THE MEANINGS OF LANGUAGE: SEMANTICS

THE MEANINGS OF LANGUAGE: SEMANTICS

5.1. To understand language we have to know:

- The meanings of words and the morphemes that compose them.

- We must also know how the meanings of words combine into phrases and sentence meanings.

- Finally, we must interpret the meaning of utterances in the context in which they are made. In other words, knowing a language means knowing how to produce and understand sentences with particular meanings.

The study of linguistic meaning of morphemes, words, phrases and sentences is called semantics.

There are several subfields of semantics:

i. Lexical semantics is concerned with the meanings of morphemes and words and the meaning relationships among words.

ii. Phrasal or sentential semantics is concerned with the meaning of syntactic units larger than the word, i.e. phrases and sentences.

iii. Pragmatics: the study of how context affects meaning in certain situations. For example, how the sentence Its cold in here could be interpreted, in a certain context, as Close the window. Pragmatics will be dealt with in chapter 7.

5.2. LEXICAL SEMANTICS (WORD MEANINGS)

The meaning of words is part of linguistic knowledge and is therefore a part of the grammar. Your mental storehouse of information about morphemes and words is what has been called the lexicon.

5.2.1. The function of lexemes

Chapter 2 dealt with the word, with its origins. But is said nothing of the function of the word, of the lexeme. Without a function a word is a mere sequence of sounds.

So what is the function of a word, of a lexeme? In the sentence The daughter of the terrorist has been caught we can identify two distinct types of word.

While we can tell somebody what a daughter or a terrorist is, we cannot tell them what a the or an of is, for while the words daughter and terrorist denote something in the real world, the words the and of do not. What these latter words do is serve the others in some way, the by specifying, of by indicating a relationship.

Daughter and terrorist are content words; the and of are function words.

Lexical content words: in English, nouns, verbs, adjectives and adverbs make up the largest part of the vocabulary. They are the content words of a language, which are sometimes called the open class words because we can and regularly do add new words to these classes. A new verb, download, which means to transfer information from one computer system to another, entered English with the computer revolution. New adverbs like weatherwise and saleswise have been added in recent years, as well as adjectives like biodegradable.

Function words or grammatical words include conjunctions, like and and or, prepositions, like in and of, the articles the and a/an, part of the class of determiners, pronouns. They have been referred to as being closed class words. It is difficult to think of new conjunctions or prepositions or articles or pronouns that have recently entered the language.

The use of content words changes as society changes but the use of function words is much more stable; nouns and verbs may come and go for various reasons but we rarely have reason to change articles and prepositions.

In this chapter our attention will be focused on content words. Function words belong rather to the field of syntax.

5.2.2. The meaning of meaning

Semantics might be described as the study of meaning. But what do we mean by meaning? What is the nature of the relationship between our utterances and the world around us?

Clearly words often specify something in the world. The word daughter denotes a younger female relative as opposed to a younger male relative, an older male relative or indeed, a house. To use terms employed by the Swiss linguist Ferdinand de Saussure, the word daughter is a signifiant, the thing that signifies, and the category of relative concerned is a signifi the thing that is signified.

In their work entitled The Meaning of Meaning C. K. Ogden and I. A. Richards maintained that the word, what they called the symbol, and the actual object, the referent, are linked only indirectly, by way of our mental perception of that object, the thought or reference (Ogden and I. A. Richards, 1985: 10-12).

Stephen Ullmann suggested that we might use more common terms: name to denote the sequence of sounds that is the physical form of the utterance, thing to denote the object or event that is being referred to, and sense to denote the information that the name conveys to the hearer (Ullmann, 1962: 57)

Other linguists have been loath to accept an intermediate conceptual stage in the process of communication. John Lyons, for example, maintains that there is no evidence to suggest that concepts are relevant to the construction of a theory of semantics (Lyons, 1992: 137)

Indeed, seemingly in contravention of their belief in the intermediary of mental perception, Ogden and Richards went on to question the role of mental images:

There are good reasons why attempts to build a theory interpretation upon

images must be hazardous. One of these is the grave doubt whether in

some minds they ever occur or ever have occurred (1985: 60).

5.2.3. Semantic range

In order to be able to deal with the world around us we have to put labels on things. This has required us to divide the world up into categories and to force an infinitely varied world into these categories. We have a category labeled dog into which we put poodles and spaniels but not foxes and wolves. We have to reach a compromise between having on the one hand an unmanageable array of categories and on the other insufficient precision.

The trade-off between manageability and precision clearly depends on how precise one has to be. Arabs may need to be able to make finer distinctions between different types of camels and different types of horses than Europeans do. Eskimos more words than others do to identify different types of snow, for example words defining fine-grained powdery snow, light snow etc.

Similarly, those who specialise in a particular sphere of activity need to make finer distinctions in that sphere than others do. For most of us the word horse and a few others like mare and foal are all that we need to be able to talk about horses, but those involved in horse racing need such terms as filly, yearling and gelding. We could, of course, refer to anything, if only by using what seems more like a dictionary definition than a lexeme: we could denote a filly using the phrase young mare.

But the more significant an object or concept is to a community, the greater the tendency to lexicalize the label used to denote it, to have a more succinct term.

The divisions that we draw in order to define our categories are very haphazard.

The English use the same verb, play, to denote the very different activities of a child amusing himself with his toys and a pianist performing at a concert. And yet the person who supervises a cricket match is given a different name from that given to supervisor of a football match, the one being an umpire and the other a referee.

When the English eat lambs they call the meat lamb, but when they eat calves they use a different word for the meat: veal.

The arbitrariness is clearly likely to be all the greater where there is little natural distinctiveness in the subject area concerned. What the English call a river can be denoted by either fleuve or rivière in French (cf. fluviu or rau in Romanian), the genral view being that the former flows into the sea while the latter flows into another river.

The colour spectrum would appear to be a prime example of arbitrary division. Russian has two equivalents of blue: гoлубoй and c . The Welsh word glas equates to blue but also overlaps with green, as it also denotes the colour of grass.

According to B. Berlin and P. Kay, however, colour terminology is not random for a study by them suggested a fairly universal hierarchy of terminology. If a language only has two colour terms they will, they claimed be based on black and white; If it has three, they will be based on black, white and red. The next distinction will be green or yellow, and so on (Berlin and Kay, 1969).

The set of items that we identify by means of a word or lexeme is the semantic range of that word or lexeme. Such sets may be grouped with others with which they share a common feature to form a semantic field.

Thus the semantic range of the English word red is that part of the colour spectrum that we denote with this word, a range considerably more restricted than that of the corresponding term in a language that distinguishes only three colours. The range of red together with that of the other colour terms can be referred to as the semantic field of colour.

5.2.4. Componential theory / analysis

Componential theory refers to the view that all lexical items can be analysed using a finite set of components (semantic features).

Componential analysis developed from a technique devised by American anthropologists in the 1950s for analyzing the kinship relations of American Indians. Basic to the approach is the assumption that individual items can be decomposed into what are termed semantic primes, or primitives. It is on the basis of these that we organize our experiential world.

Morphemes and words have meanings. The meanings of morphemes and words are defined in part by their semantic properties, whose presence or absence is indicated by the use of semantic features. Suppose someone said:

The assassin killed Thomlison.

If the word assassin is in your mental lexicon, you know that the individual to whom the word refers is human, is a murderer and is a killer of important people. These pieces of information, then, are some of the semantic properties of the word on which the speakers of the language agree.

The meaning of all nouns, verbs, adjectives and adverbs i.e. the content words - , and even some of the function words, such as prepositions (e.g. over, with) can, at least partially, be specified by such properties.

The same semantic property may be part of the meaning of many different words. Female is a semantic property that helps to define:

tigress, vixen, hen

aunt, girl, woman

The words in the second group are also distinguished by the semantic property human which is also found in:

doctor, bachelor, parent, baby, child

The meanings of the last two words are also specified as young. That is, part of the meaning of the words baby and child is that they are human and young.

The same semantic property may occur in words of different categories: Female is part of the meaning of the noun mother, of the verb breastfeed, and of the adjective pregnant.

One way of representing semantic properties is through the use of semantic features. Semantic features are a formal or notational device for expressing the presence or absence of semantic properties by pluses and minuses. Just as in the case of semantic fields, componential analysis, which defines the range of a word in terms of the presence or absence of particular components, is more easily applied to some field than to others.

Words denoting family relationship lend themselves to such a binary approach: words like daughter, son and mother denote either a male person or a female person, either an earlier generation or a later generation, and so on.

One of the commonest examples used by linguists is the set of features which are said to compose the terms woman, bachelor, spinster, wife:

woman [+female] [+adult] [+human]

bachelor [+male] [+adult] [+human] [+unmarried]

spinster [+female] [+adult] [+human] [+unmarried]

wife [+female] [+adult] [+human] [+married]

Isolating the features of these terms allows us to describe more precisely the conceptual sense of words, that is, the stable, or core, meaning which is basic to their individual identity. As a consequence, it enables us to define sense relations more closely.

In the case of hyponymy (inclusion), for example, we can see that spinster is a hyponym of woman because its feature specification contains all the features of woman. This can be expressed in the following way:

A lexical item P can be defined as a hyponym of Q if all the features of Q

are contained in the feature specification of P. (G. Finch, 2000: 155)

Incompatibility can also be dealt with in a similar fashion. The terms spinster, bachelor, and wife are all incompatible and if we look at their feature specification it is possible to see why. In each case they differ from each other in terms of one or more features, despite sharing others, or in formal terms:

Lexical items are incompatible if they share a set of features but differ from

each other by one or more contrasting features. (G. Finch, 2000: 156)

So spinster is incompatible with bachelor because of gender, and with wife because of marital status.

Its important to realize, however, that componential analysis doesnt aim to capture the entire meaning of a word. Its only concerned with conceptual, not associative meaning. So the fact that spinster has a more negative social meaning than bachelor, for example, is irrelevant here.

Despite this, semantic primes are not sufficient on their own to distinguish lexical items. There are, for example, several senses to bachelor, not all of which share the exact features above. In addition to the meaning unmarried male, bachelor can also mean one who has the first or lowest academic degree, and a young seal without a mate in the breeding season.

To cope with this the linguists J. J. Katz and J. A. Fodor (1963) make a distinction between semantic markers and distinguishers.

Semantic markers are meant to reflect systematic relations between an item and the rest of the vocabulary. Distinguishers, on the other hand, reflect what is idiosyncratic about an item. In the case of bachelor, for instance, we could separate out the feature basis of the different senses by listing the markers for each sense followed by the particular distinguisher (as all the senses share adult we can omit this feature here):

bachelor

i.            (human) (male) [one who has never been married]

ii.          (human) [one who has the first or lowest academic degree]

iii. (animal) (male) [young seal without a mate in the breeding season]

We could also set this out in a tree diagram:

BACHELOR

Noun

(human) (animal)

(male) (male)

[one who has [one who has [young seal without a

never married] the lowest degree] mate in the breeding

season]

In the diagram below the ranges of the words aunt, mother, son and daughter are defined by giving each one a positive sign or a negative sign in respect of maleness, previous generation and relationship by birth. This is enough to give each word a unique range.

Male Previous generation Related by birth

Aunt - + -

Mother - + +

Son + - +

Daughter - -

For example, the lexical entries for words such as father, girl, mare, stalk, would appear as follows:

father girl mare stalk

+male +female +female +motion

+human +human -human +slow

+parent +young -young +purposeful

+equine

Another difference between nouns may be captured by the use of the feature [+/-count]. Nouns that can be enumerated one book, two books -, or can be preceded by the quantifier many: - many books are called count nouns.

Nouns such as information, bread which cannot be enumerated or preceded by the quantifier many are called mass nouns. They may be distinguished in the lexicon with one feature [+/-count].

book information

[+count] [-count]

As usual, complications can arise. What components would one identify to define the semantic range of the word terrorist? There is no predetermined system of categories; one keeps identifying distinguishing components until one has a set unique to each word unless one believes that one is dealing with true synonyms or that the difference is one of style, formality, attitude, and so on rather than denotation.

Terrorists are people who use force to achieve a certain aim. But then the same could be said of soldiers and bank robbers. So what other components can we adduce to distinguish between them? Perhaps +authorized for soldiers and - authorized for terrorists in so far as soldiers operate within a framework established by their government whereas terrorists do not. Perhaps other components could be +political for terrorists and -political for bank robbers, thereby reflecting the aims of their actions.

Identifying the semantic features of nouns is easier than for verbs, where there is no actual entity for the process of decomposition to focus on. Most analysis of verbs proceeds by separating them into different semantic classes based on the constructions they allow and then isolating the features which are said to define them. So, for example, B. Levin (1993) isolates the following elements of meaning which material, or action, verbs are said to contain:

(change) (motion) (contact) (cause)

If we take the verbs cut, break, touch, and hit, it is possible to put them into various sets in respect of which components they possess. Thus, B. Levin analyses them in the following way:

cut (cause) (change) (contact) (motion)

break (cause) (change)

touch (contact)

hit (contact) (motion)

We can test for the existence of these components by using sample sentences and seeing whether verbs with particular features will fit. In the case of cut, touch, and hit, all of which possess the feature contact, we can say John cut / touched / hit Mike.

In addition, cut and hit are also verbs of motion and allow John cut / hit at Mike (but not * touched at).

And finally, cut can also be used in constructions which cause a change of state, as in This bread cuts easily (but not *touches / hits easily). In this respect it is similar to break. Levin summarises this by saying:

Touch is a pure verb of contact, hit is a verb of contact by motion, cut is a

verb of causing a change of state by moving something into contact with the

entity that changes state, and break is a pure verb of change of state.

(1993: 9-10)

Using this classification other verbs can be similarly assigned:

(i)           touch verbs: pat, stroke, tickle

(ii)   hit verbs: bash, kick, tap

(iii) cut verbs: hack, saw, scratch

(iv) break verbs: crack, rip, shatter

Cause is the semantic property of some verbs, such as, darken, kill, uglify, etc:

darken: cause to become dark

kill: cause to die

uglify: cause to become ugly

Other semantic properties of verbs are shown in the following table:

Semantic property verbs having it

motion bring, fall plod, walk, run

contact hit, kiss, touch

creation build, make, imagine

sense see, feel, hear

For the most part, no two words have exactly the same meaning, as we shall see in the section dealing with Synonymy. Additional semantic properties make for finer and finer distinctions in meaning. Thus, plod is distinguished from walk by the semantic property slow, i.e. to walk with heavy steps.

The underlying concern of componential analysis is to arrive at a universal inventory of semantic features which are structurally present in all language. Arguable the most consistent proponent of this has been Ray Jackendoff who, in a series of works, has developed a decompositional theory of meaning which he calls conceptual semantics. Jackendoff identifies a number of structural categories, including: Event, State, Thing (or Object), Path, Place and Property.

Loosely speaking, event and state tend to be categories present in verbs; thing /object, in nouns; path and place, in prepositional and adverbial constructions; and property, in adjectives. These categories can all be sub-categorised by reference to specific semantic components. The event category, for instance, can be broken down to include those features of cause, motion, change and contact which we have already identified.

Similarly, thing can be sub-categorised in terms of the features [ + bounded]. This will distinguish between count nouns such as table and chair and mass nouns like music and water. Nouns which are bounded are basically conceived of as units. If we dismantle a chair we cant call the individual pieces a chair. Mass nouns, however, are thought of as substances. If we only hear a few bars of a sonata we have still heard music. This is reflected in the grammar so that mass nouns, for example, cannot go into the plural, e.g. *musics, whereas count nouns can, e.g. chairs.

Note: Music has recently developed a plural form, albeit of a restricted kind, in

the phrase new musics to describe the range of cotemporary styles

available in modern music. (G. Finch, 2000: 110)

Conceptual semantics is a complex and sophisticated attempt to identify universal semantic categories and map them onto syntactic operations and structures. But in so far as it relies on componential analysis it has inevitably had its critics. A principal difficulty with componential analysis is in the identification of the semantic primes, or markers. Pinning down the sense of words is not easy, even for common items. Also, distinguishing between features which are markers and those which are distinguishers is not unproblematic. How can we be sure that we have enough features, or that they are necessarily the right ones? Consider, for example, the semantic components of chair which Katz (1972: 40) gives as:

Chair (object), (physical), (non-living), (artifact), (furniture), (portable)

(something with legs), (something with a back), (something with a seat),

(seat for one)

Clearly some of these features are open to question. Do chairs have to be portable? Is it essential they have legs, or seat only one? Does this analysis cover all chairs or only prototypical ones? Once we begin decomposing, it becomes increasingly difficult to be sure about what counts as a prerequisite feature.

None the less, despite these reservations about componential analysis as a theory of meaning, many linguists use its vocabulary and methodology as a way of examining the linguistic organization of words into groups such as semantic fields and sets.

5 Sense relations

Sense relations refer to the semantic relationships which words contract with each other within the linguistic system on the basis of their sense. The principal relationships are synonymy, antonymy, polysemy, homonymy, hyponymy, meronymy and incompatibility.

5.2.5.1. Synonymy

Synonymy refers to a sense relation that exists between words which have a similar meaning, or sense, for example, drunk /intoxicated, mad /insane. When two words have the same or nearly the same meaning but different sounds they are synonyms.

The semantic range of a word can be defined with the assistance of another word that means the same thing, a synonym or a word that means the opposite, an antonym. Thus it can help us to use the word large if we know that it means much the same as the word big or that it means the opposite of the word small.

English is particularly rich in synonyms because of the influx into it of words from a variety of languages. Royal, regal and kingly, for example, are synonymous terms which have derived from French, Latin, and Anglo-Saxon respectively.

Although, theoretically, it is possible for two words to be completely synonymous, that is to say, identical, it is very rare for this to happen. Total synonymy would mean that words were interchangeable in all linguistic environments. Such is the nature of language, however, that there is invariably some difference.

Synonyms frequently differ stylistically. Steed and nag have the same conceptual sense but belong to different styles of English: the first is poetic and rather archaic, the second slang.

Connotational differences are also fairly common. Hide and conceal, obstinate and stubborn, tight and stingy, are all synonyms but most people would feel that one term in each pair had a slightly stronger meaning than the other.

Many synonyms also differ in their collocational range, that is, the words with which they can co-occur.

It has been said that there are no perfect synonyms, that is, no two words ever have exactly the same meaning. Still, the following pair of sentences have very similar meanings:

He is sitting on the sofa.

He is sitting on the couch.

Some individuals prefer to use sofa instead of couch, but if they know the two words, thy will understand both sentences and interpret them to mean essentially the same thing.

The degree of semantic similarity between words depends to a great extent on the number of semantic properties they share.

Sofa and couch refer to the same type of object (piece of furniture) and share most of the semantic properties.

There are words that are neither synonymous, nor near synonyms yet, have many semantic properties in common: Man and boy both refer to male humans; the meaning of boy includes the additional semantic property of youth, whereby it differs from the meaning of man.

A polysemous word may share one of its meanings with another word, a kind of partial synonymy. For example, deep and profound mean the same when applied to thought, but only deep can modify water.

Due in large part to the overlay of Norman French onto Old English, the English language has an extensive vocabulary. From Old English, for example, the English have the word hide and from Old French they have the word conceal.

We can say

He was determined to hide the truth

and He was determined to conceal the truth.

Thus, as we can use either in this sentence, we might call the words hide and conceal synonyms. But if we try to replace hide by conceal in the sentence

He was determined to hide

we find that we do not get a satisfactory sentence, the reason being that , unlike hide, conceal cannot serve as an intransitive verb, cannot that is, be used without an accompanying object. Thus, the two words are not complete synonyms as they cannot substitute for each other in all circumstances. Thus, we can say that in the case of these two words hide and conceal the difference is one of syntax.

On the other hand, there a difference of reference in the case of the words high and tall. Both may be used to qualify buildings but only tall can be used to refer to the height of people.

5.2.5.2. Antonymy

Antonymy implies a sense relation which exists between words which are opposite in meaning or sense.

The meaning of a word may be partially defined by saying what it is not: male means not female, dead means not alive.

Antonyms have (share) the same semantic properties except for the one that accounts for their oppositeness. Beautiful and tall are not antonyms; beautiful and ugly or tall and short are.

The semantic property they do not share is present in one and absent in the other. In the case of antonyms we have to consider different types of relationship. The words tall and short are opposites, antonyms, as are the words alive and dead. There is, however, a fundamental difference between these two pairs, for while one person can be shorter than another, one person cannot be more dead than another. Thus, tall and short are termed gradable antonyms and alive and dead are termed complementary antonyms. Different again are pairs like buy and sell which denote, for example, two facets of an action: these are converse terms.

Antonymy can take a number of forms:

i. Gradable antonyms are terms in which the degree of opposition is said to be gradable, for example, wide and narrow, old and young, tall and short, big and small, hot and cold, fast and slow, happy and sad. In each of these pairs the opposition is not absolute. There are degrees of width, age, height, etc., so that to say a road is not narrow doesnt mean its wide, and vice versa.

Also, our definition of wide, old, and tall will vary according to the referent. That is, the meaning of adjectives in gradable pairs is related to the object they modify. The words themselves do not provide an absolute scale. Thus, a tall man is shorter than a tall building. Also, we know that a small elephant is much bigger than a large mouse.

With gradable pairs, the negative of one word is not synonymous with the other (as in the case of complementary pairs). For example, someone who is not happy is not necessarily sad.

Gradable antonyms are often found among sets of words that partition a continuum:

tiny small medium large - huge - gargantuan

euphoric elated happy so-so sad gloomy despondent

Another characteristic of certain pairs of gradable antonyms is that one item can be more widely used, or in linguistic terms is unmarked, and the other is marked.

We can say someone is 3 months old or 50 years old without meaning that they are old, but we can really only refer to them as young if they are indeed young. In this case young is more marked than old.

The unmarked member is the one used in questions of degree. We ask ordinarily, How high is the mountain? (Not How low is it?). Thus, high is the unmarked member of high / low.

Similarly, tall is the unmarked member of tall / short, fast is the unmarked member of fast / slow, etc.

ii. Complementary antonyms are different from gradable in that the opposition between them is absolute. Alive and dead, present and absent, awake and asleep, married and single, etc. have an either / or relationship. They are complementary in that to say someone is not alive means they are dead, and vice versa unlike gradable antonyms there arent degrees in between.

Having said that, however, its quite common to find colloquial instances of grading, for example, Hes very much alive.

iii. Relational antonyms or converse terms

They are also different from gradable ones in that they are not susceptible to degrees of opposition. However, unlike complementary antonyms, they are not either / or in character; we could say, for example, that husband is the opposite of wife, but not to be a wife doesnt mean you are a husband. Relational antonyms exhibit reversibility: this is a logical relationship which allows us to say that if I am your husband then you are my wife.

Similar pairs like give / receive, buy / sell, teacher / pupil, employer / employee, above/ below, lend/borrow, etc., are called relational opposites as they display symmetry in their meaning. Thus, the pair give / receive: if X gives Y to Z, then Z receives Y from X.

Or, the pair teacher / pupil: if X is Ys teacher, then Y is Xs pupil.

Pairs of words ending in er and ee are usually relational opposites: if John is Toms employer, then Tom is Johns employee. Also, payer / payee are relational opposites.

Buy / sell are relational opposites because both contain the semantic property transfer of goods or services and they differ in only one property, direction of transfer.

Comparative forms of gradable pairs of adjectives often form relational pairs.

Thus, if :

Tom is taller than Ann, than Ann is shorter than Tom.

If a Cadillac is more expensive than a Ford, then a Ford is cheaper than a

Cadillac

When looking at antonyms we should note that a word can have different antonyms in different contexts: while the opposite of a short person is a tall person, the opposite of a short walk is not *a tall walk.

5.2.5.3. Polysemy

Polysemy refers to a sense relation in which a word, or lexeme, has acquired more than one meaning. That is to say, when a word has multiple (differing) meanings that are related conceptually and historically, it is said to be polysemous.

For example, bear is polysemous, with meanings to carry, to support, to tolerate.

Bear is also a homonym. Homonyms generally have separate dictionary entries: one bear is the polysemous verb just mentioned, the other bear refers to the animal.

The term flight can mean all of the following: (i) the power of flying; (ii) an air journey; (iii) a series of steps. These senses are clearly related and it is possible to see how they might derive from the same word.

Many nouns acquire new meanings by having a literal and a metaphoric meaning, for example, parts of the body, eye, leg, hand, foot, applied to needle, chair, clock, and bed. And some nouns acquire a concrete and an abstract sense. So, thesis can be used to refer to a specific item, as in Ive had my thesis bound, or to a more general one, as in I agree with your thesis.

Also, good is polysemous: it means well behaved in good child and sound in good investments.

Context, indeed, plays a very important part in the definition of the meaning of words. If somebody tells us that a chair is yellow we think of a particular colour; if they tell us that a man is yellow we may think of cowardice.

If they tell us that the chair is comfortable we understand that we would feel relaxed if we sat on it, but if we are told that a man is comfortable we are unlikely to conclude that it would be relaxing to sit on him.

5 .4. Homonymy

When two words have the same sounds (identical pronunciation) but different meanings, they are homonyms (spelling is not relevant, only pronunciation is).

Examples of homonyms are bear and bare [bεə]

Lewis Carrolls Alices Adventures in Wonderland abounds in homonyms:

Mine is a long and sad tale said the Mouse, turning to Alice and sighing.

it is a long tail, certainly, said Alice, looking with wonder at the Mouses tail,

but why do you call it sad?

Homonyms are good candidates for humour as well as for confusion, as in the passage from Alices Adventures in Wonderland:

How is bread made?

I know that Alice said eagerly. You take some flour.

Where do you pick the flower? the White Queen asked. In a garden or in

the hedges?

Well, it isnt picked at all, Alice explained, Its ground.

How many acres of ground? asked the White Queen.

The humour of the passage is due to the two sets of homonyms:

flower and flour and the two meanings of ground. Alice means ground as the past tense of grind, whereas the White Queen is interpreting ground to mean earth.

The use of homonymy may result in ambiguity which occurs when an utterance has more than one meaning. A word or a sentence is ambiguous if it can be understood or interpreted in more than one way. The sentence:

Ill meet you by the bank

may mean Ill meet you by the financial institution or Ill meet you by the riverside.

The ambiguity is due to the two words bank with two different meanings. Sometimes additional context can help to disambiguate the sentence:

Ill meet you by the bank, in front of the automated teller machine.

Ill meet you by the bank. We can go skinny-dipping.

The meaning of the word bank becomes clear when it is preceded by either savings or gravel. The context, then, can do much to determine the meaning of a word.

Some linguists have concluded that one knows a word by the company that it keeps (J. R. Firth) or that the meaning of a word is its use in the language (Ludwig Wittgenstein).

5 Problems of analysing Homonymy and Polysemy

As we have seen, for convenience we force a uniformity on the world around us. The word dog can denote quite different animals. The word lamb can denote an animal or a dish. A persons body can similarly be alive or dead. The word play can denote a piece of writing intended to be performed by actors at a theatre, the activity of a pianist and that of a child.

Thus, the question arises of how extensive the semantic range of a word or lexeme has to be before we feel that we are dealing with two separate words that have the same form rather than with one word that denotes a variety of things.

In the case of lamb we can consider the animal and the dish to be different referents of the one lexeme.

In the case of sound in the sense of something that one hears and sound in the sense of a narrow stretch of water, on the other hand, we are likely to consider that we are dealing with two distinct lexemes that happen to have the same form. In the case of one lexeme with a variety of referents we have an example of polysemy. Two or more lexemes with the same form are homonyms.

Unlike polysemic items, lexicographers will treat homonyms as different items, giving each one the status of a separate headword.

As so often, however, the real world does not fall neatly into our categories. There can be little doubt that in the case of the two referents of the word sound that we referred to we are dealing with different lexemes: one is of Latin origin and the other is of Germanic origin.

So, too, the word bill denoting an account or invoice and that denoting the beak of a bird must be deemed to be different lexemes, the former, a cognate of the French word billet, being of Latin origin and the latter being of Germanic origin.

What about grow used intransitively, as in Your tomatoes wont grow well there, and grow used transitively, as in He grows tomatoes? The Italian equivalents (NB. the Romanian equivalents, too), for example, are distinct in this case: crescere (creste) and cultivare (cultiva), respectively.

Such cases leave the lexicographer with the problem of deciding whether to arrange words under one or several headwords.

Etymology is not always a good guide; as Mott points out, the word bank denoting a financial institution and that denoting the shore of a river, two items that are semantically very distinct, are etymologically related (Mott 1993: 119)

5.2.5.6. Hyponymy

The Definition of Semantic Range: Presented with a colour chart or a box of paints we can select a colour that most people would accept as being red. But there may be disagreement about whether something is red or orange.

Similarly, few of us would deny that a crow is a bird for it is and does what we expect a bird to be ad do; it has two legs and feathers, it flies, it builds a nest and it lays eggs. But there are birds that do not fly and there are creatures such as bats that do fly but are not birds, creatures such as snakes that lay eggs but are not birds. Thus we need a way of determining the boundaries of the semantic range of the words red, bird or any other word if we are to be able to judge when it is appropriate to use it.

We can define a word in terms of what it is not or in terms of what it subsumes. Thus, we can define the semantic range of the word dog either by saying that it is any animal that is not a wolf, a cat, a goat, and so on or by saying that it is any animal that is either a poodle, a terrier, a spaniel, and so on. We can be assisted in this by a hierarchical diagram like the one below:

animal

mammal

cat dog wolf

poodle terrier spaniel

Such a relationship between words whereby more specific terms are arranged under their more general superordinate terms is known as hyponymy.

Hyponymy refers to a hierarchical sense relation which exits between two terms in which the sense of one is included in the other. Terms such as daisy, daffodil and rose all contain the meaning of flower. That is to say, they are all hyponyms of flower. The more general term is called the superordinate or hypernym.

Much of the vocabulary is linked by such systems of inclusion: red is a hyponym of colour, flute of musical instrument, and hammer of tool.

Sometimes a word may be superordinate to itself in another sense. This is the case with animal, as shown in the diagram below. The first occurrence, opposed to vegetable, is the sense contained in the phase the animal kingdom. The second occurrence is synonymous with mammal, and the third with beast.

Living thing

Animal Vegetable

Animal Bird Insect Fish

Animal Human

Hyponymy is a vertical relationship which is fundamental to the way in which we classify things. Most dictionaries rely on it for the provision of definitions (a chair is a type of furniture, a flute is a type of musical instrument, and so on)

The set of terms which are hyponyms of the same superordinate term are co-hyponyms, for example, red, black, and yellow, in the colour system.

Another way of describing the relationship is to say that the individual colours are sisters of the parent term colour. As such they exhibit incompatibility. That is, something cannot be all red and all green, a flute cannot be a violin, or an apple a peach.

The word poodle is a hyponym of the word dog, which in turn is a hyponym of the word mammal, and so on. One can compare this arrangement to the natural history classifications by class, order, family, genus and species; the dog and the wolf both belong to the genus Canis, the dog being the Canis familiaris and the wolf the Canis lupus. The hyponyms of a word define its semantic range. A superordinate like mammal or animal can serve to designate a semantic field.

This meaning relation can also be illustrated by the following set of words: red, white, blue, etc. These are colour words, that is, their lexical representations have the feature [+colour], indicating that all belong to the same class. Such sets of words are called hyponyms.

Similarly, lion, tiger, leopard, lynx have the feature [feline].

The relationship of hyponymy is between the more general term such as colour and feline and the more specific instances of it, such as red, white; or lion, tiger, etc. Thus, red is a hyponym of colour, and lion is a hyponym of feline, or equivalently, colour has the hyponym red, and feline has the hyponym lion.

Complications can, however, arise. Consider the following figure:

mammal

cat dog

dog bitch

Firstly, as the figure shows, we may ascribe different meanings to a word with the result that that word appears at different points on the diagram: we use the word dog to indicate gender as well as to denote the species in general.

Secondly, we cannot put dog and bitch alongside poodle, terrier and spaniel as hyponyms of dog, for, while being a poodle precludes being a spaniel, it does not preclude being a bitch: we are dealing with another dimension.

Love can be presented as a hyponym of emotion. But is it an immediate hyponym or is there something else in between? Rogets Thesaurus, which adopts a hierarchical arrangement like hyponymy, has love as a hyponym of sympathetic affections, which in turn is a hyponym of affections. How does love relate to loyalty or desire? Are they related? If so, are they all on the same level or is one a hyponym of another?

Sometimes there is no single word in the language that encompasses a set of hyponyms. Thus, clarinet, guitar, piano, violin, etc. are hyponyms because they are musical instruments but there isnt a single word meaning musical instruments that has these words as its hyponym.

Clearly some semantic fields lend themselves to arrangement by hyponymy more than others do.

5.2.5.7. Meronymy

Meronymy refers to a sense relation which describes a part-whole relationship between the senses of words. Cover and page, for example, are meronyms of book. Meronymy is similar to hyponymy in reflecting a hierarchical relationship between words. A typical system might be as shown below:

Car

Wheel Engine Door Window Boot

Piston Valve etc.

5.2.5.8. Incompatibility

Incompatibility refers to a sense relation which exists between words in a semantic field where the choice of one excludes the other. So, for example, it is a contradiction to say This instrument is a piano and a violin, since the sense of piano excludes that of violin. Similarly, in the field of fruit we cannot say This fruit is an apple and a banana, or, in colour, This colour is red and black.

Because these words are mutually exclusive members of the same field they exhibit incompatibility. At the same time, however, the exclusion is not a form of antonymy. Apple is not the opposite of banana, nor red of black.

But there are occasions when the distinction between antonymy and incompatibility is a fine one. The terms woman and man, for example, are commonly thought of to be opposites, but not being a woman doesnt necessarily mean you are a man. It is more correct to label woman as incompatible with man. What we are saying, then, is that antonymy will inevitably involve incompatibility, but not the reverse. Senses can be incompatible without necessarily being antonymous.

5.2.5.3. Homographs

Different words are homographs if they are spelt the same /identically and possibly pronounced the same or differently.

Examples of homographs, i.e. words spelt the same, possibly pronounced the same and having different meanings:

trunk: i. of an elephant

ii. for storing clothes

pen: i. the writing implement

ii. the cage

are both homographs and homonyms

lead: i. [li:d] the verb

ii. [led] the metal

are homographs but not homonyms (pronounced differently)

Also, another example of homograph is wind denoting a current of air and wind denoting tortuous movement: the two lexemes may look alike but sound different.

5.2.5.9. Other relations

5.2.5.9.1. Heteronyms

Heteronyms are words spelt the same, but pronounced differently and having different meanings, such as:

sow [sau], meaning pig

sow [sou], meaning to scatter seeds

Other heteronyms are:

dove [d v], meaning bird

dove [douv], the past tense of dive.

Also, bass, bow, lead, wind.

We can summarize these relations among different words, i.e. words different in concept and historical origin:

Homonyms Homographs Heteronyms

Pronounced identically YES MAYBE NO

Spelt identically MAYBE YES YES

5.2.5.9.2. Metonymy

A metonym is a substitute word, i.e. a word used in place of another word or expression to convey the same meaning.

The use of Rome to refer to the Catholic church, or Moscow to refer to the Russian government are examples of metonymy.

Also crown is a metonym for monarchy, (the) brass is sometimes used as a metonym for military officers (especially in American English), the bottle is a metonym for alcoholic drink.

5.2.6. The Human Element of Meaning

So far we have been dealing with meaning as though there were a natural association between word and thing. But, as we saw in sectionThe Meaning of Meaning, it is widely held that the link between word and thing, between symbol and referent, is routed by way of our minds. Certainly life has given each of us a different set of experiences, a different set of attitudes, and these colour our perception of the world around us, affect the impression left by an utterance.

To a child the lexeme guinea-pig is likely to suggest the animal, while to a scientist it might suggest somebody who is subjected to an experiment.

Similarly, due to their cultural backgrounds, Britons may think of Christmas when they hear the word turkey, Americans may think of Thanksgiving.

Here we are dealing with connotation, more subjective links, links of the kind that result from word association exercises in which you are asked to say the first thing that comes into your mind when you hear a particular word. While once again we are not dealing with absolute categories, it is useful to think of the word turkey as denoting a particular kind of bird and connoting a particular festival.

Somebodys tenacious behaviour might be called stubborn or obstinate by those who are obstructed by this behaviour, while sympathizers might call it determined or resolute. As we considered with the word blond, do we attempt to account for the difference as a componential feature, in this case perhaps good and +good.

Our attitudes towards some things may be so strong that we are reluctant to refer to them directly. A word that we are loathe to use may be called a taboo word and its more acceptable substitute may be called a euphemism. In this way Americans are reluctant to refer to a farmyard bird as a cock because of its association with the male genitals and they use the word rooster instead.

We can differentiate between kill, murder and assassinate by reference to such components as +intentional and +political, but what components can we adduce to differentiate between kill and do in? Here again, the difference might be seen to lie in another dimension, in this case on an axis of formality. When in the play Pygmalion by George Bernard Shaw the flower-girl Liza Doolittle, her speech not yet perfected by the phonetician Henry Higgins, says its my belief they done the old woman in it is not only what she says that arouses some consternation but also how she says it, the register that she uses.

Words, then, do more than identify things in the world around us. Ogden and Richards gave what they referred to as a representative list of the main definitions of meaning, that list ranging from An intrinsic property to that to which the Interpreter of a symbol (a) Refers (b) Believes himself to be referring(c) Believes the User to be referring (Ogden and Richards 1985: 186-7).

Geoffrey Leech presents a simpler breakdown of meaning into seven aspects (Leech, 1974: 10-23).

The first is the fundamental denotative or, as Leech refers to it, the conceptual meaning, that which defines the meaning of a lexeme in terms of its constituent features. Thus, to use the componential analysis introduced in section 4.2.2., the conceptual meaning of the lexeme woman can be defined in terms of such properties as +human, -male and +adult. Here, then, we are dealing with a direct link between word and thing.

But, as we observed in section 5.2.2., the intermediary of the human mind often affects the nature of the semantic range of a lexeme. Different people, having been subjected to different experiences in life, have different mental images when they hear a lexeme. When different people hear the lexeme woman such qualities as beauty, compassion, practicality and emotion will feature with differing relative significance. Moreover, the relative significance of features will change as society changes; the association of housewife is much less prevalent in our society than it was fifty years ago. Such association is called connotative meaning by Leech.

A lexeme may be more appropriate in a particular style. To draw from Leechs examples again, cast is associated with literary style while chuck is colloquial. This is stylistic meaning.

Shut up in the sense of being quiet similarly has stylistic meaning, being colloquial, but in so far as it is indicative of a disrespectful attitude on the part of the speaker it also has affective meaning.

Our response to one sense of a lexeme may be affected by another sense of that lexeme. We can for example, scarcely use the word gay in its older sense of merry as it now invokes homosexuality. This is reflected meaning.

Lexemes may have a collocative meaning, requiring that they are used together with certain other lexemes but not with others. Sometimes collocation may be so restrictive that we can guess which word will follow a given word. We can refer to a flock of sheep but not a flock of cows. In some cases the association is so close that we may well be able to anticipate it. Most of us would, for example, expect She has blondto be followed by hair.

Finally Leech refers to thematic meaning, to the emphasis that attaches to a lexeme as a result of the speakers choice of grammatical structure or his use of stress. If in saying the sentence John broke the vase the speaker stresses the name John we understand that the question to be resolved was who broke the vase, whereas if he stresses the word vase we understand that the question was what it was that John broke.

5.3. PHRASAL AND SENTENTIAL SEMANTICS

(PHRASE AND SENTENCE MEANING

5.3.1.Words and morphemes are the smallest meaningful units in language. We have been studying their meaning relationships and semantic properties as lexical semantics.

For the most part, however, we communicate in phrases and sentences. The meaning of a phrase and sentence depends on both the meaning of its words and how these words are combined structurally.

Some of the semantic relationships we observed between words are also found between sentences.

For example, just as two words may be synonyms, similarly, two phrases may be paraphrases. This is possible because:

- they contain synonymous words:

She lost her handbag. / She lost her purse

-the structural differences do not affect meaning:

They ran the bill up. / They ran up the bill.

Similarly, words may be homonyms, hence ambiguous when spoken.

Sentences may be ambiguous because:

- they contain homonymous words, as in:

I need to buy a pen for Tom.

- or due to their structure, as in the sentence:

The boy saw the man with the telescope.

Words have antonyms (opposites); sentences can be negated to express opposites. Thus, the opposite of

He is alive is both: He is dead using an antonym

and: He is not alive using negation

The study of how word meanings combine into phrase and sentence meanings, and the meaning relationships among these longer units, is called Phrasal or sentential semantics, to distinguish it from lexical semantics.

5.3.2. Phrase meaning

5.3.2.1. Although it is widely believed that learning a language is merely learning the words of that language and what they mean, there is more to it than that. We comprehend phrases and sentences because we know the meanings of individual words and we know rules for combining their meanings. Languages have rules for combining the meaning of parts into the meaning of the whole.

Well illustrate this with some semantic rules for adjective-noun combinations (noun-centered meaning) and verb-noun phrase combinations (verb-centered meaning)

Noun-centered meaning

The semantic rules for adjective noun combinations are complex.

a) For example, the phrase red balloon has the semantic properties of balloon combined with the semantic properties of red in an additive manner. In other words, the semantic rule to interpret the combination red balloon adds the property redness to the properties of balloon. Also, a good friend is a kind of friend, just as a red balloon is a kind of balloon. In such examples, there is a merging of semantic properties.

But some combinations are not always additive.

b) The phrase counterfeit dollar or a false friend does not simply have the semantic properties of dollar plus something else. Also, a false friend is not any kind of friend at all.

The semantic properties of dollar or friendliness are cancelled out by the adjectives counterfeit, false. Thus, semantic rules for counterfeit, false are quite different from those for red, good.

Exemplars of class of adjectives Truth of an adjective x is an x

red, good (A red balloon is a red balloon)

true

counterfeit, false, phoney false

5.3.2.2. Sense and reference

Words, phrases and sentences have sense, which is part of their meaning. By knowing the sense of an expression, you can determine its reference, namely what it points to in the world. (The object pointed to is called its referent and the noun phrase is said to have reference).

Some meaningful expressions, for example The present king of France have sense but no reference (France has no kings).

On the other hand, other noun phrases such as proper nouns have reference but no sense.

Verb-centered meaning

In all languages the verb plays a central role in the meaning and structure of most sentences.

In English the verb determines the number of objects and limits the semantic properties of both its subject and its object.

For example, find requires an animate subject and is subcategorized for one object.

In formal, written English, a verb is necessary to have a complete sentence.

Languages of the world may be classified according to whether the verb occurs initially, medially or finally in their basic sentences. All this is evidence for the centrality of the verb.

5.3.3. Sentential meaning

5.3.3.1. Like noun phrases, sentences have sense, which is usually what we are referring to when we talk about the meaning of a sentence. Some linguists would also say that certain sentences have reference, namely ones that can be true or false. Their reference, or extension, is true if the sentence is true, and false if the sentence is false.

5.3.3.2. The truth of sentences

The sense of a declarative sentence permits you to know under what circumstances that sentence is true. Those circumstances are called the truth conditions of the sentence, that aspect of meaning that allows you to determine whether the sentence is true or false.

The reference of a declarative sentence, when it has one, is its truth value, either true or false.

The truth conditions of a declarative sentence are the same as the sense of the sentence.

In the world as we know it, the sentence

The declaration of Independence was signed in 1776

is true, and the sentence

*The declaration of Independence was signed in 1976

is false.

We know the meaning of both sentences equally well and knowing their meaning means knowing their sense or truth conditions. We compare their truth conditions with the real world or historical facts and thus say which one is true and which one false. The truth or falsehood of these sentences is their reference.

5.3.2.3. Paraphrases

We can now give a formal definition of paraphrase. Two sentences are paraphrases if they have the same truth conditions. The following sets of sentences are paraphrases. Despite subtle differences in emphasis they share the same truth conditions:

The horse threw the rider.

The rider was thrown by the horse.

It is easy to play sonatas on this piano.

Sonatas are easy to play on this piano.

This piano is easy to play sonatas on.

Booth assassinated Lincoln.

It was Booth who assassinated Lincoln.

It was Lincoln who was assassinated by Booth.

The person who assassinated Lincoln was Booth.

The teacher gave the students some books.

The teacher gave some books to the students.

5.3.2.4. Entailment

Entailment is a logical relationship between two sentences such that the truth of the second sentence necessarily follows from the truth of the first.

Sometimes knowing the truth of one sentence entails or necessarily implies the truth of another (or the falseness of another). For example, if you know it is true that

Corday killed Marat.

Then you know that it is true that

Marat died.

It is logically impossible for the former to be true and the latter false. Thus, the sentence Corday killed Marat - entails the sentence Marat died.

The entailment here is a consequence of the semantic relationship between kill and die. We know that there are various ways to die and that being killed is one of them.

In other words, an important part of the sense of killed is contained within die.

Also, the sentence The brick is red - entails the sentence The brick is not white.

Or: Smith is a bachelor entails: Smith is a male, and so on.

Much of what we know about the world comes from knowing the entailments of true sentences.

5.3.3.5. Contradiction is negative entailment, that is, where the truth of one sentence necessarily implies the falseness of another sentence. For example:

Elizabeth II is Queen of England. / Elizabeth II is a man.

Jane is a baby. / Jane is an adult.

If the first sentence of each pair is true, the second is necessarily false. The relationship is called contradiction because the truth of one sentence contradicts the truth of the other.

5.3.4. When SEMANTICS and SYNTAX meet

Syntax is concerned with how words are combined to form phrases and sentences; semantics is concerned with what these combinations mean.

The Theta criterion, discussed in a previous section is an instance in which semantics and syntax interact. The semantic constraint that no thematic role may occur more than once has the effect of restricting the NPs and PPs that may follow the verb in a verb phrase.

We have seen that the same meaning may be expressed syntactically in more than one way the phenomenon of paraphrases. Thus, the semantic property of possession may b expressed by a word in the genitive case such as Englands queen or by an of-construct such as the queen of England.

A similar situation arises with certain semantic concepts such as ability permission or obligation. These may be expressed:

- through auxiliary verbs:

He can go / He may go / He must go

they may also be expressed phrasally, without the auxiliaries:

He is able to go / He has the ability to go

He is permitted to go / He has permission to go

He is obliged to go/ He has an obligation to go

It is often possible to substitute a phrase for a word without affecting the sense of the sentence:

It is often possible to substitute a phrase for a word without affecting the sense of the sentence:

John saw Mary / John perceived Mary using his eyes.

The professor lectured the class. /The professor delivered a lecture to the

class.

Another common type of paraphrase is represented by active passive pairs:

The child found the puppy. / The puppy was found by the child.

This relationship between actives and passives is based on syntactic structure as we saw in the chapter on Syntax.

However, some active sentences do not have a well-formed passive counterpart. For example:

Jim resembled Bill.

The book cost ten dollars.

cannot undergo the passive transformation to give:

*Bill was resembled by Jim.

*Ten dollars was cost by the book.

Semantically, when the subject of an active sentence is in a state described by the verb and direct object, there is no passive paraphrase. Since Jim is in a state of resembling Bill Jim doesnt do anything the sentence fails to passivize. This shows how the semantics of verbal relationships may affect syntactic relationships.

Another example of how syntax and semantics interact has to do with reflexive pronouns. The meaning of a reflexive pronoun always refers back to some antecedent.

In Mary hit herself, herself refers to Mary.

Syntactically, reflexive pronouns and their antecedents must occur within the same sentence (in the phrase structure tree).

If a noun phrase and a non-reflexive pronoun occur within the same sentence, semantic rules cannot interpret them to be co-referential, that is having the same referent. Thus, in

Mary hit her her refers to someone other than Mary.

5.3.5. WHEN RULES ARE BROKEN

There are three kinds of rule violation in language that we will discuss:

i. Anomaly a violation of semantic rules to create nonsense.

ii. Metaphor - or non-literal meaning

iii. Idioms, in which the meaning of an expression may be unrelated to the meaning of its parts.

5.3.5.1. Anomaly: No Sense and Nonsense

Anomaly occurs in many ways in language. It may involve contradictory semantic properties, nonsense words, violation of semantic rules, etc.

Sentences are anomalous when they deviate from certain semantic rules. Thus, the sentence

My brother is an only child.

conforms to all the grammatical (syntactic) rules of the language. Nevertheless, the sentence is strange, or anomalous: it is strange because it represents a contradiction: the meaning of brother includes the fact that the individual referred to is a male human who has at least one sibling, and cannot sensibly be an only child.

The semantic properties of words determine what other words they can be combined with. One sentence that is used by linguists to illustrate this fact is

Colourless green ideas sleep furiously. (N. Chomsky: 1965)

The sentence seems to obey all the syntactic rules of English: the subject is colourless green ideas and the predicate is sleep furiously. It has the same syntactic structure as the sentence

Dark green leaves rustle furiously.

but there is obviously something wrong semantically with the sentence. The meaning of colourless includes the semantic property without colour but it is combined with the adjective green, which has the property green in colour. Other such semantic violations also occur in the sentence.

Other English sentences make no sense at all because they include words that have no meaning - nonsense words. They are uninterpretable.

L.Carrolls Jabberwocky is probably the most famous poem in which most of the content words have no meaning they do not exist in the lexicon of the grammar. Still, all the sentences sound as if they should be or could be English sentences:

Twas billig, and the slithy toves

Did gyre and gimble in the wabe

Semantic violations in poetry may form strange but interesting aesthetic images, as in Dylan Thomass phrase a grief ago. Ago is ordinarily used with words specified by some temporal semantic feature:

a week ago, a month ago, a century ago.

But not: *a table ago, *a dream ago.

When Dylan Thomas used the word grief with ago he was adding a durational feature to grief for poetic effect.

Also, in the poetry of e.e. cummings there are phrases like

The six subjunctive crumbs twitch

Children building this rainman out of snow.

Though these phrases violate some semantic rules, we can understand them; it is the breaking of the rules that creates the imagery desired.

5.3.5.2. Metaphors

Sometimes the breaking of semantic rules can be used to convey a particular idea.

Many sentences have both a literal and a non-literal or metaphorical interpretation (which can give rise to ambiguity). For instance, Walls have ears: in some sense the sentence is ambiguous, but the literal meaning is anomalous. It is so unlikely that we stretch our imagination for another interpretation: the sentence can be interpreted as meaning you can be overheard when you think nobody is listening.

That stretching of our imagination is based on semantic properties that are inferred or that provide some kind of resemblance. Such non-literal interpretations of sentences are called metaphor.

The literal meaning of a sentence such as

My new car is a lemon.

is again anomalous (e.g. the new car which may be a miniature toy carved out of a piece of citrus fruit - a lemon). The more common meaning, however, would be metaphorical and interpreted as referring to a newly purchased automobile that breaks down and requires constant repairs. The imagination stretching in this case may relate to the semantic property tastes sour that lemon possesses.

Metaphors are not necessarily anomalous when taken literally. Thus, the sentence:

John is a snake in the grass.

can be interpreted literally to refer to a pet snake on the lawn named John. Metaphorically, the sentence has nothing to do with a scaly, limbless reptile. It means someone you cannot trust, a hidden, treacherous enemy.

To interpret metaphors we need to understand both the literal meaning and facts about the world. Thus, to understand the metaphor: Time is money

it is necessary to know that in our society we are often paid according to the number o hours or days worked. This metaphor is interpreted as time should not be wasted because you lose money as a result.

A metaphor can be defined as

a semantic mapping from one conceptual domain to another, often using

anomalous or deviant language (D. Crystal: 249)

Several kinds of metaphors can be identified:

a) a conventional metaphor is one which forms a part of our everyday understanding of experience and is processed without effort, such as: a ray of hope, a storm of indignation, a shadow of a smile, floods of tears, to lose the thread (of an argument) etc.

b) a poetic metaphor combines everyday metaphors, especially for literary purposes and this is how the term is traditionally understood, in the context of poetry.

c) conceptual metaphors are those functions in speakers minds which implicitly condition their thought processes, for example, the notion that Argument is war underlies such expressed metaphors as: I attacked his views.

d) The term mixed metaphor is used for a combination of unrelated or incompatible metaphors in a single sentence, such as:

This is a virgin field, pregnant with possibilities

According to G. Finch, metaphor is

a process in which one semantic field of reference is carried over, or

transferred, to another (G. Finch, 2000: 169).

So, for example, in the sentence The ship ploughed the water, the field of farming is transferred to that of sailing. In traditional literary criticism, the field being described, in this case sailing, is referred to as the target, and the field being used for the comparison, farming, is referred to as the source. Other, more traditional terms are tenor and vehicle.

There are two main positions on the role of metaphor in language.

(1)The first, often called the classical view, can be traced back to Aristotle. Basically, this sees metaphor as a kind of decorative addition to ordinary language. Metaphor is regarded as something outside normal language, requiring special forms of interpretation. A version of this is often adopted in the literal language theory. According to this, metaphor is a form of anomaly, or deviation, which a hearer recognizes as such before employing strategies to construct the figurative, or non-literal meaning. These usually involve lifting the selection restrictions on words.

The difficulty for linguists following this approach, however is to describe the set of rules which governs when and how selection restrictions may be lifted. One of the most influential of such attempts is that provided by S. Levin (1977). Levin bases his account of metaphor on the work carried out by Katz and Fodor into semantic primes. Taking as his example The stone died, he argues that the anomaly of this sentence lies in the incompatibility of the features for stone and die. In the case of stone these are:

physical object

natural

non-living

mineral

concreted

and for die:

process with result, namely, that some living entity x ceases to be living.

[adapted from Levin]

The semantic features which are incompatible here are non-living and living. What happens in the case of metaphor, Levin argues, is that one feature transfers across and neutralises the other so allowing us to provide an acceptable interpretation. If non-living transfers to die it produces the interpretation The stone ceased to be, whereas if living transfers to stone, it produces the interpretation The natural physical object died. Levin develops a number of what he calls construal rules, which he argues govern when and how such feature transfers can occur. These typically employ processes of disjunction, conjunction, or displacement.

(2) The second approach, often called romantic, views metaphor not as an anomaly requiring special methods of interpretation, but as an integral part of language and thought. Fro this standpoint there is no real distinction between figurative and non-figurative language since all language is essentially metaphorical. An extension of this view can be found in the work of cognitive semanticists, though they usually adopt a weaker version of the romantic position in which some distinction is made between literal and figurative language. Of principal importance here is the work of George Lakoff and Mark Johnson (1980), who see metaphor as a naturally occurring feature of language and a consequence of the way in which we reason and think about the world.

Lakoff and Johnson distinguish three basic kinds of metaphors:

(i) The first consists of structural metaphors of the sort X is Y, where one thing is experienced and understood in terms of another (1980: 5). This is what most people conventionally understand as metaphor. They illustrate this with various metaphors surrounding the concept of argument, e.g.

An argument is war, An argument is a building, An argument is a container,

and so on. These metaphorical concepts, as they call them, develop through the mapping of one kind of experience onto another. They underlie such expressions as

He defended his argument, His argument is founded on, and His

argument includes the idea that

(ii) The second kind of metaphor is called orientational. These are concerned with the way in which we spatialise experience. They arise from our awareness of our own bodies and the way they function in a predominantly physical environment. So, for example, the fact that up is connected with things which are pleasant and good

His spirits rose, Im in peak condition, Things are looking up,

and down with the reverse

Hes feeling down, Sales are down, That was a low thing to do,

reflects our sense of the importance we attach as physical beings to standing up as opposed to lying down: the former being associated with activity and alertness, and the latter with sleep and death.

Similar metaphorical correspondences are constructed on the basis of our physical awareness of in/out, front/back, on/off and near/far. These correspondences allow us to project our orientation onto situations and things around us. We experience ourselves as having an inside and an outside, a front and a back, and correspondingly we talk of being on the inside of a decision, or of some activity as a front for something else.

(iii) The third kind of metaphor is ontological. These arise from our experience of objects as discrete entities or substances (Lakoff and Johnson,1980: 5). Ontological metaphors are typically used to comprehend events, actions, activities and states (: 30)

5.3.5.3. Idioms

Knowing a language includes knowing the morphemes, simple words, compound words and their meanings. In addition, it means knowing fixed phrases, consisting of more than one word, with meanings that cannot be inferred from the meanings of the individual words. The usual semantic rules for combining meanings do not apply. Such expressions are called idioms.

Idioms are phrases whose meaning is not the combination of the meanings of the individual words. All languages contain many idiomatic phrases, as in these English examples:

Cut it / that out used for telling somebody to stop doing something that you

dont like

Haul somebody over the coals speak to somebody severely because they

have done something wrong

Ill eat my hat / boots (if Im wrong)! a declaration of absolute confidence in

the correctness of ones judgement

Let ones hair down to relax completely and enjoy oneself

Put ones foot in ones mouth to accidentally say something that is

embarrassing or that upsets or annoys somebody

Sell somebody down the river betray somebody for some personal profit or

advantage

Idioms are similar in structure to ordinary phrases except that they tend to be frozen in form and do not readily enter other combination or allow the word order to change. Thus,

(i) She put her foot in her mouth

has the same structure as

(ii) She put her bracelet in her drawer.

Sentence (ii) allows passive or relative transformation:

Her bracelet was put in her drawer.

The drawer in which she put her bracelet was hers.

On the other hand, the following sentences do not have the idiomatic sense of sentence (i):

Her foot was put in her mouth.

The mouth in which she put her foot was hers.

Nevertheless, the words of some idioms can be moved without affecting the idiomatic sense:

The FBI kept tabs on radicals.

Tabs were kept on radicals by the FBI.

Radicals were kept tabs on by the FBI.

Idioms often violate co-occurrence restrictions of semantic properties, i.e. they can break the rules on combining semantic properties. Consider the idioms:

Ill eat my hat, eat ones heart out

The object of eat must usually be something with the semantic property edible, but in these idioms this restriction is violated.

Idioms, grammatically as well as semantically, have special characteristics. They must be entered into the lexicon or mental dictionary as single items with their meanings specified, and speakers must learn the special restrictions on their use in sentences.

Many idioms may have originated as metaphorical expressions that established themselves in the language and became frozen in their form and meaning.

5.3.5.4. Collocations and Idioms

When a word becomes closely associated with a particular context to the exclusion of other words with a similar meaning such that they form what is almost a set phrase, we have what linguists call collocation.

From a logical point of view we (the English) could perhaps refer to a *complete moon but we dont, we refer to a full moon.

White coffee is not white and black coffee is not black; it would seem that white and black are being used to indicate polarity rather than to give an accurate indication of colour.

We can refer to a flock of sheep or birds but not a *flock of cows.

In some cases the association is so close that we may well be able to anticipate it; most of us would, for example, expect She has blondto be followed by hair.

One might ask whether collocation should be dealt with within the context of the semantic range of a word. As blond can describe little more than hair, should we include reference to hair in the definition of the semantic range of the word (+colour, -dark, +hair, etc) or should the denotation of a word and the environment in which it occurs be kept separate? Geoffrey Leech (1974: 20) refers to collocation as simply an idiosyncratic property of individual words.

White coffee and white wine are only relatively white. In the case of white lies and blue jokes the logical link with a colour is less apparent still. With white coffee we are dealing with a set phrase where the meaning can at least be guessed at on the basis of the two constituent words. Conceivably the white in white lie suggests purity of intention. In the case of blue joke knowing what part of the spectrum is indicated by the word blue gives us not help in understanding what a blue joke is.

When the phrase can only be understood as an entity it is an idiom.

This independence of the meaning of their constituent words gives idioms great freedom. As there is no reason for smutty jokes to be blue, Spaniards are equally justified in calling them green (chistes verdes); in China it is the colour yellow (hu ngsè) that is associated with such things.

Foreigners often feature in idioms, usually unfavourably. Thus, while the English take French leave, the French filent a langlais.(cf. Romanian a o sterge englezeste)

Summary

Knowing a language means knowing how to produce and understand sentences with particular meanings. The study of linguistic meaning is called semantics. Lexical semantics is concerned with the meaning of morphemes and words; phrasal semantics with phrases and sentences. The study of how context affects meaning is called pragmatics

Words may be divided into content words, those which identify something in the world around us, and function words, those which serve to specify, link, ad so on the content words. Semantics is principally concerned with content words. The set of objects, actions and so on that such a word denotes is known as the semantic range of that word.

The meanings of morphemes and words are defined in part by their semantic properties, whose presence or absence is indicated by use of semantic features.

The semantic range of a word can be defined by such techniques as hyponymy and componential analysis, reference to synonyms and antonyms. Hyponyms are words that shre a feature indicating they all belong to the same class. When two words have the same meaning but different sounds, they are synonyms. Two words that are opposite in meaning are antonyms. There are antonymous pairs that are complementary, gradable and relational opposites.

Context contributes to the defining of the range of a word. In some contexts the meaning of a word cannot be fully determined without reference to a wider phrase. In such a phrase where there is a conventional association between the constituent words we are dealing with collocation. On the other hand, if the constituent words give little, if any, indication of the meaning of the phrase then we are dealing with idiom.

Sentences are anomalous when they deviate from certain semantic rules. Some sentences are uninterpretable because they contain nonsense words. Many sentences have both a literal an a nonliteral or metaphorical interpretation. Idioms are phrases whose meaning is not the combination of the meanings of the individual words. Idioms often violateco-occurrence restrictions of semantic properties.

It is not always clear whether two words represent two different meanings of the one lexeme or are two different lexemes. In the case of the former we are dealing with polysemy, in the case of the latter with homonymy.

Exercises

Complete the following diagram by (a) devising a category that distinguishes the word bus from the word car, and (b) giving the appropriate symbol against each component for the word motorcycle.

Powered Carries people Four-wheeled

Bus + + +

Car + + +

Van + - +

Bicycle - + -

2. Arrange the vehicles in the above exercise, together with some appropriate

superordinates and hyponyms, in a hyponymy diagram.

3. How valid do you consider the concept of synonym to be?

4. For each of the following pairs of words, state the principal reason why they may

not be considered to be synonyms:

man boy

pavement sidewalk

toilet loo

walk run

determined stubborn

5. Discuss the problem of distinguishing between homonyms and polysemic

lexemes. Give an example of the practical relevance of this distinction.

6. Indicate whether the following antonyms are complementary, gradable or

relational opposites:

good bad; expensive cheap; parent offspring; beautiful ugly; false

true; pass fail; lessor lessee; hot cold; legal illegal; larger smaller; poor

rich; fast slow; asleep awake; husband wife; rude polite

7. The following sentences consist of a verb, its noun phrase subject and various

objects. Identify the thematic role (agent, theme, location, instrument, source,

goal, experience, causative, or possessor) of each noun phrase.

Example: The boy took the books from the cupboard with a handcart.

The boy = agent; the books = theme; from the cupboard = source;

with a handcart = instrument

a. Mary found a key in the house.

b. The children ran from the playground to the wading pool.

c. The hay was loaded on the truck by the farmer.

d. The farmer loaded hay with a pitchfork.

e. The farmer loaded hay onto the truck.

8. Explain the semantic ambiguity of the following sentences by providing two or

more sentences that paraphrase the multiple meanings.

Example: She cant bear children. - can mean either

i.            She cant give birth to children.

ii.          She cant tolerate children.

a. He waited by the bank.

b. Is he really that kind?

c. He saw that gasoline can explode.

d. The long drill was boring.

Guide to exercises:

1. (a) One might distinguish between buses and cars by introducing a category

such as public (or, conversely, private).

(b) A motorcycle would be defined as +powered, +carries people, -four-wheeled,

-public.

2. Possible hyponymy diagrams for these words include the following:

vehicles vehicles

people carrying freight carrying powered non-powered

bus car bicycle motorcycle van bus car van motorcycle bicycle

3. Your answer should show awareness of the variety of factors that may prevent

words from being completely interchangeable. We have seen that syntax

prevents hide and conceal being considered synonyms. In one context we might

use either of two words but in another context we might only be able to use one,

as we saw with high and tall. Two words might be distinguished by social

register, one being more colloquial than the other, by whether they are

associated with a positive or negative point of view, and so on.

4. The words man and boy are principally distinguished by age, the words walk

and run by speed. The principal distinction between the words toilet and loo is

one of social register. Determined and stubborn are largely distinguished by

attitude a person reluctant to give up is described as determined by those who

sympathise and as stubborn by those who do not. The difference between the

words pavement and sidewalk is a matter of geography, the former being used

in Britain and the latter in America.

5. Two words with the same form may be different lexemes or may refer to

different elements of the semantic range of the one lexeme. But it is often

difficult to decide which is the case. One guide is the degree of similarity

between the sense of the two words (consideration of foreign elements might

help). Another guide is etymology: if the two words come from distinct sources

they are more likely to belong to different lexemes. The decision of whether or

not two words belong to the same lexeme is of practical significance to

lexicographers as it determines whether the words should be presented under

the one headword or as separate headwords.