Bold type 1 страницаFor technical terms and for emphasis
PART I Setting the scene CHAPTER I Metalinguistic preliminaries
INTRODUCTION
In this chapter, which constitutes the whole of Part 1, we deal with a number of concepts which are fundamental to the whole enterprise of putting linguistic semantics on a sound theoretical footing. Although it is one of the longest chapters in the book and includes several sections containing material which, at times, is quite demanding for those who are new to the subject, I have deliberately not divided it into two (or more) chapters, because I wish to emphasize the fact that everything that is dealt with here hangs together and is equally relevant throughout. Readers who find some of the material difficult on a first reading should not be too concerned about this. They can come back to it as they proceed through the following three parts of the book and see how the various technical distinctions that alre drawn here are actually used. Indeed, this is the only way of being sure that one has understood them. The fact that I have brought together, at the beginning of the book, some of the more fundamental terminological and notational distinctions which are relevant throughout should make it easier for readers to refer back to them. It should also make it easier for them to see how the conceptual and terminological framework that I am adopting compares with that adopted in other works that are referred to in 'Suggestions for further reading'. We begin and end the chapter with the most fundamental question of all, the question to which semantics, linguistic and non-linguistic, seeks to provide a theoretically and empirically
2 M etalinguistic preliminaries satisfying answer: what is meaning? This question is posed non-technically in section 1.1; in section 1.7, we look briefly at some of the general answers that have been proposed by philosophers, linguists and others in the past and more recently. Between these two sections I have inserted a section (1.2) on what I have called the metalanguage of semantics and a section (1.3) which sets out in greater detail than I have done in the Preface the scope of linguistic semantics. That there should be a section dealing with the relation between linguistic and non-linguistic semantics is only to be expected. It is important that readers should realize that there are various ways in which the subfield of linguistic semantics is defined by specialists as part of the broader fields of semantics, on the one hand, and of linguistics, on the other, and that they should be able to see from the outset the way in which my definition of 'linguistic semantics' differs from that of other authors. The term 'metalanguage' and the corresponding adjective 'metalinguistic', as we shall see in the later chapters of this book, are quite commonly employed nowadays in the discussion of particular issues in linguistic semantics. (The two terms are fully explained in section 1.2.) It is not often, however, that theorists and practitioners of linguistic semantics discuss explicitly and in general terms the relation between the everyday metalanguage of semantics and the more technical metalanguage that they use in the course of their work. I have devoted some space to this topic here because its importance, in my view, is not as widely acknowledged as it ought to be. The next three sections introduce a number of distinctions -between language and speech, 'langue' and 'parole', 'competence' and 'performance'; between form and meaning; between sentences and utterances - which, nowadays, are all more or less generally accepted as part of the linguist's stock-in-trade, though they are not always defined in exactly the same way. Once again, I have given rather more space to some of these distinctions than is customary. I have also sought to clarify what is often confused, especially in the discussion of sentences and utterances, on the one hand, and in the discussion of competence and performance, on the other. And I have explained 1.1The meaning of 'meaning' 3 these distinctions, of course, in the present context with particular reference to their application in semantics (and pragmatics) and to the use that is made of them in the organization of this book.
1.1 THE MEANING OF 'MEANING' Semantics is traditionally defined as the study of meaning; and this is1 the definition which we shall initially adopt. But do all kinds of meaning fall within the scope of semantics, or only some? What is meant by 'meaning' in this context. The noun 'meaning' and the verb 'mean', from which it is derived, are used, like many other English words, in a wide range of contexts and in several distinguishable senses. For example, to take the case of the verb: if one says (1) Mary means well, one implies that Mary is well-intentioned, that she intends no harm. This implication of intention would normally be lacking, however, in an utterance such as (2) That red flag means danger. In saying this, one would not normally be implying that the flag had plans to endanger anyone; one would be pointing out that it is being used (in accordance with a previously established convention) to indicate that there is danger in the surrounding environment, such as a crevasse on a snowy hillside or the imminent use of explosives in a nearby quarry. Similar to the red-flag use of the verb 'mean', in one respect at least, is its use in (3) Smoke means fire. In both (2) and (3) one thing is said to be a sign of something else: from the presence of the sign, a red flag or smoke, anyone with the requisite knowledge can infer the existence of what it signifies, danger or fire, as the case may be. But there is also an important difference between (2) and (3). Whereas smoke is a natural sign of fire, causally connected with what it signifies, the red flag is a conventional sign of danger: it is a culturally established symbol. These distinctions
4 Metalinguistic preliminaries between the intentional and the non-intentional, on the one hand, and between what is natural and what is conventional, or symbolic, on the other, have long played a central part in the theoretical investigation of meaning and continue to do so. That the verb 'mean' is being employed in different senses in the examples that I have used so far is evident from the fact that (4) Mary means trouble is ambiguous: it can be taken like (1) Mary means well or like (3) Smoke means fire. Indeed, with a little imagination it is possible to devise a context, or scenario, in which the verb 'mean' in (4) Mary means trouble can be plausibly interpreted in the way that it would normally be interpreted in (2) That red flag means danger. And, conversely, if we are prepared to suspend our normal onto-logical assumptions - i.e., our assumptions about the world - and to treat the red flag referred to in (2) as an animate being with its own will and intentions, we can no less plausibly interpret (2) in the way in which we would normally interpret (1). Most language-utterances, whether spoken or written, depend for their interpretation - to a greater or less degree -upon the context in which they are used. And included within the context of utterance, it must not be forgotten, are the onto-logical beliefs of the participants: many of these will be culturally determined and. though normally taken for granted, can be challenged or rejected. The vast majority of natural-language utterances, actual and potential, have a far wider range of meanings, or interpretations, than first occur to us when they are put to us out of context. This is a point which is not always given due emphasis by semanticists. Utterances containing the; verb 'mean' (or the noun 'meaning') are no different (from, other English utterances in this respect. And it is important, to remember that the verb 'mean' and the noun 'meaning' are' ordinary words of English in other respects also. It must not be assumed that all natural languages have words in their everyday vocabulary which can be put into exact correspondence with the verb 'mean' and the noun 'meaning' grammatically and semantically. This is a second important
1.1 The meaning of 'meaning' 5 point which needs to be properly emphasized, and I will come back to it later (1.2). Let us now take yet another sense (or meaning) of the verb 'mean'. If one says (5) 'Soporific' means "tending to produce sleep";, one is obviously not imputing intentionality to the English word 'soporific'. It might be argued, however, that there is an essential, though indirect, connexion between what people mean, or intend, and what the words that they use are conventionally held to mean. This point has been much discussed by philosophers of language. Since it is not relevant to the central concerns of this book, I will not pursue it here. Nor will I take up the related point, that there is also an intrinsic, and possibly more direct, connexion between what people mean and what they mean to say. On the other hand, in Chapters 8 and 9 I shall be drawing upon a particular version of the distinction between saying what one means and meaning what one says - another distinction that has been extensively discussed in the, philosophy of language. Intentionality is certainly of importance in any theoretical account that one might give of the meaning of language-utterances, even if it is not a property of the words of which these utterances are composed. For the moment, let us simply note that it is the meaning of the verb 'mean' exemplified in (5), rather than the meaning exemplified in (6) Mary didn't really mean what she said, which is of more immediate concern in linguistics. We have noted that the noun 'meaning' (and the corresponding verb 'mean') has many meanings. But the main point that I want to make in this section is, not so much that there are many meanings, or senses, of 'meaning'; it is rather that these several meanings are interconnected and shade into one another in various ways. This is why the investigation of what is referred to as meaning (in one sense or another of the English word 'meaning') is of concern to so many disciplines and does not fall wholly within any single one of them. It follows that, if semantics is
6 Metalinguistic preliminaries defined as the study of meaning, there will be many different, but intersecting, branches of semantics: philosophical semantics, psychological semantics, anthropological semantics, logical semantics, linguistic semantics, and so on. It is linguistic semantics with which we are primarily concerned in this book; and, whenever I employ the term 'semantics' without further qualification, it is to be understood as referring, more narrowly, to linguistic semantics. Similarly, whenever I employ the term 'language' without qualification, it is to be understood as referring to what are commonly called natural languages. But what is linguistic semantics and how does it differ from non-linguistic semantics? And how do so-called natural languages differ, semantically and otherwise, from other kinds of languages? These are the questions which we shall address in section 1.3. But something should first be said about terminology and style, and more generally about the technical and non-technical metalanguage of semantics. 1.2 THE METALANGUAGE OF SEMANTICS We could have gone on for a long time enumerating and discussing examples of the different meanings of 'meaning' in the preceding section. It we had done so and if we had then tried to translate all our examples into other natural languages (French, German, Russian, etc.), we would soon have come to appreciate the force of one of the points made there, that 'meaning' (and the verb from which it is derived) is a word of English which has no exact equivalent in other, quite familiar, languages. We would also have seen that there are contexts in which the noun 'meaning' and the verb 'mean' are not in correspondence with one another. But this is not a peculiarity of English or of these two words. As we shall see later, most everyday, non-technical, words and expessions in all natural languages are like the noun 'meaning' or the verb 'mean' in that they have several meanings which cannot always be sharply distinguished from one another (or alternatively a range of meaning within which several distinctions can be drawn) and may be somewhat vague or indeterminate. One of the most important tasks that we have to
1.2 The metalanguage of semantics 7 accomplish in the course of this book is to furnish oursejves with a technical vocabulary which is, as far as possible, precise and unambiguous. In doing so, we shall be constructing what semanticists refer to as a metalanguage: i.e., a language which is used to describe language. Now it is a commonplace of philosophical semantics that natural languages (in contrast with many non-natural, or artificial, formal languages) contain their own metalanguage: they may be used to describe, not only other languages (and language in general), but also themselves. The property by virtue of which a language may be used to refer to itself (in whole or in part) I will call reflexivity. Philosophical problems that can be caused by this kind of reflexivity will not be of direct concern to us here. But there are other aspects of reflexivity, and more generally of the metalinguistic function of natural languages, which do need to be discussed. The metalanguage that we have used so far and shall continue to use throughout this book is English: to be more precise, it is more or less ordinary (but non-colloquial) Standard English (which differs in various ways from other kinds of English). And whenever I use the term 'English' without further qualification this is the language (or dialect) to which I am referring. Ordinary (Standard) English is not of course absolutely uniform throughout the world or across all social groups in any one English-speaking country or region, but such differences of vocabulary and grammatical structure as there are between one variety of Standard English (British, American, Australian, etc.) and another are relatively unimportant in the present context and should not cause problems. We have now explicitly adopted English as our metalanguage. But if we arc aiming for precision and clarity, English, like other natural languages, cannot be used for metalinguistic purposes without modification. As far as the metalinguistic vocabulary of natural languages is concerned, there are two kinds of modification open to us: regimentation and extension. We can take existing everyday words, such as 'language', 'sentence', 'word', 'meaning' or 'sense', and subject them to strict control (i.e., regiment their use), defining them or
8 Metalinguistic preliminaries re-defining them for our own purposes (just as physicists re-define 'force' or 'energy' for their specialized purposes). Alternatively, we can extend the everyday vocabulary by introducing into it technical terms which are not normally used in everyday discourse. In the preceding section, we noted that the everyday English word 'meaning' has a range of distinguishable, but interconnected, meanings. It would be open to us at this point to do what many semanticists writing in English do these days: we could regiment the use of the word 'meaning' by deliberately assigning to it a narrower, more specialized, sense than it bears in normal everyday discourse. And we could then employ this narrower, more specialized, definition of 'meaning' to restrict the field of semantics to only part of what is traditionally covered by the term 'semantics' in linguistics, philosophy and other disciplines. In this book, we shall adopt the alternative strategy. We shall continue to use both the noun 'meaning' and the verb 'mean' as non-technical terms, with their full range of everyday meanings (or senses). And for the time being we shall continue to operate with a correspondingly broad definition of 'semantics': until such time as it is re-defined, semantics for us will continue to be, by definition, the study of meaning. It should be mentioned, however, that nowadays many authorities adopt a rather narrower definition of 'semantics', based on the regimentation of the word 'meaning' (or, one of its near-equivalents) in other languages. I will come back to this point (see 1.6). Although the ordinary language word 'meaning' will be retained without re-definition in the metalanguage which we are now constructing, several composite expressions containing the word 'meaning' will be introduced and defined as we proceed and will be used thereafter as technical terms. For instance, later in this chapter distinctions will be introduced between prepositional and non-propositional meaning, on the one hand, and between sentence-meaning and utterance-meaning, on the other; and these will be subsequently related, with various other distinctions, to the distinction that is commonly drawn nowadays between semantics (in the narrow sense) and pragmatics. In Chapter 3, sense and denotation will be distinguished as interdependent aspects or dimensions of the meaning of words and phrases. . 1.2 The metalanguage of semantics 9 Reference will be distinguished from denotation initially in Chapter 3 and then in more detail in Chapter 10. Once again, until they are formally defined 'or re-defined these three terms — and especially the word 'sense' - will be used non-technically. So too will all other words and expressions of ordinary everyday. English (including the nouns 'language' and 'speech' and such seman-, tically related verbs as 'speak', 'say' and 'utter', which will be dealt with in some detail in section 1.4). As will be explained in a later chapter, in recent years linguists and logicians have constructed various highly formalized (i.e., mathematically precise) non-natural metalanguages in order to be able to describe natural languages as precisely as possible. It will be important for us to take a view, in due course, about the relation between the formal, non-natural, metalanguages of logical semantics and the regimented and extended, more or less ordinary, metalanguage with which we are operating. Which, if either, is more basic than the other? And what does 'basic' mean in this context? It is of course written English that we are using as our metalanguage; and we are using it to refer to both written and spoken language, and also (when this is appropriate) to refer to languages and to language-utterances considered independently of the medium in which they are realized. In our regimentation of ordinary written English for metalinguistic purposes, it will be useful to establish a'number of notational conventions, which will enable us to refer unambiguously to a variety of linguistic units. Such more or less ordinary notational conventions as are employed metalinguistically in this book (italics, quotation marks, etc.) will be formally introduced in section 1.5 (see also the list of symbols and typographical conventions on p. xvii). As far as the everyday metalinguistic use of the spoken language is concerned, there are certain rules and conventions which all native speakers follow without ever having been taught them and without normally being conscious of them. But these have not been fully codified and cannot prevent misunderstanding in all contexts. Phoneticians have developed 10 Metalinguistic preliminaries special notational systems for the representation of spoken utterances with great precision. However, in the everyday, nontechnical, use of English (and other natural languages) there is no conventionally accepted written representation of intonation, rhythm, stress and other non-verbal features, which are a normal and essential part of speech. As we shall see later, such features have many communicative and expressive functions. Here, I want to draw attention to the fact that they may also have a metalinguistic function. For example, (7) John said it was raining can be pronounced in various ways. In particular it can be uttered with a characteristic prosodic transition between said and it, which would distinguish in speech what is conventionally distinguished in the written language as (8) John said [that] it was raining (9) John said, "It was raining";. In this case, there is a more or less generally accepted convention - the use of quotation-marks - which serves to distinguish direct from indirect discourse in written English. But there are recognized alternatives to the use of quotation-marks. And even when quotation-marks are used, the conventions for using them are not fully codified or universally accepted: for example, different writers and different printing houses have their own rules for the use of single and double quotation-marks. As I have already mentioned, my own conventions for the metalinguistic use of single and double quotation-marks (arid for the metalinguistic use of italics) will be explained in a later section (1.5). There are many ordinary-language metalinguistic statements which are unambiguous when spoken, but not necessarily when written. Conversely, because there is nothing in normal speech that is in direct one-to-one corespondence with the punctuation marks and diacritics of written language (underlining, italics or bold type for emphasis, quotation-marks, capital letters,
1.3 Linguistic and non-linguistic semantics 11 etc.), there are many ordinary-language metalinguistic statements which are unambiguous when written, but not when spoken. For example, (10) I can't stand Sebastian (11) I can't stand 'Sebastian', in that (10) might be interpreted as a statement about a person whose name happens to be 'Sebastian' and (11) as a statement about the name 'Sebastian' itself. But the conventional use of quotation-marks for such purposes in ordinary written English is not obligatory. And, as we shall see presently, it needs to be properly regimented (as does the use of other notational diacritics) if it is to do the job we want it to do as part of the metalanguage of semantics.
1.3 LINGUISTIC AND NON-L I NGU I STI C SEMANTICSNTICS
The English adjective 'linguistic' is ambiguous. It can be understood as meaning either "pertaining to language" or "pertaining to linguistics". The term 'linguistic semantics' is correspondingly ambiguous. Given that semantics is the study of meaning, 'linguistic semantics' can be held to refer either to the study of meaning in so far as this is expressed in language or, alternatively, to the study of meaning within linguistics. It is being employed here, and throughout this book, in the second of these two senses. Linguistic semantics, then, is a branch of linguistics, just as philosophical semantics is a branch of philosophy, psychological semantics is a branch of psychology, and so on. Since linguistics is generally defined as the study of language, it might be thought that the distinction which I have just drawn between the two senses of 'linguistic semantics' is a distinction without a difference. But this is not so. Linguistics does not aim to deal with everything that falls within the scope bf the word 'language'. Like all academic disciplines, it establishes its own theoretical framework. As I have already explained in
12 Metalinguistic preliminaries respect of the word 'meaning', linguistics reserves the right to redefine for its own purposes everyday words such as 'language' and does not necessarily employ them in the way in which they are employed, whether technically or non-technically, outside linguistics. Moreover, as we shall see in the following section, the English word 'language' is ambiguous, so that the phrase 'the study of meaning in language' is open to two quite different interpretations. There are therefore, in principle, not just two, but three, ways in which the term 'linguistic semantics' can be interpreted. And the same is true of the phrase 'linguistic meaning' (for the same reason). This point also will be developed in the following section. Meanwhile, I will continue to employ the everyday English word 'language’ without specialized restriction or re-definition. Of all the disciplines with an interest in meaning, linguistics is perhaps the one to which it is 6f greatest concern. Meaningful-ness, or semanticity, is generally taken to be one of the defining properties of language; and there is no reason to challenge this view. It is also generally taken for granted by linguists that natural languages are, of their 'essence, communicative: i.e., that they have developed or evolved - that they have been, as it were, designed - for the purpose of communication and interaction and that their so-called design-properties - and, more particularly, their grammatical and semantic structure - fit them for this purpose and are otherwise mysterious and inexplicable. This view has been challenged recently within linguistics and philosophy. For the purposes of this book we can remain neutral on this issue. I will continue to assume, as most linguists do. that natural languages are properly described as communication-systems. I must emphasize however that nothing of consequence turns on this assumption. Although many kinds of behaviour can be described as meaningful, the range, diversity and complexity of meaning expressed in language is unmatched in any other kind of human or non-human communicative behaviour. Part of the difference between communication by means of language and other kinds of communicative behaviour derives from the properties of intentionality and conventionality, referred to in section 1.1. 1.3 Linguistic and non-linguistic semantics. 13 A non-human animal normally expresses its feelings or attitudes by means of behaviour which appears to be non-intentional and non-conventional. For example, a crab will signal aggression by waving a large claw. Human beings, on the other hand, will only rarely express their anger, whether intentionally or not, by shaking their fist. More often, they will convey feelings such as aggression by means of language-utterances such as (12) You'II be sorry for this (13) I'll sue you or (14) How dare you behave like that!. True, the tone of the utterance will generally be recognizably aggressive; and it may also be accompanied with a recognizably aggressive gesture or facial expression. But as far as the words which are used are concerned, it is clear that there is no natural, non-conventional, link between their form and their meaning: as we noted in the preceding section, the words are, in this sense, arbitrary. So too is much of the grammatical structure of natural languages which serves to express meaning. And, as we shall see throughout this book, there is much more to accounting for the semanticity of language - its capacity to express meaning - than simply saying what each word means. It should also be emphasized at this point that, although much of the structure of natural-language utterances is arbitrary, or conventional, there is also a good deal of non-arbitrariness in them. One kind of non-arbitrariness is commonly referred to these days as iconicity. Roughly speaking, an iconic sign is one whose form is explicable in terms of similarity between the form of the sign and what it signifies: signs which lack this property of similarity are non-iconic. As linguists have been aware for centuries, in all natural languages there are words which are traditionally described as onomatopoeic, such as splash, bang, crash or cuckoo, peewit, etc. in English; they are nowadays classified under the more general term 'iconic'. But these are relatively
14 Metalinguistic preliminaries few in number. More important for us is the fact that, although much of the grammatical structure of natural languages is arbitrary, far more of it is iconic than standard textbooks of linguistics are prepared to concede. Most important of all, however, from this point of view, is the partial iconicity of the non-verbal component of natural-language utterances.
|
