As you will have gathered from the title of my talk, I have a complaint to make about the science of linguistics. Let me explain what is bothering me.

   The task of linguistics, I hardly need to remind you, is to study language. The task of language, in its turn, is to enable us to communicate. A language is a system of visual or audial signs, each of which means something or other. The advantage of having such a system is that we can make our thoughts accessible to our neighbours. If we write or utter words which, in a given language, have the appropriate meaning, then anybody who is familiar with that language, and cares to read or listen, can see what's on our mind.

   The linguist makes it his business to explain how a language such as English works, how it manages to provide this remarkable service. One might be tempted to say that there is not a great deal to be explained. The meaning of an English sentence is completely arbitrary: there is no natural connection between, say, the string of letters 'It is raining' and the weather phenomenon it refers to. In order to master the language, one simply has to learn the gratuitous convention which binds the two together.

   A moment's reflection reveals, however, that this cannot be the entire story. We cannot have a separate convention for every English sentence. There are infinitely many English sentences, most of which have never been uttered or even considered by anybody. When a creative writer produces a novel English sentence, no new convention is needed to endow it with meaning: the sentence has had its meaning all along despite the fact that nobody ever used it before. And when read or hear the sentence, we do not have to learn what it means; our existing knowledge of the language enables us to see immediately what the sentence says.

   Hence meaning must be assigned to English sentences not piecemeal but rather by means of general rules. Only on this assumption can we explain the fact that a sentence which has never entered anyone's head --or mouth-- can nevertheless be meaningful: the general rules endow it with meaning regardless of whether anybody has ever bothered to apply the rules to see what that meaning is. It is because we have internalized those rules that we can comprehended a sentence we have never seen before.

next page

contents

   Now the tantalizing thing about language is that while most English speakers are consummate experts on applying those rules, not even the most eloquent among them know those rules. When I first arrived in England many years ago with a rather rudimentary knowledge of the language, for some time I was naive enough to ask natives speakers questions like: Why is what I have just said ungrammatical? What rule have I violated? Why is it that I can say 'John shaved himself' but not 'Himself shaved John'? The answer was invariably 'Well, this is simply way not the way we say it', or something illuminating. At the first I was a little annoyed. It seemed to me that my English friends were deliberately keeping me in dark, that they conspired to perpetue my linguistic competence. But gradually it dawned on me what the real situation as: they simply did not have a clue how to answer my question.

   The situation may seem paradoxical. How can we able to successfully apply grammatical rues if we do not know what they are? But in fact the situation is typical rather than extraordinary. Birds are virtuoso flyers, responding flawlessly to atmospheric conditions, but nobody expects them to know the laws of aerodynamics. Or, to turn to humans, the best cyclists would probably get stuck if challenged to explain exactly what they do to keep balance on a bike. There is no end of feats we know how to accomplish without having a clue how we accomplish them. Linguistic ability belongs to this category. We know how to associate sentences with their meanings without being able to spell out the precepts we follow in doing it. We know, for example, that the fact that Jade shaved Fred can be stated by saying 'Fred was shaved by Jade', whereas the fact that Jade shaved herself cannot be stated by saying 'Herself was shaved by Jade'. But nobody really knows why it should be so.

   A linguist is someone who wants to find out. His ambitions is to ferret out the grammatical rules which we store in the depths of our minds and yet are completely ignorant of. His aim is to produce a grammar, a description of the arcane apparatus which underlies our linguistic competence.

   A grammar of English is a system which determines which English expressions are well-formed and what each of them means. In other words, it is a generator of expression-meaning pairs. This is indeed how grammar is frequently defined. Noam Chomsky, the uncontested guru of modern linguistics, says that

previous page

next page

contents

(a) person who has learned a language has constructed a system of rules and principles --a grammar-- determining a sound-meaning relation of some sort over an infinite domain. The linguist's grammar is a theory of this attained competence... The grammar determines sound-meaning relations.

   Now the interesting thing is that although books on linguistics are so abundant that one could soundproof walls with them, no sentence-meaning pairs, let alone systems for generating such pairs, are to be found in them. Given the programmatic statements like the one by Chomsky, this state of affairs is certainly surprising enough. But it is not quite what I have in mind when I speak of the scandal of linguistics. Maybe the task of matching sentences with their meanings is so tricky that, despite sustained effort on the part of hundreds of practising linguists, the business has not yet gotten off the ground.

   Closer inspection of the literature soon reveals, however, that this is not how the situation is perceived by the linguists themselves. On the contrary, they seem to be quite sanguine, even self-congratulatory, about the state of their discipline. It is this complacency in the face of total failure that I find scandalous.

   One factor responsible for this bizarre situation is the endemic view that linguistic research can be compartmentalized into two separate domains; syntax, the study of the way words are put together, and semantics, the study of meaning. The former, according to the view, can be pursued quite independently of the latter. Chomsky, who, as we have seen, is one of those who have coined the definition of grammar as a generator of sentence-meaning pairs, is also a proponent of something called autonomous syntax, a research program based on the thesis 'that the question of what the syntactic structure of a sentence is entirely independent of the question of what its semantic structure may be'. There is a series of books edited by Chomsky which has a general preamble saying in so many words

preference is given to ... research within the framework of the thesis ... that syntax can be essentially defined without reference to interpretation.

previous page

next page

contents

   It is not easy to see how anybody can see an advantage in separating the study of syntax from the study of meaning. The composition of any artefact is dictated by the purpose it is to serve. The shape of a piece of clothing, for example, is dictated by the purpose that it is to cover a certain part of the human body. Clearly no theory of clothing can explain why a shirt, say, takes the geometric shapes it does, without reference to that fact. By the same token, one would have thought, the syntactical shape of a sentence is dictated by the fact that it encodes a definite message. No theory of language can explain why a sentence like 'Jade shaved Fred' takes the shapes it does without reference to the fact that it serves to convey a message whereby a definite relationship, that of shaving obtained between two individuals, Jade and Fred.

   The same applies even if one accepts Chomsky's view that a language is not entirely an artefact and that sentences are products of what he calls a 'language organ', something that has been fashioned by biological evolution. Any explanation of why the sentences produced by the 'language organ' take the shape they do will clearly have to refer to their adaptive function, just as any explanation of the shape of a bird's wing will have to refer to the fact that it serves to keep the bird aloft. Since the adaptive function of a sentence is clearly to serve as a vehicle of information exchange between members of the species, any explanatory account of it must refer to that function.

   A major source of this cageiness concerning meaning is the absence of any consensus as to what sort of things meaning are. Indeed there is no universal consensus as to whether there are such things as meanings in the first place. Here much damage has been done by philosophers and their frivolous games at denying the obvious. Twentieth-century philosophers vie with one another in propounding ever bolder theses of the form 'There are not such things as ...'. Values, sensations, times, causes, laws of nature, possibilities, numbers, attributes, all have been targets of parsimonious crusades by sanguinary philosophers wielding Occam's Razor. At a time when number-crunching machines are spearheading a second industrial revolution, philosophers have the nerve to publish books with titles like 'Science Without Numbers'. It would not surprise me if a sequel to that book appeared in the bookshops entitled 'Zoology without animals'. Meanings, needless to say, have not

previous page

next page

contents

been spared the carnage. One of the most influential living philosophers informs us that 'there are not such things as meanings over and against their verbal embodiments'.

   One can be excused for thinking that a philosopher who declared that there are no such things as meanings can hardly avoid the consequence that all linguistic expressions are therefore meaningless. This, however, is a somewhat embarrassing consequence to embrace because it makes the philosopher's own declaration that there are no meaning, meaningless. So to prevent malicious critics from making this obvious point the philosopher defiantly proclaims his right to have it both ways: no contradiction is involved, he heroically declares, in saying both that there are no such things as meanings and that a given sentence is meaningful. One may as well argue that no contradiction is involved in saying both that there are no such things as cars and that Fred is a car-owner. What is surprising is not that philosophers make such declarations but rather how often they get away with it.

   Most linguists stop short of embracing this fundamentalist semantic nihilism, sensing rightly that by doing so they might do themselves out of a job. But they have been left with a lingering feeling that there is something not quite kosher about the notion of meaning, and that they will be best advised to steer clear of the subject. If you ask a linguist what meaning is, don't expect a straight answer. Be prepared instead for a lengthy sermon designed to talk you out of asking the question. Jerold Katz, for example, is of the opinion that

the question 'What is meaning' does not admit of a direct 'this and that' answer; its answer is instead a whole theory. ...[I]t is a theoretical question, like 'What is matter?', 'What is electricity?', 'What is light?'

Katz makes it sound as if by propounding a theory one is somehow exempt from the task of giving 'this and that' answer. Yet the accepted theory of electricity would be hardly illuminating if it did not yield the 'this and that' definition of electricity as a stream of electrons, and the accepted theory of light would be hardly illuminating if it did not yield the 'this and that' definition of light as electromagnetic waves in a certain frequency band.

previous page

next page

contents

   Bent Jacobsen suggests that we get rid of the obnoxious question by 'breaking it down':

[Transformational-generative semanticists have argued from the beginning that the question 'what is meaning?' asked at one swoop is not likely to result in an adequate semantic theory. Rather, this question must be broken down into smaller, more modest ones of the following kind: 'What is the sameness of meaning (i.e. synonymy)?' --'What is semantic ambiguity?' --'What is contradictoriness?'

One may well wonder what makes the said semanticists a priori certain that an answer to the question 'What is meaning?' would yield an inadequate semantic theory. It would seem that once that question were satisfactorily answered, all the 'more modest' questions would be easily answered in its wake. Once we knew what meanings were, synonymy, ambiguity, etc. would present no further problems. We could simply say that expressions are synonymous if they have the same meaning and that an expression is ambiguous if it is laden with at least two different meanings. Nor is the conceptual situation symmetrical: merely knowing that two expressions are synonymous does not by itself tell one what either of its readings. Suppose that you want to learn a foreign language, say Swahili. Your instructor gives you a Swahili thesaurus of synonyms and homonyms, and you learn it by heart. Clearly you still don't understand a single word of Swahili. A mere theory of synonymy and ambiguity hardly amounts to a theory of meaning.

   Not to put too fine a point on it, all these evasions are so many attempts to conceal the fact that when asked what meanings are the linguists are simply out of their depth. This, however, should not worry them too much. Strictly speaking, the question 'What is meaning?' is one for a philosopher to answer, not for a linguist. So perhaps instead of judging the linguists by what they say when they wax philosophical and theorize about meaning in general, we should attend to what they do when they actually ply their trade. After all, quite a few of them describe themselves as semanticists (to set themselves apart from the autonomous

previous page

next page

contents

syntacticians), and offer what they call semantic analyses of English sentences. What notion of meaning can be distilled from the way they go about this task?

   The methods used by various authors differ greatly in detail but there is something they all have in common. They are all based on the unspoken assumption that to give the meaning of an English expression is to produce another expression.

   Some years ago I supervised an M.A. dissertation openly defending the thesis that the meaning of a sentence is another sentence. The student's line of argument was very straightforward. When I ask you, he argued, what the meaning of a sentence is, what do you do? You utter another sentence. For example, if I ask you, what the sentence 'Bovine are angulate' means, you say, 'Cow have hooves.'. Does this not show that the meaning of a sentence is another sentence?

   I did my best to talk my student out of this. I pointed out that by the same token it could be argued that the United States is presided over by a couple of words, for if someone in the know is asked who the U.S. president is, he responds by uttering two words:'George Bush'. It is true enough that one cannot say anything without using words, but surely it does not follow that we never speak about anything other than words. In particular, when I say that bovine are angulate means that cows have hooves I am not speaking about the sentence 'Cow have hooves'. Rather, I am exploiting of my interlocutor's familiarity with its meaning to tell him that the meaning of the sentence 'Bovine are angulate' is the same.

   I also made the following point to my student. If the meaning of the sentence 'Bovine are angulate' is the sentence 'Cows have hooves', what is the meaning of 'Cows have hooves'?, 'Bovine are angulate'? That would mean that two sentences have different meanings. Yet the very usefulness of the sentence 'Cows have hooves' in explaining what 'Bovine are angulate' means seems to turn on the fact that they have the same meaning. But nothing washed with the student and I soon gave up trying to change his mind.

   Generative linguisticians do not subscribe to my student's view officially but they do so in practice. When it comes to assigning meaning to a sentence they simply offer another sentence as a translation. Those who follow Chomsky call the translation the 'logical form' of the given sentence. Others call it its 'deep structure'. But in either

previous page

next page

contents

case they refer to a linguistic expression: a string or a tree of symbols of an artificial language.

   So far the linguists have produced precious few concrete translations of this sort and hardly anything in the way of a general method for associating 'logical forms' or 'deep structure' with English sentences. All we have are a few disconnected examples and a patchwork of heuristic hints and apercus. Some might argue, however, that there is nothing wrong with the very programme, provided that the meanings of the symbolic expressions have bee seen as assigning meanings to English expressions indirectly by way of their translations into the symbolic formalism. They can be construed as proceeding after the pattern of the bovine-are-angulate case I have mentioned a minute ago: to indicate the meaning of an unfamiliar sentence, they produce another sentence whose meaning is the same and familiar. Or so some might be inclined to argue.

   The question is, however, are the meanings of the symbolic formulas familiar? The linguists do not consider it part of their brief to inquire into this matter. They borrow the formal languages, lock, stock, and barrel, from logicians and presume that the logicians know perfectly well what the symbolic formulas of those languages

mean. Is the presumption correct?

   We have seen that there are two different senses in which one can be said to know the meaning of a sentence. One can know it in the sense of understanding the sentence, that is, in the sense of knowing what it takes for the sentence to be true. In this sense, the logician (or any layman who has been through a logic course) knows what the symbolic formulas mean. This, however, is not the sense which is relevant here. For in this sense we all understand the English sentences which prompt the linguist's inquiry in the first place. If we did not understand English sentences in this sense we could not use them as vehicles of communication. We do not need a linguist to translate it into the logician's formalese in order to appreciate the meaning of an English sentence in this sense.

   Not so long ago philosophical logicians espoused the view that the meaning of a

sentence simply is its truth condition. They believed that to know the meaning of a sentence is tantamount to knowing what the world would have to be like if the sentence were to be true. The idea is that the world we live in, or, as it is also called,

previous page

next page

contents

the actual world, is only one of a whole range of possible worlds, possible all-encompassing states of affairs. And so it undoubtedly is. The world we live in is one in which George Bush is the U.S. president, for example. But many people find this circumstance deplorable and wish that somebody else, say Michael Dukakis were president. This wish would be nugatory if there were no such possibility as Dukakis' being the president instead of Bush. But of course there is such a possibility: George Bush is not the U.S. president by necessity he only happens to be the U.S. president, which is just another way of saying that there are possible, or conceivable, states of affairs, in which he is not. It is customary to speak of such conceivable, albeit unrealized, states of affairs somewhat dramatically as possible worlds.

   Every English sentence divides the class of possible worlds into those in which it is true, and those in which it is false. The idea underlying truth-conditional semantics was to identify the meaning of a sentence with the division, or dichotomy, it effects.    The idea seems plausible enough as long as one confines oneself to sentences reporting matters of fact, like the sentence 'Bush is the U.S. president.' For example, it gives a neat account of the fact that the sentences 'Bush is the U.S. president' and 'Bush is the husband of Barbara Bush,' although both true, differ in meaning. The explanation is this. Although the two sentences have the same truth-value in the actual world, there are many possible worlds where their truth-values diverge: possible worlds in which George Bush is the U.S. president but is not married to Barbara and possible worlds in which he is married to Barbara but is not the U.S. president. Hence the sentences, although both true, cut the pie of possible worlds in two different ways. And this, according to truth-functional semantics, is just another way of saying that their meanings are distinct.

   It has been known for some time, however, that the equation of meanings with

truth-conditions won't do. For one thing, sentences are not the only expressions

endowed with meaning. Parts of sentences are also meaningful: noun phrases, such as 'the U.S. President', verb phrases, such as 'shaved Mary' are not empty sounds; yet they are not the sort of expressions that can be assigned truth-conditions.

   But the equation of meanings with truth-conditions is indefensible even as applied to sentences. This is because there are many sentences which obviously differ in meaning and yet have the same truth-conditions. It is enough to take any two distinct

previous page

next page

contents

mathematical truths, such as '1+1=6/3' and '7 is a prime number'. These are clearly not synonymous, they are not alternative ways to say one and the same thing. Yet their truth-conditions are identical: both are true come what may. Or, to put it in the possible-worlds terminology, both are true in exactly the same possible worlds, namely, in all of them without exception.

   Nor is this the peculiarity of the necessary truths that are the stock-in-trade of the

mathematicians. Consider the sentences 'If Fred sold half of his cars, he would have only one car left' and 'If Fred doubled the number of cars he owns he would have four cars'. These are not necessary truths, they are not synonymous, and yet their truth-conditions are identical: both are true in and only in possible worlds where Fred owns exactly two cars.

   The examples show clearly that meanings are much finer-grained affairs than truth-conditions. What sort of entities are they then?

   Consider the statement 'Fred calculated 1+1'. Here Fred is being related not to the expression '1+1', but to what the expression signifies. Now if the expression signified the number two then the expression '6/3' would also signify the number two and it would be impossible for Fred to calculate 1+1 without also calculating 6/3. Since it clearly is possible, '1+1' must stand for something finer than a number. And it is not difficult to see what this finer thing is. It is a definite calculation, that of adding the number 1 to itself. It is this calculation which the statement 'Fred calculated 1+1' relates Fred to. The statement 'Fred calculated 6/3' relates Fred to another calculation, that of dividing the number 6 by three. These calculations are distinct despite the fact that they yield one and the same number, namely 2. Thus it is that it might be true that Fred calculated 1+1 without it being true that Fred calculated 6/3, or vice versa.

   Now consider the statement 'Fred knows that 1+1=2.' Here again, Fred is being related not to the expression '1+1=2', but rather to what the expression signifies. Now if the expression signified its own truth-condition then it would be impossible for Fred to know that 1+1=2 without also knowing that 6/3=2, for the truth-condition of '1+1=2' is the same as the truth-condition of '6/3=2'. Since it clearly is possible, the two sentences must stand for something finer than their truth-conditions. What the sentence '1+1=2' stands for is a definite way of arriving at its truth-condition

previous page

next page

contents

in terms of the numbers one and two and the addition function. What the sentence '6/3=2' stands for is a definite way of arriving at its truth-condition in terms of the numbers six, three, and two and the division function. These are two different ways of arriving at, or constructing, a truth-condition despite the fact that they construct one and the same truth-condition.

   We thus see that linguistic expressions express what may be called constructions:

ways of arriving at objects in terms of other objects. A construction as such is a non-linguistic object. It is best looked upon as a cognitive itinerary: it starts at some given objects and leads, by way of some operations, to another object.

   A linguistic expression is a typographic or audial diagram of such a journey. The way words are combined into the expression depict the way objects and operations are combined into the corresponding construction. In particular, the way words are combined into a sentence depicts a particular construction of the sentence's truth-condition. To give a semantic analysis of a sentence is to specify the construction the sentence codifies. To present a grammar of a language is to describe the general codification method peculiar to that language.

   Now for all the work done recently in the field of linguistics, we are as far from cracking the natural-language code as we have ever been. But as I said before, what I find disturbing is not the lack of progress towards this goal but the fact that linguists do not seem to be worried by it. They don't seem to see their task as that of code-cracking, that is, as that of uncovering the way in which our words hook on to things in the world. They are rather content to stay within the realm of language itself, explaining linguistic phenomena incestuously in terms of each other. They seem to treat langauge as a self-contained game which, like the game of chess, is governed by rules unrelated to anything outside its own domain.

   But language is not a game. It is one the most important weapons in our struggle for survival. Our ability to cope with the world depends crucially on our ability to exchange messages about matters of brute fact. The job of the linguist is to explain how such messages get conveyed trough the medium of natural language. Until progress is made in this direction, the enterprise of linguistics has not really gotten off the ground.

previous page

contents