly similar; I find the stimulus in each case the same, and the response the same. You must also, in some primitive sense, find my pointings similar; the evidence for this is your similar responses. But there is nothing in the offing to let you tell whether or not your reactions are relevantly similar. No matter what the stimuli, your similar reactions will indicate that you found something similar in the situations; and apparently dissimilar responses to the same stimulus can equally be taken to show that you took the stimulus to be different, or that for you this is a similar response. As Wittgenstein says, by yourself you can't tell the difference between the situations seeming the same and being the same. (Wittgenstein, many commentators hold, thought this point applies only when the stimulus is private; I think it holds for all cases.)16 If you and I can each correlate the other's responses with the occurrence of a shared stimulus, however, an entirely new element is introduced. Once the correlation is established it provides each of us with a ground for distinguishing the cases in which it fails. Failed natural inductions can now be taken as revealing a difference between getting it right and getting it wrong, going on as before, or deviating, having a grasp of the concepts of truth and falsity. A grasp of the concept of truth, of the distinction between thinking something is so and its being so, depends on the norm that can be provided only by interpersonal communication; and of course interpersonal communication, and, indeed, the possession of any propositional attitude, depends on a grasp of the concept of objective truth.
Those who insist that shared practices are essential to meaning are half right: there must be an interacting group for meaning — even propositional thought, I would say — to emerge. Interaction of the needed sort demands that each individual perceives others as reacting to the shared environment much as he does; only then can teaching take place and appropriate expectations be aroused. It follows that meaning
I have argued this in [32] and in a number of subsequent articles.
16
DONALD DAVIDSON
something requires that by and large one follows a practice of one's own, a practice that can be understood by others. But there is no fundamental reason why practices must be shared.
The University of California at Berkeley
SINGULAR TERMS
Bob Hale
1.    INTRODUCTION:  THE NEED   FOR CRITERIA, AND  DUMMETT'S   PROPOSAL
In Frege: Philosophy of Language, ch.4, Michael Dummett argues, with great clarity and force, that it is "essential for Frege to be able to maintain that each expression may be recognised as belonging to its logical category or type from a knowledge of the way in which it is employed in the language" and, more specifically, that it is "essential, if Frege's whole philosophy of language and the ontology that depends upon it are to be even viable, that it should be possible to give clear and exact criteria, relating to their functioning within language, for discriminating proper names from expressions of other kinds."1 With these claims I am in complete agreement; indeed, although I shall not defend it here, I would make a stronger claim — provision of such criteria is an essential prerequisite, not only to any full defence of Frege's own views, but to progress on a wide range of issues in the philosophies of language and mathematics and in general metaphysics. One important, and of course controversial, part of the dependent ontology in Frege's case is his belief in the existence of a range of objects — (cardinal) numbers — to which simple numerals and other kinds of numerical expressions are to be understood as making reference. This belief may be seen as grounded in two claims: first, that such expressions function in certain statements as singular terms, and second, that the statements in which they so function are (some of them, anyway) true. The effect of viewing the matter in this way is to reduce the question whether numbers are
1 See [41], pp. 57-58. (This paper is a sequel to [80], a draft of which was presented at the conference Fundierungaprobleme bei Frege und in der Logik der Gegenwart, held in Munich during July 1991. A hnal version of that paper is to appear in a conference volume edited by the organiser, Matthias Schirn. A summary of its principal claims, which should enable the present paper to be read independently, is given towards the end of Section 1.)
17
B. McGuinness and G. Oliveri (eds.). The Philosophy of Michael Dummett, 17-44. © 1994 Kluwer Academic Publishers. Printed in the Netherlands.
,
i
'.    i
-í
-
'
r
REPLY  TO  DAVIDSON
The issue between Donald Davidson and myself is, as he states in his first sentence, whether the idiolect or the common language is primary in the order of philosophical explanation. The issue remains entangled, however, with another which I now think to be irrelevant: whether Davidson was right to deny that there are such things as languages, if a language is anything like what many philosophers and linguists have supposed. I admit, of course, that, in some of the remarks I made at Rutgers, I was teasing; not, however, with complete frivolity, since I suspected that there was no place in Davidson's intellectual landscape for any normal concept of a language. At any rate, I felt certain that the concept of a language had no philosophical importance for him, and with this I strongly disagreed. Let us look at how Davidson delineates that concept of a language that he opposes. "It was this", he says in his present paper:
In learning a language, a person acquires the ability to operate in accord with a precise and specifiable set of syntactic and semantic rules; verbal communication depends on speaker and hearer sharing such an ability, and it requires no more than this. I argued that sharing such a previously mastered ability was neither necessary nor sufficient for successful linguistic communication.
Now Davidson is quite right that sharing such an ability is neither necessary nor sufficient for communication, and he is right for the right reasons. In classing himself as having been among the philosophers he is criticising, I hope he is doing himself an injustice; at any rate, I do not think that I can be classed among them. I have frequently remarked that, to understand the speakers of a language, their utterances must be seen as rational acts, that is, as ones after whose motive or purpose we may ask. If we attempted to make sense of them on a stimulus/response model, as part of the behaviour of some natural objects, we should fail to understand them. In understanding other people's utterances, we have
257
B. McGuinness and G. Oliveri (eds.), The Philosophy of Michael Dummett, 257-262. © 1994 Kluwer Academic Publishers. Printed in the Netherlands.
258
MICHAEL DUMMETT
always to assess their intended purpose, the point they were meant to have. This we largely do without reflection, but sometimes consciously. If we did not do it at all, however, we should not so much as have the idea of distinguishing between a remark meant to be taken literally and one made ironically, or between one intended as relevant to the foregoing conversation and one meant as a change of subject, since no rules distinguish these; and then we should understand very little of what was said to us. Thus a shared knowledge of the syntax and semantics of the language will not suffice for effectively participating in communicating with other speakers.
It is also not always necessary, for essentially the same reason. If I mispronounce a word, or use it wrongly, I may still succeed in communicating. My hearer may well guess what mistake I am making, and hence what I intended to convey: he does so precisely because he is not simply reacting to the sounds he hears, but trying to discern the intention underlying them. Of course, to say that a mastery of syntax and semantics is not necessary for communication does not mean that a lack of it may not sometimes block communication: if I use the wrong word or grammatical construction, my hearer may not understand me as I intended, or at all. Likewise, in certain contexts such a mastery may be sufficient, without the need for any discernment of underlying intention: I can very probably ask the way to the station, buy a ticket and enquire when the next train leaves for York and whether I have to change, and understand the answers solely by my knowledge of the syntax of the language and (in a broad sense) of its semantics.
I do not know whether any philosophers or linguists have claimed a knowledge of the syntax and semantics of a language to be always sufficient or always necessary for successful communication with speakers of that language; but why should the falsity of such a claim justify rejecting that concept of a language delineated in the foregoing quotation as needed for a philosophical account of meaning? After all, my estimate of your intention in saying what you did, though no part of my knowledge of the language, rests on it; I have to discern why, at that moment, you said something with that meaning. Likewise, if I use words wrongly, your recognition of what I intended depends on your knowledge of the language, which tells you whether my utterance meant anything at all, and, if so, whether something I could possibly have been trying to convey, and enables you to guess what mistake I have made. "The plain of Argos is as flat as a pikestaff', the late Professor Fraenkel once said in a lecture;
REPLY TO DAVIDSON                                259
it needs a knowledge of two English idioms to enable you to tell what he meant.
These two features of verbal communication are, however, largely independent of our central issue, whether the idiolect or the common language should be taken as the primary unit for philosophical purposes. The first is entirely independent of it, for it does not relate to how we grasp what someone means, but to his point in saying something with that meaning. If communication is described in terms of idiolects, the hearer will grasp the speaker's meaning provided that their idiolects overlap at the relevant point; or, on Davidson's scheme, provided that the relevant fragment of his theory for interpreting the speaker is correct. This still leaves it necessary for the hearer to discern the speaker's purpose in saying something that has that meaning: this necessity is independent of whether we conceive of the interchange in terms of a common language or of idiolects, whether or not the latter are further relativised to the interlocutor. The second feature does require some adjustment in our description of it. In terms of idiolects, it relates to a situation in which the speaker uses an expression that either belongs to the hearer's idiolect, or theory of interpretation of that speaker, but patently not as subject to the same syntactic or semantic principles as in that idiolect or theory, or does not belong to that idiolect or theory at all. The hearer has therefore to guess at the relevant feature of the speaker's idiolect, or at how to revise his theory of interpretation. In doing so, he can appeal, not only to his own idiolect, but to other idiolects of which he has a partial knowledge, or to his theories of interpretation of other speakers. His means of divining the speaker's intention have had to be described in a slightly different way from before: the process is essentially the same.
Thus Davidson's ground for denying that there are languages, in the sense allegedly supposed by many philosophers and linguists, does not bear in any way upon the issue whether the common language or the idiolect is primary in the order of philosophical explanation. His objection to that conception of a language is indeed sound: but it concerns features of linguistic communication that must be acknowledged whichever of the two we take to be primary, and can be equally readily acknowledged in either case.
It may be objected that I have ignored another component of the conception of a language Davidson is opposing: that it is governed by "a precise and specifiable set of syntactic and semantic rules"; I shall
260
MICHAEL DUMMETT
take it that the adjective 'precise' is here intended to apply to the rules, rather than to the set. We should first note that Davidson does not make it part of the philosophical conception he is opposing that the notion he characterises fit what we ordinarily regard as the languages of mankind — Hungarian, Turkish, Swahili, etc.. A philosopher adhering to that conception of a language might regard it as applying, not to languages in that everyday sense, but only to idiolects. Moreover, Davidson would score too easy a victory if he intended it to be a part of the conception he was opposing to require, for communication to take place at all between two speakers, that they share exactly the same set of syntactic and semantic rules that apply to the words and sentences they utter in the course of the conversation. Thus, if the conception under attack is taken to characterise, not a common language, but an idiolect, it is part of the conception that two individuals can communicate only to the extent that their idiolects overlap.
If these assumptions are correct, I have to confess that I do not know
what an imprecise rule is. (People often speak of 'rigid rules'; but I do
not know what a flexible rule is, either.) It is true that my former college
was addicted to passing bye-lawa and resolutions containing the words
'shall in normal circumstances', but this is scarcely relevant: a college
meeting was always required to determine whether circumstances that
arose were normal or not, and such bye-laws would be better described
as guidelines than as rules. One may of course say that some speakers
observe a certain rule and others do not: but then the linguistic practices
of both sets of speakers would not qualify as defining a single language,
according to the conception under attack.   I do not think that there
are any imprecise rules: there are only rules that have been imprecisely
stated. The rule that personal pronouns are in the nominative when in
subject position governs the linguistic practice of one or more English
speakers in so far as they never say, e.g., "Me went ..." or "Him went
...", but only "I went ..." and "He went ..." But if it transpires that
they do say, "Me and him went...'", the rule has been wrongly stated: it
applied only when the pronoun stood alone as subject. Discovery of an
exception to a putative rule merely proves that the rule has been stated
imprecisely: if there are rules at all, they can only be precise rules.
Well, then, if only a precise rule is to count as a rule, may we not claim that languages do not have rules at all? Or could it be retorted that there could not be a language without rules? This depends, of course, on what we are prepared to count as a language.   Frege and
REPLY TO DAVIDSON
261
Tarski both criticised natural languages for being defective in the sense of not perfectly fulfilling what a language should be required to do; and we should admit that all natural languages ore in part defective, at least to the extent that they fail to guard against all ambiguities. In playing a card game, an unusual situation may arise in which the rules fail to provide for which card has won the trick, or in which they both require a certain player to play a particular card and leave him free not to. The rules are defective; there is no reason why this may not happen in language. I suppose that Davidson wishes to deny that rules can determine precisely whether a word belongs to the language, or what it means: but let us consider only syntax. You want to learn a certain little-known language, and buy a Teach-Yourself book. On page 1 it says, "The reader will be glad to learn that this language has no grammar: he has only to learn the vocabulary". You learn it, and try out your first conversation: but how are you to know whether the speaker was saying, "The soldier hit the quiet woman", "The woman hit the soldier quietly", or any one of four other things? If you can tell, the language has a grammar after all. If you cannot, then perhaps it really does lack a grammar; whether we then say that it does not deserve to be called a language, or merely that it is a grossly defective one, matters little.
Thus I sympathise partly, but not wholly, with Davidson's criticism of the conception he repudiates; but has he or anyone else ever really held it? To the extent that it involves the two theses about communication, probably not; but in respect of requiring a language to be governed by a precise set of syntactic and semantic rules, almost certainly yes: we have only to recall Frege's absurd dictum that, in virtue of their identifying the bearer of some personal proper name in different ways, two people do not speak the same language. Among those who have thus conceived of a language, some have taken a shared language as the primary unit; they have not supposed the conception to give an accurate image of actual natural languages, however. They have regarded it, rather, as an idealisation; for theoretical simplicity, we prescind from linguistic flux and differences of accent, of dialect and of individual competence. Other exponents of the conception have taken an idiolect as primary; but, again, they have not regarded the conception as yielding a realistic description of an individual speaker's linguistic competence. What a speaker knows of his language is far from being exhausted by how he personally is disposed to use it. He knows a great many ways
262
MICHAEL DUMMETT
of speaking and writing that he himself would never use: because they are archaic, or, in his view, too stilted, pedantic, coarse, euphemistic, pompous, precious, patronising, slangy, ambiguous, offensive, ungainly, obscure or ungrammatical, or simply not to his taste. On the one side we have the speaker's own active idiolect; on the other a range of passive idiolects, for the most part only vaguely distinguished from one another. Of those who employ the conception Davidson criticises, therefore, neither those who treat as primary a common language, nor those who so treat an idiolect, should be in any degree worried by Davidson's declaration that there are no languages that accord with that conception; they already know that. What should disturb them is only his view that their conception is not even of any theoretical value. The most pressing concern of the philosopher is to explain what meaning is. Proponents of the conception of a language under fire believe that the complications of which account must be taken if we are realistically to describe either a natural language or an individual speaker's imperfect mastery of it can be disregarded when our aim is to explain this concept; and I have to say that, while the complications need to be acknowledged and are interesting to characterise, Davidson has left me unpersuaded that such idealisation is out of order.
The issue I have been discussing may possibly remain in contention between Davidson and me: I have argued that it is irrelevant to the quite different issue whether the idiolect or the common language is primary. It is more tempting to assimilate that issue to yet a third, namely which of the two roles of language is primary, that of an instrument of communication and that of a vehicle of thought. This third issue is by far the most important of the three. It is essential to avoid confusing it with the issue that does divide Davidson and me. I greatly fear that I have in the past confused the two: Davidson's paper is especially helpful in making so clear a distinction between them. For on the third issue he and I are at one: and neither of us intends any equivocation about communication, for instance by admitting communication between a speaker's present and his future self. As Davidson-says (p. 5), "we both insist that verbal behaviour is necessarily social .. .there couldn't be anything like a language without more than one person". I believe that language could not serve as a vehicle of thought unless it were first an instrument of communication; unless I have grossly misunderstood him, Davidson is of the same opinion. When he maintains that the idiolect is conceptually primary, and the common language secondary, he does not mean
REPLY TO DAVIDSON
263
someone's idiolect as a vehicle of inner thought, but as an instrument whereby he communicates with others.
Davidson's contention is that, while communication is the primary function of language, there is no necessity that speakers communicate with one another in the same language. His picture of what is essential to communication is given by his experience of replying in English to people who write to him in Spanish. (This is possible in speech, too: many years ago, when my Italian was still more rudimentary than my French, I made a confession to a priest in French, and he answered me in Italian.) Suppose a Pole and a Chinese marooned on a desert island. They have no common language: so the Pole teaches the Chinese Polish, and the Chinese teaches the Pole Chinese. Each conveys to the other only a passive knowledge of his language, however: so, when the learning process is completed, the Pole speaks to the Chinese in Polish, and the Chinese replies in Chinese. Their linguistic interchanges do not take place in any one language, and they still cannot be said to share a language: but communication is perfect.
Now, do I object to this? It is a fantastic possibility, certainly; but it is in principle possible. A passive and an active knowledge of a language are different abilities, as I believe certain brain injuries make distressingly apparent: even without any such injury, someone may readily understand most Polish words when he hears or reads them, but be incapable of calling to mind the Polish for anything he wants to say when he comes to speak. Most of us over 60 are familiar with this phenomenon as it applies to personal names. The fantasy is after all little more than an exaggeration of what happens in an interchange between people speaking different dialects, in one sense of 'dialect', about which I said in Tie Logical Basis of Metaphysics that these were two ways of speaking the same language. If a Scotsman says something to me about 'wee bairns', he expects me to understand him, but he does not expect me to use that expression: this is quite different from what happens when I am attempting to converse with someone in his language. The question is not whether the fantasy is possible, but whether it yields a picture of language capable of assuming the role of explanatory primacy.
It may be remarked, first, that both the Pole and the Chinese in the story already have the concept of a language, which they have acquired by learning languages they shared with others around them. To see whether the fantasy yields a primitive picture of the use of language in communication, it must be generalised to a whole community. So let us
264
MICHAEL DUMMETT
imagine a community within which twenty very diverse languages are spoken. The children are each brought up to speak just one of these, in roughly equal numbers; but they are all taught to understand all twenty. Of course, we know that this is not in practice possible, but it is well within the bounds of theoretical possibility. Is there not, in the community, perfect communication, without anything that can be called a shared language?
I conceded that active knowledge of a language differs from passive knowledge;   but. we need to look at that more closely.    I am afraid I do not know the observational results; but I strongly suspect that while those suffering from brain injuries may lose their active command of their mother tongues, but retain their passive command, the converse does not happen.  What would it be to be able to speak and write English, but unable to understand it when heard or read, although suffering from no defect of hearing or sight?  Someone could be in such a condition only if, when he spoke, he did not know what he was saying; and that would be a form of mental dissociation that is not what we ordinarily understand by 'being able to speak English'.   Knowledge is a store, from which we retrieve items when we need them; if they are equally extensive, the same store serves an active and a passive knowledge of a language (even if there are expressions that the speaker could use, but doesn't): you have to know just the same things in order to speak and to understand a language. What has happened to someone who has only a passive, but not an active, knowledge of a language is that his retrieval mechanism has gone wrong; there is nothing that he has permanently forgotten. In the light of this, we have to ask what it is that prevents each member of the imaginary community from speaking any language but that which he was trained as a child to speak.  It cannot be inability, either a lack of knowledge or a failure to grasp a possibility, since he cannot regard the languages he does not himself speak only as a means of being communicated with, and not as a means of communication: it can only be inhibition, psychological or social. So we have essentially just a community whose members all know twenty languages; or, if none of the twenty languages is spoken by any other group of people, we may prefer to describe it as speaking just one extraordinary language having a remarkable variety of different forms. In either case, communication rests on knowledge, by both speakers and hearers, of the language or languages being used, even if the language switches with each change of speaker.
REPLY TO DAVIDSON                                 265
Thus the primary unit is still a shared language, known to all participants in a conversation; and the prototypical case is that in which they all use that language in the same way. Davidson is impressed by the fact that not all speakers of what we ordinarily count as a language do use it in the same way; in my view, over impressed. We take in our stride the pilot's announcement "The plane will be landing momentarily", he remarks (p. 6). No doubt: but if I had never heard that particular misuse, I might misunderstand a doctor when he said, "You will feel a momentary pain", while, if I had heard it, I should be anxiously uncertain what he meant. To repeat: the philosopher's main concern with language is to explain what meaning is; and none of Davidson's examples overturns the basic principle that an expression of a language has the meaning that it does because the speakers concur in using sentences containing it in a certain way.
The pilot is no opponent of this view. He is not playing Humpty-Dumpty to the passengers' Alice: he thinks 'momentarily' means 'in a moment', rather than 'for a moment'. Davidson is an opponent, because he is impressed by being able to understand the pilot. Certainly a full description of linguistic understanding must take account of our understanding of dialects not our own and of deviant uses like the pilot's; and certainly a full description of language must take account of the fact that, if enough people come to use 'momentarily' in that way, it will ipso facto cease to be a mistake, and the word will then mean 'in a moment', rather than 'for a moment'. These concessions in no way shake the fact that, if we want to explain what meaning is, we have no option but to begin with the prototypical case of converse between people who speak the same language in the same way.
It remains somewhat obscure to me how far apart Davidson and I really are on the strictly philosophical issues; but I think we do differ sharply in our attitudes to language. He quotes me as saying that, if we cannot be held responsible to the socially accepted uses of our words, we shall fail to communicate, and comments that, if the penalty is a failure to communicate, responsibility is irrelevant (p. 9); he therefore emends the antecedent to 'if we do not use words in their socially accepted way'. I concede that no one is an authority on what he (long ago) meant; all the same, I think Davidson's skills as an interpreter have failed him here. If I do not use words in the socially accepted ways, then I may or may not succeed in communicating: but, if I decline to be held responsible to their socially accepted meanings, without giving notice of any specific
266
MICHAEL DUMMETT
deviations (a qualification I expressly inserted), I am adopting Humpty-Dumpty's position, and no one will be able to tell what I mean, nor greatly care.
Let us nevertheless consider my statement as Davidson interprets it. It is then false as it stands, inasmuch as one who uses language incorrectly will often still make himself understood. Davidson concludes that, provided we make ourselves understood, we have no obligation to speak correctly; indeed, he labours the point.   "Any obligation we owe conformity is contingent on the desire to be understood", he says (p.  9). "If we can make ourselves understood while deviating from the social norm, any further obligation has nothing to do with meaning or successful communication ... so far as the point of language is concerned, our only 'obligation' .. .is to speak in such a way as to accomplish our purpose by being understood as we expect and intend". He quotes Aristotle as calling it absurd to wish good for wine, and, in the same spirit, calls it absurd to be obligated to a language.
These are dangerous half-truths.  In enunciating them, Davidson is unwittingly allying himself with a great body of English speakers who hold observance of linguistic norms in contempt, and justify doing so precisely by the plea that the sole purpose of speech is to convey one's meaning. (I say 'English speakers' advisedly, because I believe the phenomenon to be far more widespread among them than among speakers of any other language.) But a language is an instrument. If a youth uses his father's finely honed razor to cut a piece of string, he will not gain much sympathy if he says that the purpose was to cut the string, and that that purpose was accomplished, and still less if he quotes Aristotle to show it absurd to feel obligated to a razor. When pilots say, "The 'plane will be landing momentarily", they are understood as they expect and intend, all right; but they are causing me subsequent uncertainty whether I understand the doctor, and they may later cause the doctor uncertainty how to speak so as to be understood.
Perhaps, strictly speaking, we have no obligation to our language as such: but we have an obligation to others who use it to avoid damaging its effectiveness as an instrument of communication. Each generation of speakers has an obligation to future ones to leave the language with as great an expressive power as when they inherited it. This is not to say that all linguistic change should be resisted. Some changes add to the expressive power of a language, while others diminish it: we should resist the latter and encourage the former.
REPLY TO DAVIDSON                                 267
Davidson shies away from such a view for fear that it is 'elitist'; but it is not. All today's languages are products of the co-operative efforts of many generations, to which their speakers, of all social classes, have contributed: they are as subtle instruments of expression as they are because those speakers have cherished them. By destroying that love and respect, the doctrine that it does not matter how you express your meaning, as long as you convey it, serves to destroy the language's immune system; that is why it must be combated. Greatly as I admire Davidson's work as a philosopher, I regret finding myself on the opposite side of this conflict.
x /Contributors
tive and positive approach characterises his work in philosophy generally.
In September 1996, an international group of about twenty philosophers gathered in the Bohemian spa, Karlovy Vary, to participate in the immensely exciting and enjoyable entreprise of interpreting and criticising Davidson. A selection of their papers, together with Davidson's opening paper 'Externalisms', are presented in this book. The participants greatly appreciated Davidson's careful replies to all the papers, and the editors are happy that Davidson's reactions, included in the last chapter of this volume, will now also be available to the public. The succinctness, comprehensiveness and systematic structure of Davidson's concluding essay make it unnecessary to try to characterise in this Foreword Davidson's philosophical views or the topics discussed in this
book.
Although Davidson's article is structured by topics rather than persons, one name does appear    in a subtitle: not surprisingly, it is Quine's. The relation between Davidsonian and Quinean philosophy is a fascinating issue and one much discussed; as such it appears in several articles collected in this volume. Willard Van Orman Quine was among the participants of the Karlovy Vary symposium on Davidson's philosophy and contributed substantially to the discussion. He would undoubtedly have had a lot to say about the views expressed in this book. Sadly enough, he will not be among its readers.
There are four persons, apart from the contributors themselves, that the editors wish to thank especially for help with bringing about this volume: Dagfinn Fellesdal, Oslo and Stanford, John Perry, Stanford, and Gunnar Svensson and Henrik Ahlenius of the department of philosophy at Stockholm University.
Petr Kotatko Peter Pagin Gabriel Segal
1
Externalisms
Donald Davidson
Our beliefs are objective, not, of course, in being unprejudiced and formed in the light of all the evidence, but in the sense that they are true or false, and that, with few exceptions, their truth depends on matters independent of us. Our other thoughts and attitudes, in so far as they have a propositional content, are also objective: it is an objective question whether or not our intentions or desires will be fulfilled, whether our hopes or fears are realized, whether our suspicions are justified, whether our orders will be obeyed, whether we will find what we seek.
Thought is objective, and we know that it is. But it is not obvious what it is about thought that makes this possible. This question is not, in itself, epistemological. It is not the question how knowledge of the world is possible, or what justifies the beliefs we have. Skeptics think beliefs are objective, that is, that they are objectively true or false, for unless at least some of our beliefs were objectively true or false, there would be nothing to doubt. What skeptics doubt is whether we have good reason to hold those of our beliefs that happen to be true, that is, whether we have knowledge. If we cannot justify our beliefs, the skeptic maintains, then we must question whether the world is at all like what we believe it to be.
If we are not to be skeptics about the possibility of knowledge of the external world or other minds we must reject the view that all knowledge of the world depends on objects or phenomena that are directly present in in-
1