Words, Meaning and Vocabulary An introduction to modern English lexicology Second edition W £ W ":! ' * jstfedni knihovny PdF MU Bm Howard Jackson Etienne Ze Amvela continuum The word Th 6 VyOrd This chapter covers: • the notion of 'word' in general • word meaning • polysemy - the same word with two or more different meanings • homonymy - two or more words with the same shape • multiword lexemes. 3.1 The notion of word Difficulties in the definition of the word The word defined Characteristics of words Ambiguity in the notion of word 3.2 Word meaning The linguistic sign The word as a linguistic sign Denotation and reference Denotation and sense Denotation and connotation 3.3 Polysemy Definition Problems inherent in the concept of polysemy 3.4 Homonymy Definition Homonym clashes 3.5 Multiword lexemes The lexeme Multiword verbs Idioms 3.6 Summary 57 65 68 72 74 79 3.1 The notion of word The basic question we shall attempt to answer in this section is: 'What exactly do we mean by the term "word" in lexicology?' Before proposing the definition which we intend to adopt in this book, we shall first discuss the difficulties involved in the definition of the word. Then we shall discuss in turn the characteristics of the word and ambiguity in the notion of word. 3.1.1 Difficulties in the definition of the word The term 'word' is used to designate an intermediate structure smaller than a whole phrase and yet generally larger than a single sound segment. However, the word may be defined differently depending on whether we focus on its representation, the thought which it expresses, or purely formal criteria. The first type of definition relies mainly on writing traditions that separate by spaces sequences of letters or characters. It must be pointed out that these separations do not always correspond to functional realities. For example, in 'a new waste paper basket', the first two spaces do not have the same value as the last two because the group 'waste paper basket', although represented by three words, constitutes a semantic unit, while such a unit does not exist for the group 'a new waste'. Consequently, a definition based on writing traditions alone cannot be entirely satisfactory. The second type of definition considers the indivisible unit of thought as the most essential criterion. The main problem faced by this view of the word is that of 'delimitation', which offers three possible alternatives. (a) The word as represented in writing represents a thought unit or a psychological unit: this is the most common case, the easiest to observe and one which, unfortunately, may make us forget the others. Examples are names of objects: table, house; 57 r Words, Meaning and Vocabulary The word abstractions: courage, faith, intelligence; adjectives: tall, short; verbs: eat, sleep. (b) The word forms one block but includes two units of thought: e.g. farmer, rethink, spoonful. (c) The psychological unit exceeds the limit of the graphological unit and spreads over several words; the word is only an element of the real unity, which is then a more complex unit: e.g. all of a sudden, as usual, coconut. The third type of definition stems neither from writing traditions nor from thought units and relies only on purely formal criteria. Bloomfield was the first to suggest a formal definition of the word. He contrasted the word with other significant units: the morpheme or minimal meaningful unit and the syntagme or structure, consisting potentially of more than one word. For him, a minimal form is a morpheme; its meaning a sememe. A form which may occur alone is free. A form which may not occur alone is bound. For example, book and man are free; -er and -ing as in singer and writing are bound forms. A word is a minimal free form. Hence, a word is viewed as a form which can occur in isolation and have meaning but which cannot be analysed into elements which can all occur alone and also have meaning (see Bloomfield 1933/5: 178). From Bloomfield's analysis, it follows that lexis, which consists of an infinite number of elements, excludes relational words or grammatical morphemes. In fact, the latter may be considered as an integral part of the grammatical system of the language. Some of them occur as free forms, e.g. and. by, while others can only be bound forms, e.g. -ing as in sleeping, -est in slowest. A further difficulty in the use of formal criteria is that the word may be defined from the phonological, lexical, and grammatical points of view. For example, the phonological word /'falndz/ and the orthographic word finds correspond to the grammatical word 'third person singular of find'. The word find as the base form without any modification is the lexical word. The lexical word is also referred to as a 'lexeme', e.g. book, fast, open, student. Lexical words are different from grammatical words. The latter are forms like a, an, but, that, which cannot usually occur alone as minimal utterances. For more details on the distinction between lexical and grammatical words, see 1.3.4, and 3.1.2 below. 3.1.2 The word defined Most fluent speakers of English seem to know what a word is. They know, for example, that words are listed in dictionaries, that they are separated in writing by spaces, and that they may be separated in speech by pauses. But it is one thing to identify words and another to suggest a definition that will apply to all types of word in English. We shall not go into a theoretical discussion of how words can best be defined. Instead, we shall propose the following definition, which will serve as a working tool in this book. We shall consider the word as an uninterruptible unit of structure consisting of one or more morphemes and which typically occurs in the structure of phrases. The morphemes are the ultimate grammatical constituents, the minimal meaningful units of language. For example, the different forms of the verb speak, i.e. speak, speaks, speaking, spoke and spoken are separate words grammatically. Equally, grave (noun) and grave (adjective) are different grammatical words. So are the plural, the plural possessive, and the possessive of the word boy, all represented by the pronunciation /boiz/, but spelt boys, boys' and boy's respectively. For further discussion of the different senses of the term 'word', see for example, Matthews (1974: 20-7), Quirk et al. (1985: 68-70), Jackson (1988: Iff), Lipka (1990: 72-3). We shall now come back to the distinction often made between 'lexical' and 'grammatical' words (1.3.4). In most general terms, lexical words are nouns, verbs, adjectives, and adverbs. They have fairly independent meanings and may be meaningful even in isolation or in a series. For example, bottle has an independent meaning; and so does the series boy, break, bottle, stone. On the other hand, a word such as a, with, or a series such as a, the, to, with does not automatically suggest any identifiable meaning. Classes of lexical words contain hundreds or even thousands of members, and they form open classes. For example, the English vocabulary has thousands of nouns, and new items can always be added to the list. Grammatical words are elements like prepositions, articles, conjunctions, forms indicating number or tense, and so on. Such classes contain relatively few members and the addition of new members is rather rare. They constitute closed classes or closed sets. However, the distinction between grammatical and lexical words must be handled with care. There is no clear-cut dividing line between the two types of word. We should, in fact, speak of a continuum ranging from words with semantic content such as exam, students, to words devoid of semantic content such as it and that in a sentence like 'It is obvious that some students will pass this exam.' However, 58 59 Words, Meaning and Vocabulary although prepositions may be classed as grammatical words, they are not completely empty of semantic content. The sentence 'The book is on the table' has quite a different meaning when on is replaced by under, near, off, etc. Similarly, the co-ordinators and, or, and but are not mutually interchangeable, because they are not synonymous. 3.1.3 Characteristics of words The following four characteristics are considered essential in our definition of the word in English. First, the word is an uninterruptible unit. When elements are added to a word to modify its meaning, they are never included within that word. They respect the internal stability of the word and are added either at the beginning as prefixes of the word or at the end as suffixes. For example, the prefix un- and the suffix -able may be added to the words aware and drink and give unaware and drinkable respectively. Note, however, that an affix may also occur not at the beginning or at the end, but simultaneously with the word; we then speak of a suprafix. Compare for example the words 'export (noun) and ex'port (verb); they differ only in the position of the primary stress represented by the symbol ('). The stress pattern may be referred to as a suprafix. (We shall not discuss the use of infixes because such affixes do not occur in English.) The word to which affixes are added and which carries the basic meaning of the resulting complex word is known as the 'stem', which may consist of one or more morphemes. The label 'root' is used to refer to a stem consisting of a single morpheme. Secondly, the word may consist of one or more morphemes. When it consists of one morpheme only, then it cannot be broken down into smaller meaningful units, e.g. dog, hand, man, out, work. These are called 'simple' words, which are typically 'minimum free forms', in the sense that they may stand by themselves and yet act as minimally complete utterances, e.g. in answer to a question. When words consist of more than one morpheme, they may be either complex or compound. Complex words may be broken down into one free form and one or more bound forms: e.g. dogs, happi-Iy, quick-er, work-ing, whereas compound words consist of more than one free form: e.g. birth + day, black + bird, candle + stick, coat + hanger. We also need to mention cases which incorporate the characteristics of both complex and compound words: e.g. gentle-man-ly consists of the compound word gentle + man and the suffix -ly; wind + shield + wip-er consists of the compound word wind + shield and the complex word wip-er. For a general discussion of compounding, see 4.5. However, it is not always obvious whether or not a given sound The word sequence should be considered a morpheme. For instance, should window and woman be said to consist of wir d + ow and wo + man respectively? This would allow us to isolate wind and man as identifiable morphemes, but leave -ow and wo-, which are no longer morphemes in English. The morpheme may have ceased to be recognizable because of linguistic change, as in the case of -ow in window (related to 'eye') or the -fer of refer (from Latin ferre, 'carry'). In these cases, we shall say that unless a word can be completely analysed into morphemes, it should be regarded as unanalysable. It must also be noted that a sound sequence that is a morpheme in some words does not necessarily constitute a morpheme in all its occurrences. For example, the suffix -er is a morpheme in gardener and speaker, but it is not a morpheme in never or consider. Thirdly, the word occurs typically in the structure of phrases. According to the hierarchy adopted in this book, morphemes are used to build words, words to build phrases, phrases to build clauses, and clauses to build sentences. This is the typical mapping of lower level into higher level units. However, in atypical mapping, a higher level unit may be used in a lower level unit. For example, a clause such as who came late may be used like an adjective (word) to modify the head noun man in a sentence such as 'The man who came late was my brother'. We shall still regard such a unit as a sequence of words; it has merely shifted levels. Finally, it is also an important characteristic of each word that it should belong to a specific word class or part of speech (1.3.4). Where the same form appears in more than one class, as frequently happens in English, we regard the various occurrences as separate words (for example, smoke (verb) as distinct from smoke (noun)). It may even bo suggested that a word is defined by two factors: its semantic 'nucleus' and the class to which it belongs. EXERCISE 3/1 Count the number of words in the following sentence. You should ask: 'Which type of word am I to count?' Make a separate count of: (a) orthographic words; (b) grammatical word forms; (c) lexical words (lexemes). In their conceptual world, words are at once containers, tools and weapons, just as in the physical world a bag is a container, a screw-driver is a tool, and a gun is a weapon. (Adapted from McArthur 1998: 38) 60 61 Words, Meaning and Vocabulary f } The word 3.1.4 Ambiguity in the notion of word No matter how careful we are in our definition of the word, we are bound to accommodate a certain amount of vagueness or ambiguity, which is inherent in the very nature of language in general and that of the word in particular. Following Ullmann, the most important sources of ambiguity are: the generic character of the word, the multiplicity of aspects in every word, the lack of clear-cut boundaries in the non-linguistic world, and the lack of familiarity with the referent of the words (see Ullmann 1962: 118ff). We shall discuss these various sources of ambiguity in turn, and close the section with a word on 'emotive overtones'. In most general terms, a word can be regarded as 'generic' when it has 'abstract' reference, i.e. is more schematic, poorer than particular words in its ability to distinguish specific features. Generic terms apply to a wide range of items but tell us little about them. The word mammal, for example, is more generic and therefore more abstract than cow; in the same way, animal is more abstract than mammal. The same relationship exists between the words plant, tree, and acacia, the last one being the least abstract and consequently the least vague. Except for rigorously defined scientific or technical terms, proper names, and for a small number of common nouns referring to unique objects, the words denote, not single items, but classes of things or events bound together by some common element. There is always a certain amount of generalization, which inevitably involves an element of vagueness or some degree of ambiguity. In order to generalize, we must always discriminate between 'distinctive' and 'non-distinctive' features of words. For example, the word table will be used to refer to a certain class of objects, irrespective of certain non-distinctive features such as shape, size, colour, and number of legs. In other words, a table could be square, rectangular, oblong, round, etc.; it could have any dimension and/or any colour we wish to give it; it could have one or more legs, and yet it could still be a table, as opposed to a chair or a bed, for instance; provided it has a certain number of distinctive features which characterize all tables. From the above discussion, it may be said that most English words are generic to a lesser or greater extent. The words we use are never completely homogenous in their meaning: all of them have a number of facets or aspects depending on the context and situation in which they are used and also on the personality of the speaker using them. Consider the word table again. Its significance will vary according to the context in which it occurs; e.g. a table for two in a restaurant, the table of contents in a book, a multiplication table at school, to give money under the table. It will also vary according to the user; e.g. a carpenter, a waiter, a member of parliament, a pupil. With abstract words such as democracy, equality, freedom, immortality, such differences in application are even more marked. Note that even proper names, which may be considered the most concrete of all words, are also subject to such 'shifts in application': only the context will specify which aspect of a person, which phase in their development, which side of their activities we have in mind. Such shifts in application can easily lead to multiple meaning, ambiguity or even in extreme cases, misunderstanding. The nature of the non-linguistic world itself may be a source of ambiguity. In our physical environment we are often faced with phenomena which merge into one another and which we have to divide up into discrete units. For example, the colour spectrum is a continuum; however, each language introduces into it a certain number of more or less arbitrary distinctions. This lack of boundaries is even more conspicuous when we consider abstract phenomena. Such phenomena involve distinctions that are largely imposed, because they have no concrete existence without the linguistic form used to refer to them. For example, while speakers may point at an object to specify the particular shade of green they have in mind, they would have no such alternative in order to specify the particular aspect of the word equality which they have in mind. The specialists are aware of this difficulty and always strive to define their terms clearly and to distinguish them sharply from one another. The precise definition and limitation of abstract words can lead to endless and often heated discussions in philosophical works, in law courts and in conferences. A lawyer will avoid any confusion between crime and offence, a psychiatrist between neurosis and psychosis, and a linguist between morpheme and allomorph. For the layman, however, 'abstract' terms of ordinary language are far less precise than 'concrete' terms. One of the functions of the dictionary is to attempt to establish boundaries between the overlapping uses of words in order to reduce cases of potential ambiguity and misunderstanding. Lack of familiarity with the 'referent' of a word is, of course, a highly variable factor, since it depends on the general knowledge and the special interest of each individual. For example, many university students and lecturers will have rather hazy notions about the meanings of tools and objects which will be perfectly clear to any mechanic. Similarly, town-dwellers will not be familiar with plant names or agricultural terms which will be common knowledge for any gardener 62 63 Words, Meaning and Vocabulary or farmer. Since the vocabulary of any language is open-ended, i.e. new words can always be added to it, and since a single individual, no matter how learned, cannot be a specialist in all fields, nobody can pretend to be familiar with all the words in a given language. This lack of familiarity can therefore be another source of ambiguity and may in some cases cause a serious breakdown in communication. Emotive overtones refer to the use of language either to express emotions or to arouse them in others. This use of language may be opposed to the purely communicative one, which is basically symbolic or referential. We assume that both the communicative and the emotive uses of language contribute to the meaning of all utterances. However, one of these uses may be particularly dominant in a given utterance. For example, some words are used purely for evaluative purposes, e.g. the adjectives good and bad, and the adverb well; but we cannot normally assume that such words have no cognitive meaning. As Delacroix (1924: 41) puts it in his book Le Langage et la Pensee: 'All language has some emotive value: if what I say were indifferent to me I would not say it. At the same time, all language aims at communicating something. If one had absolutely nothing to say one would say nothing' (quoted in Ullmann 1962: 128). In lexicology, the consequence of this dual aspect of language use is that the meaning of any word may be modified by emotive overtones. Although this modified meaning may be considered secondary, it is nevertheless as important as the denotational meaning of the word. In certain contexts, it may even be the determining factor for the most appropriate interpretation of a given message. However, because emotive overtones may be added to any word, this aspect of language does not lend itself to objective analysis. Consequently, we shall not elaborate on the influence of emotive overtones in lexicology. EXERCISE 3/2 What does the word 'bull' mean in each of the following sentences? 1. Beware of the bull! 2. I think the elephant is a bull. 3. Stop acting like a bull in a china shop! 4. There was a bull market on the stock exchange today. 5. Well done! You've hit the bull's eye. 6. Don't give me all that bull. 7. I'm afraid that you'll just have to take the bull by the horns. The word 3.2 Word meaning Before we examine the most common terms used in the discussion of word meaning, we shall first define 'linguistic sign' and then discuss the word as a linguistic sign. 3.2.1 The linguistic sign Following de Saussure (1959), the linguistic sign is a mental unit consisting of two faces, which cannot be separated: a concept and an acoustic image. The term 'sign' is quite a general expression which can refer to sentences, clauses, phrases, words, or morphemes. De Saussure later referred to 'concept' as 'signifie' or 'thing meant' and to 'acoustic image' as 'signifiant' or 'signifier'. These have since become accepted technical terms in modern linguistics. De Saussure pointed out that an alteration in the acoustic image must make a difference in the concept and vice versa. But this view does not appear to take homonyms into account (see 3.4 below). However, since the linguistic sign has both form and meaning, it follows that, when dealing with words, we can focus either on the form or on the meaning. 3.2.2 The word as a linguistic sign Since the word is a linguistic sign, a discussion of 'word meaning' focuses on the relationship between the two faces of the sign, viz. the acoustic image or 'signifiant', i.e. the signifier, on the one hand, and the concept or 'signifie', i.e. the thing meant, on the other. A major difficulty in this task is how to accommodate both the fuzzy nature of meaning and the ambiguity inherent in the notion of word. We cannot go into the intricacies of the various aspects of meaning in an introductory book of this nature. Instead, we shall limit our discussion to an examination of some of the most common terms associated with word meaning; those that will be useful not only in our discussion of the different types of relationship that exist between words, but also in our study of sense relations (Chapter 5). We shall consider in turn denotation, connotation, reference and sense. However, to ease comparison and cross-references, we shall discuss these terms in pairs as follows: denotation and reference, denotation and sense, and finally denotation and connotation. 64 65 Words, Meaning and Vocabulary The word 3.2.3 Denotation and reference We need the concept of 'lexeme' to clarify the distinction between denotation and reference. This concept, which was coined by Lyons in analogy to 'phoneme' and 'morpheme', is considered an abstract linguistic unit (spelt in capitals) with different variants (e.g. SING as against sang, sung). Thus, the relation of denotation holds between a lexeme and a whole class of extra-linguistic objects. For example, Lyons defines the denotation of a lexeme as 'the relationship that holds between that lexeme and persons, things, places, properties, processes and activities external to the language system' (Lyons 1977: 207). It is therefore difficult to give concrete examples of denotation since this relation holds between an abstract linguistic unit and a whole class of extra-linguistic objects. As opposed to denotation, the relationship of reference holds between an expression and what that expression stands for on particular occasions of its utterance (Lyons 1977: 207). Lyons further points out that reference depends on concrete utterances, not on abstract sentences. It is a property only of expressions. It cannot relate single lexemes to extra-linguistic objects, since it is an utterance-dependent notion. Furthermore, reference is not generally applicable to single word forms and it is never applicable to single lexemes (Lyons 1977: 197). For example, expressions such as the computer, John's computer, or the two portable computers on the table may be used to establish a relationship of reference with specific items as referents. In this case, the reference of these expressions containing computer is partly determined by the denotation of the lexeme COMPUTER in the overall system of the English language. 3.2.4 Denotation and sense We have already defined denotation following Lyons (see 3.2.3). His definition of sense also evolved with time. Initially, he defined the sense of a word as its 'place in a system of relationships which it contracts with other words in the vocabulary' (Lyons 1968: 427). Later (Lyons 1977: 206), he defines sense as a relationship 'between the words or expressions of a single language, independently of the relationship, if any, which holds between those words or expressions and their referents or denotata'. It follows that sense is a relationship which is internal to the language system, i.e. a language-immanent relationship. Both individual lexemes and larger expressions have sense. However, the sense of an expression is a function of the sense of the lexemes it contains and their occurrences in a particular grammatical construction. The sense of the word table will vary in the following sentences: 'Don't put your feet on the tahleV and 'It was finalized under the table.' A comparison between denotation and sense shows that the two relations are dependent on each other. According to Lyons, some words may have no specific denotation and still have sense. To use an often quoted example, consider the following pair of sentences: There is no such animal as a unicorn. There is no such book as a unicorn. While the first is perfectly acceptable, the second is semantically odd. Furthermore, this double observation proves that, whereas the lexemes book and unicorn are incompatible, animal and unicorn are somehow related in sense. Such examples can be multiplied easily. The important point here is that a word may have sense but have no denotation. (See further Chapter 5.) 3.2.5 Denotation and connotation Specialists are not in agreement in their treatment of the distinction between denotation and connotation, or denotative vs. connotative meaning. Some, such as Ullmann (1962: 74) make a binary distinction between the two terms. Others, like Lyons (1977: 287) and Leech (1981: 12ff), do not accept the binary distinction and prefer to use denotation and connotation in a rather specific sense. However, these divergent views cannot conceal the fact that connotation is closely associated with synonymy. In other words, synonyms may have the same denotation, i.e. cognitive, or conceptual meaning, but differ in connotation. However, it is sometimes difficult to draw the line between 'connotation', which is essentially stylistic, and 'denotation'. To illustrate this point, Lyons (1977: 618f) argues that in Scots English, the word loch (vs. lake) can be regarded either as a dialectally marked variant of the Standard English word or as a different lexeme, with a specific 'descriptive meaning' or 'denotation', and a connotation of 'Scottishness'. For the supporters of the binary distinction between denotation and connotation, denotative meaning refers to the relationship between a linguistic sign and its denotatum or referent. However, connotations constitute additional properties of lexemes, e.g. poetic, slang, baby language, biblical, casual, colloquial, formal, humorous, legal, literary, rhetorical. Denotation and connotation are both important in order to determine word meaning in a given context. 66 67 Words, Meaning and Vocabulary As a final observation, it must be acknowledged that words are not normally used in isolation, but are combined with other words to form larger units expressing various relationships. These units constitute the linguistic context in which a specific word operates. EXERCISE 3/3 What are the connotations of the italicized words in the following sentences? 1. We are away to sunnier climes. 2. I had to fork out a lot for that present. 3. Give mummy the doggie then. 4. If we do that, they'll call down fire and brimstone on us. 5. Let's get rid of this little beastie. 6. I suppose muggins will have to do it. 7. Let me get you the aforementioned. 8. We'll meet upon the greensward. 3.3 Polysemy We shall first define polysemy in 3.3.1, before discussing some of the problems inherent in the concept of polysemy in 3.3.2. i 3.3. I Definition Polysemy refers to the situation where the same word has two or more different meanings (from Greek poly, 'many' + semeion, 'sign'). For instance, the noun board is said to be polysemous because it may mean: (1] a long thin flat piece of cut wood, (2) a flat surface with patterns, used for playing a game on, (3) a flat piece of hard material used for putting food on, (4) a flat piece of hard material fastened to the wall in a public place to pin notices on, (5) the cost of meals, (6) a committee or association, as of company directors or government officials, set up for a special responsibility [Longman Dictionary of Contemporary English 1978: 105). Similarly, the word flight is defined in at least the following ways: (1) the act of flying, (2) the distance covered or course followed by a flying object, (3) a trip by plane, (4) the aircraft making the journey, (5) a group of birds or aircraft flying together, (6) an effort that goes beyond the usual limits, (7) a set of stairs as between floors, (8) swift movement or passage (see LDOCE 1978: 421). The word In most cases, only one of the meanings of a polysemous word will fit into a given context, but occasionally ambiguity may also arise. For instance, consider the words bat and bank in the following contexts: Look at that bat under the tree. Susan may go to the bank today. Ambiguity results from the fact that bat may mean either 'flying mammal' or 'implement used to hit the ball in cricket', while bank may mean either 'river bank' or 'the place that deals with money'. 3.3.2 Problems inherent in the concept of polysemy Despite its apparent simplicity, the concept of polysemy is complex and involves a certain number of problems. We shall consider in turn the number of meanings, transference of meanings, and difficulty in recognizing polysemy. Since one meaning cannot always be delimited and distinguished from another, it is not easy to say without hesitation whether two meanings are the same or different. Consequently, we cannot determine exactly how many meanings a polysemous word has. Consider the verb ear. Most dictionaries distinguish the 'literal' sense of 'taking in through the mouth and swallowing' and the derived meaning of 'use up, damage, or destroy something, especially by chemical action', which tends to suggest that the verb may have at least two different meanings. However, in the literal sense, we can also distinguish between eating nuts and eating soup, the former with fingers and the latter with a spoon. Moreover, we can talk of drinking soup as well as eating it. It may therefore be said that in this sense at least, eat corresponds to drink, since the latter involves the 'swallowing of liquids'. We can push the analysis further by asking whether eating an orange (which can involve sucking) is the same thing as eating an apple (which involves only chewing). It goes without saying that if we push this analysis too far, we may end up deciding that the verb eat has a different meaning for every type of food that we 'eat'. The above discussion shows that there is no clear criterion for either difference or sameness of meaning. Consequently, it would seem futile to attempt an exhaustive count of the number of possible meanings which a given word may have. The point of view adopted in this book is that the meaning of a given word is bound to vary according to the specific context in a wide semantic field, part of which overlaps with that of other words. For instance, the semantic field of eat overlaps with that of drink when referring to a soup, since you can 68 69 Words, Meaning and Vocabulary either eat or drink a soup, but there is no overlapping when dealing with nuts, since nuts can only be eaten, not drunk. As suggested in the case of the verb eat, a word may have both a 'literal' meaning and one or more 'transferred' meanings, although we cannot determine with precision how many different meanings a given word may have altogether. We shall first discuss metaphor, which is the most familiar kind of transference, before turning to other kinds of transference. The basic difference between metaphor on the one hand and the other types of transference on the other is that metaphor is 'irregular', because it applies to individual lexical items, whereas the other kinds may be considered more 'regular', in the sense that they do not apply just to individual lexical items but to several members of a specific class, e.g. a group of nouns or adjectives. These characteristics will be made more explicit below. The term 'metaphor' refers to cases where a word appears to have both a 'literal' and a 'transferred' meaning. The words for parts of the body provide the best illustration of metaphor. For example, we speak of the hands and face of a clock, the foot of a bed or of a mountain, the leg of a chair or table, the tongue of a shoe, the eye of a needle, etc. Intuitively, we assume that words such as eye, face, foot, hand, leg and tongue apply first to the body, from which they derive their literal sense. This intuition is supported by the fact that the whole set of words applies only to the body, while only some of them can be transferred to certain objects. For instance, the clock has no tongue, the bed no eyes, the chair no feet and the mountain no legs. It should, however, be said that metaphor is rather haphazard not only within specific languages, but also when we compare the use of the same metaphor across languages. It is from these two points of view that metaphor is considered 'irregular'. For example, it may seem obvious that foot is appropriate to a mountain, or eye to a needle, but a look at French will show that, although a mountain also has a 'foot' (French pied), the needle does not have an 'eye', but a 'hole' [trou); furthermore, a clock does not have 'hands', but 'needles' {aiguilles), chairs and tables do not have 'legs' but 'feet' [les pieds de la table/ chaise). The label 'metaphor' can also be applied to other cases of transference, but only in a rather loose sense, because it is not always clear which meaning should be considered literal and which transferred. However, this second kind of transference is fairly productive because it involves the transfer of meaning in a predictable manner. Thus, many adjectives may be used either literally for the quality they refer to or with the transferred meaning of being the source of the quality. For The word instance, in the literal sense, we may say that 'John is sad' (he feels sadness), 'a blanket is warm' (it is of a certain degree of temperature). But in the transferred sense, when we say that a book or film or story is sad, we do not imply that 'it feels sadness', rather, we mean that it causes someone else to feel sad. Note that this possibility of transfer of meaning may result in ambiguity. For instance, a blanket or a coat may be warm in two senses: either that it is of a certain temperature as mentioned above, or that it keeps one warm. Similarly, many nouns may have a concrete and an abstract meaning. Thus, we may compare 'The thesis is on the desk' and 'The thesis is not supported by objective evidence'. The word thesis has, of course, a concrete meaning in the first sentence and an abstract one in the second. Similar contrasts may be established for bible, book, score and table, for instance. When we refer to the difficulty in recognizing polysemy, we are dealing with the relationship between polysemy (i.e. one word with several meanings) and homonymy (i.e. several words with the same shape - spelling and/or pronunciation). The problem is to decide when we have polysemy and when we have homonymy. In other words, assuming that we have a written form with two meanings, should we consider it as one word with different meanings (polysemy), or as two different words with the same shape (homonymy)? Dictionaries have to decide whether a particular item is to be handled in terms of polysemy or homonymy, because a polysemous word will be treated as a single entry, while a homonymous one will have a separate entry for each of the homonyms. For a discussion of how dictionaries decide, see 8.3. As a final observation, it must be said that far from being a defect of language, polysemy is an essential condition for its efficiency. If it were not possible to attach several senses to the same word, this would mean a crushing burden on our memory; we would have to possess separate terms for every conceivable 'object' we might wish to talk about, and be absolutely precise in our choice of words. Consequently, polysemy must be considered an invaluable factor of economy and flexibility in language. EXERCISE 3/4 List all the meanings that you can think of for the lexemes lemon (noun) and review (verb). Then compare your list with that in the Key to Exercises. 70 71 r Words, Meaning and Vocabulary 3.4 Homonymy Section 3.4.1 below defines homonymy, while 3.4.2 discusses the devices used to avoid homonymous conflicts in English. 3.4.1 Definition Homonymy refers to a situation where we have two or more words with the same shape. Although they have the same shape, homonyms are considered distinct lexemes, mainly because they have unrelated meanings and different etymologies (see 8.3). There is, however, some difficulty in the establishment of 'sameness' of shape, owing to the fact that we do not make the same distinctions in both speech and writing. Thus, lead (metal) and lead (dog's lead) are spelt the same but pronounced differently; while right, rite and write are spelt differently but pronounced the same. For the first case, the term 'homograph' (same spelling) may be used; for the second 'homophone' (same sound) is the appropriate term. In addition to the difference in meaning, homonyms may also be kept apart by syntactic differences. For example, when homonyms belong to different word classes, as in the case of tender, which has different lexemes as adjective, verb and noun, each homonym has not only a distinct meaning, but also a different grammatical function. The same observation applies to pairs of words such as bear (noun) and bear (verb), grave (adjective) and grave (noun), hail (noun) and hail (verb), hoarse (adjective) and horse (noun). 3.4.2 Homonym clashes Because of the sameness of shape, there is a danger of homonymous conflict or clashes in the sense that two homonyms with totally different meanings may both make sense in the same utterance. For example, The route was very long The root was very long; Helen didn't see the bat (animal) Helen didn't see the bat (wooden implement). However, there are at least two different safeguards against any possibility of confusion: the difference in word class and the difference in spelling, besides the difference in overall context. Many homonyms exist only in theory, since in practice there is no risk of any confusion, because they belong to different word classes. The word Consider the pairs of homophones knows (verb) and nose (noun), rights (noun) and writes (verb). Apart from differences in meaning, it is difficult to imagine a context in which both members of a given pair might occur interchangeably. They are in 'complementary distribution', in the sense that where one occurs the other cannot occur. However, it must be specified that since the members of each pair differ in word class, the choice of one homonym instead of the other is determined mainly by the rules of syntax, not those of lexicology. Similar types of restriction also apply to pairs of homonyms which are identical in spelling and pronunciation, e.g. grave (adjective) versus grave (noun), stick (verb) versus stick (noun). This analysis shows that difference in grammatical class contributes to a substantial reduction in the number of 'effective' homonyms in English. However, it must also be acknowledged that difference in class alone does not automatically rule out all possibilities of confusion. English has a non-phonetic writing system, in the sense that there is no absolute one-to-one correspondence between the letters of writing and the sounds in the pronunciation of words. Consequently, spelling will often help to differentiate between words which are identical in sound. This aspect also reduces the number of homonyms on the written and printed page; it may also be useful in spoken language because it provides a quick and easy way of removing confusion. For example, if there is any doubt in the listener's mind whether we mean rite or write, route or root, it may be much simpler to spell the words out than to define their meanings. This discussion of the elimination of homonym clashes shows, among other things, that, in this respect, English writing is more intelligible than speech and that homonymy in the language as a whole, spoken as well as written, is reduced by writing conventions. It also shows that even if we focus on individual words, grammatical and graphological considerations play an important role in the distinction between homonyms. Before we turn to multiword lexemes, it is important to note that there is no clear-cut dividing line between polysemy and homonymy. The major difficulty, as we have seen, is that it is not at all clear how far meanings need to diverge before we treat the words representing them as separate. However, according to Lehrer (1974), the results of experiments suggest that native speakers are generally in agreement over a fair range of examples of homonymy and polysemy, although there is still a considerable residue of borderline cases (quoted in Lyons 1977: 550). 72 73 Words, Meaning and Vocabulary The word EXERCISE 3/5 As a test of Lehrer's thesis, which of the following pairs do you think are homonyms, and which are cases of polysemy? It is perhaps not always as easy as we think. barge - noun (boat), verb (intervene); court - noun (entourage), verb (woo); dart - noun (missile), verb (move quickly); fleet -noun (ships), adjective (fast); jam - noun (preserve), verb (block); pad - noun (thick material), verb (walk softly); steep -adjective (of gradient), verb (immerse); stem - noun (of plant), verb (stop); stuff - verb (fill), noun (material); watch - verb (observe), noun (timepiece) 3.5 Multiword lexemes In 3.1.1 the lexeme was simply referred to as 'lexical word' in opposition to 'grammatical word'. We also gave examples of lexemes consisting of single words. In this section we want to expand on the discussion in 3.1.1 by examining with appropriate examples two types of multiword lexeme. But first, let us revisit our definition of lexeme. 3.5.1 The lexeme Following Crystal (1995: 118), a lexeme or lexical item is a 'unit of lexical meaning, which exists regardless of any inflectional endings it may have or the number of words it may contain'. Crystal adds that 'the headwords in a dictionary are lexemes'. This definition shows clearly that a lexeme may consist of one word, such as big, boy, break, down, quick; but it may also contain more than one word, e.g. away from, brother-in-law, cut down on, hurry up, in front of, switch on, steam iron. For some words, such as adverbs or prepositions, which have no grammatical variants, the headword consists of only one form. But in most cases, the headword is considered as the base form or citation form of the word, from which all the other related word forms may be derived. For example, speak is the lexeme, the base form; while speaks, spoke, speaking and spoken are all derived forms. We shall come back to the way dictionaries treat multiword lexemes (8.3.5). For now, let us discuss two of the main types of multiword lexeme, viz. multiword verbs and idioms. For economy of presentation, the other main type of multiword lexeme, i.e. compounds [steam iron), will be dealt with in Chapter 4. 3.5.2 Multiword verbs In multiword verbs, the main verb and one or two particles can be analysed as constituents of a single unit. Following Quirk et al. (1985: 1150), we shall make a distinction between 'phrasal verbs', 'prepositional verbs' and 'phrasal-prepositional verbs'. We shall first discuss the criteria for the classification of multiword verbs into sub-classes, then we shall examine these sub-classes in turn. We shall use two main criteria in our identification of the different sub-classes of multiword verbs: first, the notion of 'transitivity' and the relative position of the direct object will establish the distinction between prepositional and phrasal verbs; secondly, the number of particles following the main verb will help distinguish between prepositional and phrasal verbs on the one hand and phrasal-prepositional verbs on the other. Prepositional verbs are always followed by an object, i.e. they are all transitive, e.g. call for (John), look at (him). But they are all characterized by the fact that the object cannot occur between the particle and the main verb: hence '*call John for', '*look him at are both ungrammatical (*has its usual meaning of an ungrammatical form or structure). Phrasal verbs may be followed by an object, i.e. they may be either transitive, such as bring up, look up, or intransitive, such as give in, sit down. By contrast with prepositional verbs, transitive phrasal verbs are characterized by the fact that the object may occur between the main verb and the particle without resulting in unacceptable structures, e.g. 'bring them up', 'look John up'. Phrasal-prepositional verbs constitute a bridge class between phrasal and prepositional verbs. Like all prepositional (and some phrasal) verbs, phrasal-prepositional verbs are transitive. But since they can easily be distinguished by the fact that they have two particles, transitivity is not used as a distinctive feature for this sub-group of multiword verbs; e.g. check up on (my friend), get away with (that), stand up for (your rights). Other examples of prepositional verbs include ask for, believe in, care for, deal with, refer to, write about. Such verbs vary in the extent to which the combination preserves the individual meanings of verb and particle. In cases such as ask for and refer to, the meaning of the multiword verb can be derived from that of its constituents. But in cases such as go into (a problem), 'investigate', come by (the book), 'obtain', the multiword verb is best treated as an idiomatic expression. 74 75 r Words, Meaning and Vocabulary Intransitive phrasal verbs consist of a main verb followed by a particle. Most of the particles are adverbials of place. Normally the particle cannot be separated from its verb. Hence, '*broke again down' is ungrammatical. However, particles referring to directions can be modified by intensifiers, e.g. come right back, go straight ahead, go straight on. Other examples of intransitive phrasal verbs are: blow up, catch on, get up, play around, stand up, take off, turn up. This list is enough to show that phrasal verbs vary in the extent to which the combination preserves the individual meanings of the verb and particle. In several cases the meaning of the phrasal verb may be derived from that of its constituents; but in instances such as give in, 'surrender', catch on, 'understand', and turn up, 'appear', it is clear that the meaning of the construction cannot be predicted from the meanings of the verb and particle in isolation. With most transitive phrasal verbs, the particle can either precede or follow the direct object, e.g. They switched on the light. They switched the light on. However, the particle cannot precede personal pronouns, e.g. They switched it on, and not *They switched on it. As a general rule, the particle tends to precede the object if the object is long or if the intention is that the object should receive end-focus. Other examples of phrasal-prepositional verbs include keep out of, stay away from, look down on, 'despise', look up to, 'respect'. Like phrasal and prepositional verbs, some phrasal-prepositional verbs are more idiomatic than others. Some, like stay away from or keep out of, both meaning 'avoid', are easily understood from their individual elements, though many have figurative meaning, e.g. stand up for, 'support', walk away with, 'steal and take away'. Others are fused combinations and it is difficult, if not impossible, to derive the meaning of the multiword verb from that of its constituents, e.g. put up with, 'tolerate', walk out on, 'desert'. There are still others where there is a fusion of the verb with one of the particles. For example, put up with can mean 'tolerate' as in I can't put up with heavy smokers but it can also mean 'stay with' and in that sense put up by itself stands for the unit 'stay'. Similarly, keep up with (the Joneses) 'to compete with one's neighbours or stay level with social changes' may be analysed as consisting of the prepositional verb keep up plus the The word preposition with. Another example is given by the series check, check on, check up on, which consists of three transitive verbs of similar meaning, i.e. 'investigate'. 3.5.3 Idioms Idioms may be treated as a type of collocation (see 5.6) involving two or more words in context. However, since the meaning of an idiom cannot be predicted from the meanings of its constituents, we may also consider idioms as a type of multiword lexeme. Before we discuss their characteristics, we will define and give appropriate examples of idioms in English. In most general terms, an idiom may be defined as a phrase, the meaning of which cannot be predicted from the individual meanings of the morphemes it comprises. For example, when we say that someone 'kicked the bucket', we do not imply that they necessarily hit a certain type of container for liquids with their foot; what we mean is that they died. Similarly, when we say 'Don't beat a dead horse' we do not imply that the carcass of a certain kind of animal is involved; what we mean is that the person should not waste time harping on about an issue that has already been decided. When we say that John used a 'red herring' in his argument, we do not imply that he made use of a specific type of fish called 'herring'; instead we mean that John introduced an irrelevant question to turn attention away from the main issue. Other examples of idiom are 'bury the hatchet', 'come up smelling like a rose', 'have an axe to grind', 'have a bone to pick with somebody', 'have second thoughts', 'hit the nail right on the head', 'hit the sack', 'let the cat out of the bag', 'on the straight and narrow', 'straight from the horse's mouth', 'take the bull by the horns', 'wash one's dirty linen in public'. From the above examples, chosen at random, it is apparent that most idioms are easily recognized as 'frozen' metaphors. However, once they are established as fixed lexical units, frozen metaphors tend to lose their vividness, and speakers often lose sight of their metaphorical origins. For instance, the metaphorical origin of 'kick the bucket' is not readily apparent to most speakers of modern English, and is in any case disputed. Besides 'full' idioms, lexicologists also identify what are called 'partial idioms'. In such idioms, some of the words have their usual meaning while the others have meanings that are peculiar to that particular structure. Thus, in 'red hair' the word hair has its usual meaning because it does refer to the fine filaments growing from the human head; but red is idiomatic in the sense that it does not refer to 76 77 Words, Meaning and Vocabulary the strict colour term. Similarly, in 'to make a bed', a bed is not idiomatic because it does refer to the piece of furniture used to sleep on; however, to make is not used in the usual sense of 'to manufacture'. An interesting set of partial idioms involves the word white, since 'white coffee' is brown in colour, 'white wine' is usually yellow, and 'white people' are generally off-pink. While the words coffee, wine, and people have their usual meanings, white is perhaps idiomatic at least to some degree: it could be interpreted as 'the lightest in colour of that usually to be found'. Black is, of course, used as its antonym for 'coffee' and 'people' (though again neither is black in colour terms); yet it is not used for wine. Thus, we may say that even partial idiomaticity can be a matter of degree and may in some cases be little more than a matter of collocational restriction. For instance, we can say that black collocates with coffee and people but not with wine. (For more on collocation, see 5.1.2, 5.6, and 8.4.3.) In conclusion, what is and what is not an idiom is, then, a matter of degree (Fernando and Flavell 1981). Consider, for instance, 'make up a story', 'make up a fire', and 'make up one's face'. The first expression is used in its literal meaning, the second is a partial idiom, while the last is fully idiomatic. Idioms may be characterized by several features; but they may be summarized under two main headings: ambiguity, and syntactic peculiarities. Since most idioms are constructed from morphemes that are also used non-idiomatically, they may have either a literal or an idiomatic meaning; hence, their ambiguity. For example, if someone 'beats a dead horse', they may in fact be in the process of striking the carcass of an animal (literal meaning); or they may be wasting time discussing a matter that has already been closed (idiomatic meaning). Similarly, if someone 'hits the sack', they could be engaged in striking a sack (literal meaning) or they could be going to bed (idiomatic meaning). However, once a reader or hearer realizes that a given expression is an idiom, even if they are not sure of its exact meaning, they will automatically discard the literal meaning of the expression, and seek an idiomatic meaning (Abdullah and Jackson 1999). Furthermore, the context in which the idiom occurs also plays an important role in the reduction of this potential ambiguity. In addition to the fact that idioms differ semantically from the corresponding strings of morphemes taken in the literal sense, idioms also have special syntactic properties. Consider 'John kicked the bucket'. In the literal sense, this expression has a passive variant, viz. 'The bucket was kicked by John'. But in its idiomatic use, i.e. when 'kicked the bucket' means 'died', the expression does not allow the use of the corresponding passive alternative. Also, as a general rule, in an The word idiomatic expression, none of the words may be replaced by a synonym. For example, in 'we look forward to meeting you', look cannot be replaced by see or watch, for instance. Hence '*we see/watch forward to meeting you' is not idiomatic. Similarly, in 'wash one's dirty linen in public', linen cannot be replaced by socks, neither can thought be replaced by idea in 'have second thoughts'. Furthermore, none of the words in an idiomatic expression may normally be omitted. So, '*we look forward seeing you' is unacceptable because to is omitted. Similarly, '*straight from horse's mouth' and '*turn a new leaf are both unacceptable because the and over have been omitted. There are other syntactic restrictions which characterize idioms. But we need not elaborate on such cases, especially as syntactic restrictions vary from one idiom to another, some idioms being more restricted or frozen than others (see Fernando and Flavell 1981). EXERCISE 3/6 Examine the following definition of 'word' from Collins English Dictionary (1992): one of the units of speech or writing that native speakers of a language usually regard as the smallest isolable meaningful element of the language, although linguists would analyse these further into morphemes. In the light of the discussion in this chapter, what considerations have been ignored by this definition? 3.6 Summary This chapter has shown the central importance of the word in lexicology. In so doing, it has first provided an answer to the fundamental question: 'What exactly is meant by "word" in lexicology?' Secondly, it has examined the notion of 'word meaning', before discussing two of the relationships that may be established between words, viz. polysemy and homonymy. Finally, it has revisited the notion of 'lexeme'; and with appropriate examples, discussed two of the main types of multiword lexeme, viz. multiword verbs and idioms. 78 79