Bold type 4 страница

It must not be thought that all languages have the same gram­matical structure. As we shall see later, there are many natural languages which do not have interrogative or declarative sen­tences. This does not mean, of course, that it is impossible to ask questions or to make statements in those languages. Questions might be distinguished from statements, as utterances,,by super­imposing upon the same string of words a distinctive prosodic or paralinguistic contour in speech and distinctive punctuation-marks or diacritics in writing. For example, (29) — the product of the utterance of sentence (26) - may be uttered in the spoken language with a particular intonation-pattern which marks the utterance as a question (indicative perhaps also of surprise or indignation, etc.) and in the written language with a question-mark:

 

(29) Exercise is good for you?

 

But exactly the same sentence can be uttered with a different intonation pattern in the spoken language or differently punctu­ated in the written language (e.g., with a full stop, or period, rather than a question-mark) in order to make a statement.

Sentence-meaning, then, is related to utterance-meaning by virtue of the notion of characteristic use, but it differs from it in that the meaning of a sentence is independent of the particular context in which it may be uttered. To determine the meaning of an utterance, on the other hand, we have to take contextual

40 Metalinguistic preliminaries

factors into account. This point will be developed in greater detail later. But what has been said here will do for the purpose

of organizing the content of this book between Parts 3 and 4.

Meanwhile, the notational distinctions we have adopted will enable us to keep distinct sentences and utterances and to distin­guish the meaning of the sentence itself from the meaning of an utterance which results from the use of that sentence in particu­lar contexts.

 

1.7 THEORIES OF MEANING AND KINDS OF MEANING

 

There are several distinguishable, and more or less well-known philosophical, theories of meaning: theories which seek to pro­vide an answer to the question What is meaning? Among them, one might mention the following:

the referential (or denotational) theory ("the meaning of an expression is what it refers to (or denotes), or stands for"; e.g., 'Fido' means Fido, 'dog' means either the general class of dogs or the essential property which they all share);

the ideational, or mentalistic. theory ("the meaning of an expression is the idea, or concept, associated with it in the mind of anyone who knows and understands the expression");

the behaviourist theory ("the meaning of an expression is either the stimulus that evokes it or the response that it evokes, or a combination of both, On particular occasions of utterance");

the meaning-is-use theory ("the meaning of an expres­sion is determined by, if not identical with, its use in the language");

the verificationist theory ("the meaning of an expres­sion, if it has one, is determined by the verifiability of the sentences, or propositions, containing it");

the truth-conditional theory ("the meaning of an expres­sion is its contribution to the truth-conditions of the sen­tences containing it").

1.7 Theories of meaning and kinds of meaning 41

None of these, in my view, will serve alone as the basis for a com­prehensive and empirically well-motivated theory of linguistic semantics. But each of them has contributed in one way or another to the background assumptions of those who are cur­rently working towards the construction of such a theory. I will not go into the details of any of those theories of meaning listed above. However, I will make reference to some of the key- concepts which distinguish them in the course of the chapters that fpllow, and I will explain these concepts in the context in which they are invoked and applied. Limitations of space will prevent me from discussing the historical connexions among the several theories or the philosophical issues asso- ciated with them. I should add that the list I have given is by no means complete and that the definitions in brackets have in certain cases been deliberately simplified. It is now worth noting that one philosophically defensible response to the question What is meaning? is There is no such thing as meaning. This was the response, for example, of the later Witt­genstein (1953); and it has to be taken seriously. It clearly makes sense to enquire about the meaning of words, sentences and utterances, just as it makes sense to ask what they mean. In doing so, we are using the English words 'meaning' and 'mean' in one of their everyday metalinguistic functions. As we saw earlier, there are also other everyday meanings, or uses, of the noun 'meaning' and the verb 'mean'; and some philosophers at least have held these to be intimately connected with and perhaps more basic than the one that has just been exemplified. Interestingly enough they cannot always be matched one-to-one with the meanings or uses of otherwise comparable expres­sions in such familiar European languages as French, German, Italian, Russian or Spanish. For example, the following two English sentences,

 

(30) 'What is the meaning of 'concept'?'
and

(31) 'What do you mean by the word 'concept'?',

 

might be translated into French as

42 Metalinguistic preliminaries

(30a) 'Quel est le sens de 'concept' [en anglais]?' and

(3la) 'Qu'est-ce que tu veux dire par le mot [anglais] 'con­cept'?'

(and comparably into Italian and Spanish), respectively; into

German as

(30b) 'Was ist die Bedeutung von 'concept' [auf Englisch]?'

and

(31b) 'Was meinst du mit dem [englischen] Wort 'concept'?';

into Russian as

(30c) 'Cto znacit [anglijskije slovo] 'concept'?'

and '

(,31c) 'Cto vy podrazumyvaete pod [anglijskym] slovom 'con­cept'?';

and so on.

In supplying these translations I have not translated the Eng­lish word 'concept', because I have assumed that French, German and Russian are being used metalinguistically with reference to English. There are, of course, other possibilities, especially in the case of (31a-c). In fact, there is a whole range of possibilities, as anyone who has any practical experience of translation will be aware. But we need not go into these in the present context. What these examples show, on the basis of translation into just a few other languages, is that, in each case, the second of the two trans­lated examples, (31a-c), uses an expression which reveals, at least etymologically, a sense of the English verb 'to mean' - utterer's meaning as it is sometimes called - which relates it either to communicative intention (French 'vouloindire', Ger­man 'meinen') or to understanding and interpretation (Russian 'podrazumevatj'). There are those who have seen utterer's mean­ing as being, ultimately, the basis for linguistic meaning.

For the present, however, I am concerned to make the simple point that we cannot inferjthe existence of meaning or meanings from the existence and meaningfulness of the everyday English word 'meaning'. Moreover, even if there is such a thing as

1. 7 Theories of meaning and kinds of meaning 43

meaning (whatever 'thing' means in this context), its onto-logical and psychological status is surely more questionable than that of form. We shall come back to this point.

It was part of the later Wittgenstein's purpose to emphasize the diversity of the communicative functions fulfilled by language. His slogan "Don't look for the meaning, look for the use" (which docs not necessarily lead to the- incauing-is-use theory, though it is commonly so interpreted) must be under­stood with reference to this purpose. Like the so-called ordinary-language Oxford philosophers, such as J. L. Austin (whose theory of speech acts we shall be looking at in Part 4), he pointed out that the question What is meaning? tends to attract answers which are either so general as to be almost vac­uous or so narrow in their definition of meaning as to leave out of account much of what ordinary users of a language think is relevant when one puts to them more specific questions about the meaning of this or that expression in their language.

In this book we are taking a fairly broad view of meaning. We are also assuming that there is an intrinsic connexion between meaning and communication. As was noted earlier, this assump_: tion is not uncontroversial. It has been strongly challenged, for example, by Chomsky, but it is one that is commonly made by philosophers, psychologists and linguists. It enables us to give a better account of the relation between form and meaning in nat­ural languages than does any currently available alternative. And I would emphasize that, although I have referred here to various philosophical theories of meaning and shall draw freely upon them throughout, I am not concerned with philosophical issues as such but with the theoretical and practical problems that arise in the description of natural languages.

So far we have been talking, in a preliminary way, about the meaning of words, sentences and utterances. We have also seen that there are distinguishable senses of the English word 'mean­ing' which may well correspond to different, but related, kinds of meaning. But how many kinds of meaning are there? Are they all of concern to the linguist? And how do they correlate with the distinction we have drawn between lexical meaning and sentence-meaning (including, as we shall see, grammatical

44 Metalinguistic preliminaries

meaning), on the one hand, and between sentence-meaning and utterance-meaning, on the other?

In this book, I will make no attempt to provide a comprehen­sive classification of the different kinds of meaning that a linguis­tic theory of semantics (and pragmatics) should cover. However, it might be helpful to draw even now one very broad distinction which can be developed in more detail later. This is the distinction between descriptive (or prepositional) and non-descriptive (or non-propositional) meaning. (Alter­native terms, more or less equivalent with 'descriptive' and 'prepositional', are 'cognitive' and 'referential'.) With regard to descriptive meaning, it is a universally acknowledged fact that languages can be used to make descriptive statements which are true or false according to whether the propositions that they express are true or false. This fact is given particular prominence in the truth-conditional theory of semantics, which has been extremely important in recent years.

Non-descriptive meaning is more heterogeneous and, in the view of many philosophers and linguists, less central. It includes what I will refer to as an expressive component. (Alternative more or less equivalent terms are 'affective', 'attitudinal' and 'emotive'.) Expressive meaning - i.e., the kind of meaning by virtue of which speakers express, rather than describe, their beliefs, attitudes and feelings - is often held to fall within the scope of stylistics or pragmatics. It will be demonstrated in Part 3, however, that some kinds of expressive meaning are unques­tionably a part of sentence-meaning. It follows from this fact that for anyone who draws the distinction between semantics and pragmatics in terms of the distinction between sentences and utterances, expressive meaning falls, at least in part, within semantics. It also follows, as we shall see in due course, that sentence-meaning is not wholly truth-conditional.

Natural languages vary considerably in the degree to which they grammaticalize expressive meaning. English does so to a relatively low degree. For example, it does not have a rich system of grammatical moods (subjunctive, optative, dubitative, etc.) as many languages do. Like all natural languages, however, it encodes expressive meaning in much of its vocabulary and in

1.7 Theories of meaning and kinds of meaning 45

the prosodic structure of spoken utterances. We are, of course, taking the view (which, as I have noted, is not widely shared) that the meaning of sentences (in contrast with the meaning of utterances) is independent of the prosodic contour with which they are uttered: i.e., that the same sentence can be uttered with various, significantly different, prosodic contours. It can also be argued that exclamatory and contextualizing particles of the kind that one finds in many languages, are not constituents of the sentence, but of utterances which result from the use of the sentence. But expressive meaning is also lexicalized in combi­nation with descriptive meaning, as we shall see, in many ordin­ary nouns, verbs and adjectives.

Other kinds of non-propositional meaning may be left until later. It is worth emphasizing, however, that the expressive func­tions of language cannot be sharply differentiated from their social and instrumental functions. Human beings are social beings with socially prescribed and socially sanctioned purposes. They may not always be consciously projecting one kind of self-image rather than another; they may not be deliberately expres­sing the feelings and attitudes that they do express in order to manipulate the hearer and achieve one goal rather than another. Nevertheless, it is impossible for them to express their feelings and attitudes in language, however personal and spon­taneous these attitudes and feelings might be, otherwise than in terms of the distinctions that are encoded in particular language-systems. As we shall see throughout this book, but more especially in Part 4, expressive meaning necessarily merges with what many authors have referred to as interpersonal, instrumental, social or conative, mean­ing. In other words, as far as the structure and function of natural languages are concerned, the expressive is necessarily socio-expressive and the personal is necessarily interper­sonal. Unless this fact is appreciated, it would seem to be impossible to give a proper semantic account of even such common, though not universal, grammatical categories as tense, pronouns or mood.

 

 

PART 2

Lexical meaning

CHAPTER 2

Words as meaningful units

 

2.0 INTRODUCTION

As we saw in the preceding chapter, it is generally agreed that the words, phrases and sentences of natural languages have meaning, that sentences are composed of words (and phrases), and that the meaning of a sentence is the product of the words (and phrases) of which it is composed.

But what is a word? And do all natural languages, in fact, have words? These questions are not as easy to answer as they might appear to be at first sight. One reason is that the term 'word' is ambiguous, both in everyday usage and also as it is employed technically by linguists. Words may be considered purely as forms, whether spoken or written, or, alternatively, as compo­site expressions, which combine form and meaning. To com­plicate matters further, the term 'form' is employed in several different, though related, senses in linguistics. One of my prin­cipal aims in this chapter is to sort out these different senses of 'word' and 'form' and to establish notational and terminological conventions for avoiding ambiguity and confusion.

Another reason why it is not as easy to say whether something is or is not a word as non-linguists might think - or to say whether all natural languages have words - is that several dif­ferent criteria come into play in the definition of words, both as forms and as expressions, and these criteria are often in conflict. Moreover, some of the criteria employed by linguists, taken ' separately, are such that they do not sharply divide words from non-words.

 

2.0 Introduction 47

In this book, we are concerned primarily with words as expressions: i.e., as composite units that have both form and meaning (more precisely, as we shall see, as units which have, typically, a set of forms and a set of meanings). Whenever the term 'word' is used without further qualification, this is the sense in which it is to be understood. In fact, as will be explained in this chapter, the term 'word' will generally be used through­out this book, and especially in Part 2, to refer to what may be called, non-technically, dictionary-words (or vocabulary-words): i.e., in the sense in which it is used in the everyday meta­language when one says, for example, that a comprehensive dic­tionary of a given language contains, in the ideal, all the words in the vocabulary of that language. In this sense of 'word', all languages do have words.

The technical term that we shall be using for what I have just called dictionary-words is 'lexeme'. The noun 'lexeme' is of course related to the words 'lexical' and 'lexicon'. (We can think of 'lexicon' as having the same meaning as 'vocabulary' or 'dictionary'.) A lexeme is a lexical unit: a unit of the lexicon. The lexical structure of a language is the structure of its lexicon, or vocabulary; and the term 'lexical meaning', which has been used as the title of Part 2, is therefore equivalent to the commonly used, less technical (but ambiguous), term 'word-meaning'. The reasons for extending our metalanguage by introducing the more technical terms 'lexeme' and 'lexical meaning' (in accordance with the principles outlined in section 1.2) will be explained in this chapter. As we shall see, not all words are lexemes and, conversely, not all lexemes are words. We shall also see that, far from being novel or paradoxical, this is something which anyone who consults a conventional diction­ary simply takes for granted, without necessarily reflecting upon its implications for semantic (and grammatical) theory.

When we look at words (and phrases) as meaningful units we also have to deal with the fact that, on the one hand, a single form may be combined with several meanings and, on the other, the same meaning may be combined with several word-forms'. This fact is well recognized in traditional grammar and lexicography and will be discussed here from a fairly traditional

 

 

48 W ords as meaningful units

point of view, in terms of the concepts of homonymy, poly­semy and synonymy.

Finally, as far as this chapter is concerned, we shall look at the distinction between lexical and grammatical meaning, which derives from the distinction that is traditionally drawn between the vocabulary of a language and its grammar. The way in which this distinction is developed and formalized will vary according to the particular theoretical framework within which it is elaborated. There will be a major difference, for example, between formulations of the distinctions that operate with a morpheme-based grammar and those that operate with the more traditional word-based grammar (which we are using). But at the relatively elementary level at which we are discussing the question in this book, nothing is seriously affected by the dif­ferences between these two different models, or theories, of grammatical structure; and it would be a useful exercise for stu­dents who have a sufficient background in grammatical theory, traditional and modern, to check that this is so and to reformu­late what I have to say about form and meaning with reference to morphemes (and combinations of morphemes), rather than words.

As to the effect of adopting a model of linguistic analysis which draws the distinction between the vocabulary (or lexicon) and the grammar at a different place from the place at which it is drawn in traditional grammar and lexicography, this too is rela­tively unimportant in the context of the present book. Adjust­ments can easily be made by those who are familiar with current developments in grammatical theory. The really impor­tant point is that, however one draws the distinction between grammar and vocabulary, in general linguistic theory and in the description of particular languages, the two must be seen as complementary and interdependent. That this is so will be made clear as we move from Part 2 to Part 3.

 

2.1 F O R M S A N D E X P R E S S I O N S

One of the assumptions that was made explicit in Chapter 1 was that the meaning of a sentence depends, in part, upon the

 

 

2.1 Forms and expressions 49

meaning of the words of which it is composed (1.6). This assump­tion now needs to be considered more carefully. We have already noted that the word 'word' is ambiguous: that words may be con­sidered either as forms or as expressions (1.5). Let us begin then by asking in what sense of 'word' it is true to say that sentences are composed of words.

There are, in fact, two quite different distinctions to be taken into account, as we address this question. It is important not to confuse the one with the other. The first is what the American philosopher C. S. Peirce (1839-1914) referred to as the distinc­tion between words as tokens and words as types. This is read­ily explained by means of a simple example. Consider the following sentence:

(1) 'He who laughs last laughs longest'.

From one point of view, it can be said to contain six words: it is six words long. From another point of view, however, it can be said to contain only five words, since two of the words — the third and the fifth (laughs) - are identical: they are different tokens (or instances) of the same type. Put like this, the notion of type/token identity is not difficult to grasp. And, generally speaking, it is clear enough in everyday life when the term 'word' is to be understood in the one sense rather than the other with respect to Peirce's distinction.

There is, however, a second distinction to be taken into account, which is more relevant to our present concerns. This distinction too may be explained by means of a simple example. How many words are there in the following sentence:

(2) 'If he is right and I am wrong, we are both in trouble'?

Once again, there are two correct answers to the question. But the fact that this is so has nothing to do with the type/token dis­tinction (although it is sometimes confused with it in general; works on semantics). It rests upon the difference between words as forms and words as expressions. There are thir­teen forms in the sentence in question, and each of them instantiates (is an instance, or token, of) a different type. From this point of view, however, three of the words — is, am,

50 Words as meaningful units

and are - would traditionally be regarded as different forms of the same word. In one sense of 'word', therefore, sentence (2) is composed of thirteen words; in another, equally common and equally correct, sense of the term, it is composed of only eleven ' words. Let us express this difference in the meaning of'word' by saying that the sentence is composed of thirteen word-forms and eleven word-expressions. It ii word-expressions, not ' word-forms, that are listed and defined in a conventional diction­ary. And they are listed, as we saw in Chapter 1, according to an alphabetic ordering of their citation-forms: i.e., what are com­monly referred to as the headwords of dictionary entries (1.5). In order to assign a meaning to the word-forms of which a sen­tence is composed, we must be able to identify them, not merely as tokens, or instances, of particular types, but as forms of parti­cular expressions. And tokens of the same type are not necess­arily forms of the same expression. For example, in the sentence

(3) 'They have found it impossible to found hospitals or chari­table institutions of any kind without breaking the law',

the third and seventh word-tokens (found) are tokens of the, same type, but not forms of the same expression.

It is the distinction between forms and expressions, rather than the distinction between forms as tokens and forms as types, which I had in mind when I drew attention to the ambiguity of the word 'word'. As I have already mentioned, whenever it is used without further qualification,^;'word' will mean "word-expression", rather than "word-form''; throughout the present work.

Not all the expressions listed in a dictionary, however, are words. Some of them are what are traditionally called phrases; and phrasal expressions, like word-expressions, must be distin­guished in principle from the form or forms with which they are put into correspondence by the inflectional rules of the language. For example, 'pass muster' is a phrasal expression, whose forms are pass muster, passes muster, passed muster, etc. It is tokens of these forms that occur in utterances of the language.

The expressions of a language fall into two sets. One set, finite in number, is made up of lexically simple expressions:

 

2.1 Forms and expressions 51

lexemes. These are the expressions that one would expect to find listed in a dictionary: they are the vocabulary-units of a language, out of which the members of the second set, lexically composite expressions, are constructed by means of the gram­matical (i.e., syntactic and morphological) rules of the language. In terms of this distinction, 'pass muster' is a lexeme, whereas 'pass the examination' is lexically composite. Most word-expressions, in all languages that have words, are lexically simple. However, in many languages, there are productive (derivational) rules for what is traditionally called word-formation, which enable their users to construct new word-expressions out of pre-existing lexically simpler expressions. For example, 'politeness' is constructed from the lexically simpler expression, 'polite', by means of a productive rule of English word-formation. Although many conventional diction­aries do in fact list 'politeness' as a vocabulary-unit (i.e., provide for it a separate entry with its own headword and definition), it is unnecessary to do so, since both its meaning and its grammati­cal properties (as well as its pronunciation) are fully predictable by rule.

Most phrasal expressions, in contrast with word-expressions, are lexically composite. Indeed, all natural languages would appear to contain rules for the construction of an infinite number of lexically composite phrasal expressions. And, as we shall see later, it is an important principle of modern formal semantics that the meaning of all such lexically composite expressions should be systematically determinable on the basis of the mean­ing of the simpler expressions of which they are composed. Lexi­cally simple phrasal expressions (i.e., phrasal lexemes) include, not only such examples as 'pass muster' mentioned above (which has no corresponding lexically composite homonym formed by productive rules of the language), but also idiomatic phrasal lexemes such as 'red herring', which is formally identical with the lexically composite phrase 'red herring' (formed by the productive rules of the language) meaning "herring which is red". The meaning of the lexically simple, idiomatic, phrase (let us call it 'red herring₁'), like that of'pass muster', but unlike that of the lexically composite, non-idiomatic, phrases 'red herring₂'

52 Words as meaningful units

and 'pass the examination', is not systematically determinable by rule from the meaning of its constituent lexemes.

The distinction that has just been drawn between lexically simple expressions (lexemes) and lexically composite expressions is not as straightforward, in practice, as I have made it appear. Just where the distinction is drawn will depend upon the model or theory of grammar with which the linguist is operating. But at whatever point the distinction is drawn between the grammar of a language and its vocabulary (or lexicon), there will always be borderline cases of expressions which can be classified, with equal justification, as lexically simple or lexically composite. But some such distinction is, and must be, drawn in the gramma­tical and semantic analysis of natural languages.

It is lexemes and lexical meaning that will be at the centre of our attention in this and the next two chapters. But forms, in so far as they are forms of particular lexemes, are also of concern to the semanticist. Different forms of the same lexeme will gener­ally, though not necessarily, differ in meaning: they will share the same lexical meaning, but differ in respect of their gram­matical meaning. For example, the forms girl and girls have the same lexical meaning (or meanings); but they differ in respect of their grammatical meaning, in that one is the singular form (of a noun of a particular subclass) and the other is the plural form (of a noun of a particular subclass); and the differ­ence between singular forms and plural forms, or - to take another example - the difference between the past, present and future forms of verbs, is semantically relevant: it affects sen­tence-meaning. The meaning of a sentence, it will be recalled, is determined partly by the meaning of the words (i.e., lexemes) of which it is composed and partly by its grammatical meaning.