This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
, and . Distinguishing graphemes and phonemes in this way is important because the systems from which they derive, although 25
FROM SHAPES TO WORDS
theoretically parallel in many respects, employ different physical resources—resources which are governed by their own specific sets of rules and conventions. In short, graphemes simply do not exhibit an exact one-to-one correspondence with phonemes. For instance, the grapheme in English may yield two different pronunciations depending on the word in which it occurs: in ‘clue’ it is /k/ but in ‘juice’ it corresponds to /s/. Conversely, the phoneme /z/ can be represented not only by the grapheme in a word like ‘size’, but also by the grapheme <s> in ‘rise’. A related theoretical point concerns the manner in which phonemes and graphemes are produced. The grapheme, for instance, can be written in a host of different typefaces and fonts such as , or . These individual realisations of a single grapheme are normally called allographs. Although they all exhibit variation, they are still representative of a general category; transitions between one and the other will not therefore alter the meaning of the word in which they occur. The same could not be said of variation between, say, <s>, and . The ‘allo-’ principle also applies to the spoken system of language. A phoneme like /t/ may be pronounced in a host of different ways depending on a range of contextual factors, of which the most important are the social and geographical origins of the speaker. To return to an earlier example, the acoustic quality of the /t/ in ‘bit’ will vary depending on whether it is spoken with, say, the Cockney accent of London, the ‘Scouse’ accent of Merseyside or the accent of English spoken in South Africa. In the first variety, the /t/ is not produced as a familiar ‘t’ sound as such, but rather as a glottal stop. In the second variety, the /t/ would be accompanied by a considerable amount of breathiness—indeed, the resulting word would sound almost like ‘biss’. The South African /t/ would be produced with no aspiration or breathiness at all, contributing to what is often perceived as the ‘dry’ quality which distinguishes this accent of English. These three variant pronunciations of /t/ are referred to as allophones of /t/. They are, furthermore, only three of possibly dozens of allophonic variants of this phoneme. Despite this variation in the actual ways in which /t/ is produced, however, none of the variants is sufficient to alter the meaning of the word in which they occur. Allophones, just like their allograph counterparts, are not distinct units of meaning.1 Aside from the theoretical parallels between the systems of speech and writing, there are substantial points of departure between the 26
FROM SHAPES TO WORDS
two modes. This divergence normally reflects the different functions which the two modes serve. Writing, for instance, permits contextual displacement where writer and reader can be separated in time and space. Spoken language, on the other hand, is normally channelled through a physical context which is shared between speaker and hearer. Where writing is characterised by permanence, speech is ephemeral. Features of one system, moreover, may have no direct counterpart in the other. For example, the written mode contains a set of logograms—graphemes like and —which ‘stand for’ other words. Strictly speaking, logograms have no spoken form: the ‘&’ symbol cannot be read aloud without prior conversion to the word ‘and’. The stylistic exploitation of the system of graphology is often considered the preserve of poets rather than novelists. Indeed, the stylistic effects created within the genre of writing known as ‘concrete poetry’ rely almost exclusively on manipulation of the visual medium of language. Nevertheless, writers of prose fiction have, in principle, those same techniques at their disposal, and as early as the eighteenth century the novelist Laurence Sterne, author of Tristram Shandy, was exploring the potential of the written system to the full. For a more recent example of graphological innovation in the context of prose fiction, consider the opening of the second chapter of Samuel Beckett’s novel Murphy (1938).2 This sequence is devoted to the introduction of the character Celia, a prostitute, who is to feature as Murphy’s confidante throughout the story. Until this point in the novel, the text has been written entirely within conventional, left to right, connected prose: Age Head Eyes Complexion Hair Features Neck Upper Arm Forearm Wrist Bust Waist
Unimportant Small and Round Green White Yellow Mobile 13 3/4" 11" 9 1/2" 6" 34" 27" 27
FROM SHAPES TO WORDS
Here, details of the new character are presented not in fully grammatical sentences but in a format which is reminiscent of official documents like passports. In keeping with the discourse type which it echoes, fixed categories are listed in the left-hand column while specific, individual characteristics are provided in the right. It is worth considering how a more conventional introduction to Celia might look: Celia’s green eyes were set in her small and round head. Her complexion was white, her hair yellow… The columnar format adopted in the Beckett also restricts the elaboration of physical characteristics: none of the adjectives used is modified with intensifiers like ‘rather’ (‘rather white’) or ‘quite’ (‘quite small and round’). Of course, the usefulness of the details which are provided is highly questionable. Where age, for instance, is arguably one of the most important pieces of information required on such documentation, forearm size is most certainly not. The criterion for describing someone’s features is normally ‘distinguishability’, not ‘mobility’. It could be added that this type of graphological deviation subverts textual norms on two planes. First, it constitutes a break with the canonical prose format which represents the norm for the novel as a whole. Second, it subverts an extraneous register, ‘officialese’, by appropriating it and then recontextualising it into the context of prose fiction. The linguistic system of graphology interacts in subtle and sophisticated ways with the cognitive systems of information processing and working memory. Graphology, in other words, exerts a psycholinguistic influence on the reading process. Much of the activity of reading relies on informed guesswork; efficient readers do not read words, they read meanings (Smith 1973:188). Children, in fact, often develop this skill instinctively. They quickly learn not to linger over every individual word, but learn instead to search for grammaticality, to form hypotheses—in short, to read for sense. This psycholinguistic ‘guessing game’ relies on the scanning of visual information to produce coherent readings, while simultaneously suppressing ambiguous readings. However, in those genres of writing that exploit the resources of language—of which literature is a preeminent example—this visual medium provides an excellent opportunity for controlled and motivated ambiguity. Before we 28
FROM SHAPES TO WORDS
examine some examples from literary texts, it is worth undertaking a couple of brief experiments in order to illustrate more clearly the psycholinguistic potential of the visual medium of language. The first experiment will require a little piece of paper or card and a commitment from you not to turn over the page, or read the next paragraph on this page until instructed to do so! What you will find overleaf is a single sentence which is built up through the progressive accumulation of units. The first word, ‘Businessmen’, occurs on a line of its own; the next line adds on the second word, the next line the third and so on. This pattern is continued until the sentence is complete. As you read through it, note your impressions of how the sentence develops as a unit of meaning and how your process of reading employs hypothesis formation and re-formation. Now read the sentence one word at a time, using your piece of paper or card to blot out the lines below. The interpretative strategies which this example warrants should illustrate how reading is often a process of conceptual reorientation and revision. Our search for ‘sense’ depends on the interpretation of chunks of language as coherent units of meaning, and frequently this means taking psycholinguistic ‘short cuts’. These short cuts, which are normally referred to as perceptual strategies, are an integral part of the reading process. They involve forming hypotheses about linguistic units in order to reduce ambiguity, as well as discarding other hypotheses in the light of new information. For instance, the strategies used to process the sentence overleaf will require progressive revision as each new element is added on. The third line is the first which ‘looks’ like a coherent unit of meaning and would yield a satisfactory (if sexist!) interpretation. This hypothesis has to be revised, however, when in the fourth line it becomes clear that ‘like’ is to be interpreted as a comparative item and not as a finite verb. Readers may have experienced further reorientations at the fifth line (‘Businessmen like secretaries are difficult’) where a reading is suggested along the lines that both types of people tend to be recalcitrant or tetchy. However, the subsequent text reveals that this reading too has to be jettisoned. Thus, processing the sentence rests on a string of interpretations which are progressively revised until the guessing game is over. One of the consequences of the interaction between visual information and cognitive processing is that it often interferes with a reader’s intuitive judgments about the grammaticality of a particular 29
FROM SHAPES TO WORDS
Businessmen Businessmen Businessmen Businessmen Businessmen Businessmen Businessmen
like like like like like like
secretaries secretaries secretaries secretaries secretaries
are are difficult are difficult to are difficult to fool
sentence. Perceptual strategies can, in other words, often lead to mis-analysis and misinterpretation. Here is a good example of a sentence which would pose such a problem: 1 The train left at midnight crashed. It is likely that this sentence would be judged ungrammatical by many readers and would simply be interpreted as a clumsily expressed attempt at The train left at midnight and crashed’ or The train which left at midnight crashed.’ Yet a more thorough inspection will reveal that there is, in fact, nothing ill-formed about this example at all. This is best explained by invoking a sentence with a superficially parallel structure. Consider the following, which should cause no problems: 2 The baby abandoned at midnight cried.
¶ 30
This parallel structure will help ‘disambiguate’ our problematic example. What has happened in (1) is that the intuitively coherent grammatical unit (‘The train left at midnight’) has interfered with our perception of the sentence as a whole. In the second example, however, the superficially similar sequence (The baby abandoned at midnight’) creates no such obstacle—it simply cannot be interpreted as a complete grammatical unit. This second sequence thus reinforms our interpretation of the parallel construction in the first so that it can now be successfully reinterpreted along the lines of The train which was left (i.e. in a siding, in a tunnel) at midnight crashed.’ The point is that it is the perceptual strategy—the grammatical short cut which is so central to the act of reading—that triggers the misanalysis in the first place.
FROM SHAPES TO WORDS
Informal experiments like these can highlight the ways in which visual information is encoded and retrieved. They also illustrate the linguistic complexity of the written system. Indeed, one of the benefits of graphological innovation in general is that stylistic effects can be created on more than one level. Where a line ending in poetry may, for instance, suggest a pause, it may also function as a subtle conceptual break with other grammatical structures. An excellent illustration of this aspect of the visual medium of language is offered by Roger Fowler (1986:45) in his discussion of William Carlos Williams’s poem ‘The Right of Way’. Fowler is particularly interested in the line boundaries in the last of the following three stanzas: Why bother where I went for I went spinning on the four wheels of my car along the wet road until I saw a girl with one leg over the rail of a balcony The structure of the final stanza creates a striking linguistic trompe l’oeil. The reader first assumes that the penultimate line refers to an amputee, only to have the missing leg ‘restored’ in the final line. These two potential readings, Fowler argues, are contradictory: the penultimate line suggests pathos, violence, the grotesque, whereas the girl on the balcony implies relaxation and confidence. The result is a visual ambiguity where, although the second of the two projected readings is preferred, the impression left by the first never completely disappears. Manipulation of the visual system of language is not the exclusive preserve of literary communication. Indeed, graphological exploitation is a resource which is available to other genres of discourse. There is not the space here to develop this point in detail, but a brief illustration from advertising language should none the less prove useful. The extract below is part of an advertisement which ran regularly in the weekly magazine TV Times. The product advertised is a beverage called ‘Cranberry Classic’. The pictorial detail which forms the backcloth to this written text comprises a tranquil village cricket game, set in the late afternoon and photographed in warm and sumptuous colours. 31
FROM SHAPES TO WORDS
The message superimposed on to this salubrious scene takes the following format: ALL THE JUICE GOODNESS OF FRUIT. YET THERE’S NOTHING TRADITIONAL ABOUT IT. CRANBERRY CLASSIC IS THE JUICE DRINK THAT GIVES YOU BOTH GOODNESS AND AN EXCITING, DRY, REFRESHING TASTE. The layout of the text is a crucial determinant of the interpretative path it encourages. Although the product is a ‘juice drink’—the main ingredients of which are normally sugar, water and fruit flavouringone might have assumed that the text actually advertises ‘fruit juice’. The graphological presentation of the first sentence is especially telling in this respect. Significantly, the sentence is split into the two sequences ALL THE JUICE and GOODNESS OF FRUIT, serving to create two information units within a single grammatical structure. The placement of these units on separate lines suggests that the product is ‘all juice’, so to speak, while at the same time it emphasises the ‘goodness’ that fruit contains. Reassembling the sentence offers a very different message: ALL THE JUICE GOODNESS OF FRUIT Here the product emerges more as an approximation to juice; as a beverage which exhibits some of the characteristics of juice. Only later in the advertisement is this aspect of the product divulged: CRANBERRY CLASSIC IS THE JUICE DRINK and, noticeably, this sequence is followed quickly by echoes of the ‘goodness’ theme which opened the extract. The manner by which a text presents information can often be distinguished from the content of that information. As far as one can 32
FROM SHAPES TO WORDS
tell, there is nothing factually untrue or inaccurate about the claims made on behalf of the product. What is at issue here is the technique by which copywriters shape messages in strategically motivated ways. Some elements are played down while others are foregrounded. The visual organisation of the text, in particular, serves to manipulate perceptual strategies, creating cognitive maps which influence the way we process and assimilate information. This section has explored the system of graphology and assessed the creative potential which it offers writers. For the moment, we will leave graphology, in order to introduce some new concepts in language. However, we will return to it later where it will be integrated with the framework of morphology which forms the focus of attention in the next section.
2.3 Morphemes and words This section introduces some of the basic principles relating to the study of words and vocabulary. The general term which is normally reserved for the ‘pool’ of words which form the basis of any language is the lexicon. Like most technical terms in linguistics, this one is derived from Classical Greek: from lexis ‘word’ into lexikon ‘inventory of words’. In most surveys of the English lexicon, it is normal practice to adhere to an important basic distinction between two different types of words. These are known as content words and grammatical words. This distinction is based not only on what particular words mean but also according to what their function is within a sentence. Here is a simple—if unimaginative—example of a sentence which contains both content and grammatical words: 3 The cat sat on the mat Content words, as the term indicates, are those which carry substantial meaning and in this sentence they are ‘cat’, ‘sat’ and ‘mat’. The function of content words is to name the many objects, qualities, actions and events which are manifest in everyday life. Not surprisingly, this class of words is enormous, with new content words being added to the language all the time. The addenda of any recent dictionary, with everything from computer-age ‘megabyte’ to populist ‘bonk’, illustrates how the lexicon is being expanded to 33
FROM SHAPES TO WORDS
account for new developments in a host of social and technological domains. Nouns, adjectives, verbs and adverbs are the parts of speech which typically contribute to the class of content words. Grammatical words, on the other hand, are realised by parts of speech such as articles (‘the’, ‘a’), prepositions (‘in’, ‘by’), pronouns (‘I’, ‘you’, ‘she’) and modal verbs (‘would’, ‘do’, ‘may’). One of the most important functions of these items is to bind content words together to form coherent, grammatical units. Grammatical words thus provide the structural foundations upon which the building blocks of the lexicon—the content words—can be assembled. Unlike content words, grammatical words form a closed system: they comprise a small and stable class. It is rare in any language for new grammatical words to be invented or borrowed from another language. Indeed, the last significant change that affected this part of the English language happened when the Anglo-Saxon and Old Norse pronoun systems partially merged after the Viking invasions of the eighth and ninth centuries. Another criterion which distinguishes the two classes of words is relative length. Grammatical words tend to be short: they are normally of one syllable and many are represented in spelling by less than three graphemes (‘I’, ‘he’, ‘do’, ‘on’, ‘or’). Content words are longer and, with the exception of ‘ox’ and American English’s ‘ax’, are spelt with a minimum of three graphemes. This criterion of length can also be extended to the production of the two sets of words in connected speech. Here grammatical words are often unstressed or generally de-emphasised in pronunciation. For instance, the modal verb ‘would’ and the negative particle ‘not’ can both be contracted to shortened forms: ‘He’d hate it’; ‘She isn’t there.’ Furthermore, in some varieties of writing, grammatical words may be deleted from sentences altogether, on the proviso, of course, that their basic sense can still be inferred in the context. Headlines in newspapers are prime candidates for this, as the following example from the Independent on Sunday3 should demonstrate: Collapse of holiday giant feared by travel industry Those grammatical words which are ‘expendable’ have been removed. Yet it is easy to construct a version with all the deleted items reinserted: ‘The collapse of a holiday giant is feared by the travel industry.’ Notice, however, that ‘of’ and ‘by’ need to be 34
FROM SHAPES TO WORDS
retained in the headline as their removal would lead to ambiguity. This type of linguistic reduction is normally displayed by writing varieties which are constrained by space or where economy of language is at a premium. In addition to newspaper headlines, it is not surprising therefore to find such reduction in the linguistic style of telegrams and small advertisements. This division between grammatical and content words discussed thus far only partly explains the way in which the’lexicon of English is structured. Leaving aside grammatical words for the moment, we need to look more closely at content words. Consider the following set of straightforward content words: happy; walk; horse; like Now consider the following set: unhappy; walked; horses; likable Although the second set is clearly not the same as the first, we would still not want to argue that it comprises ‘different’ words. Rather, it could be proposed that the items in the second list are systematically related to equivalent items in the first. What separates the two are the little particles nesting within the words in the second set. These particles (un-, -ed, -s, -able) have an important grammatical function which in many respects resembles the function of the grammatical words discussed above. However, unlike grammatical words, these particles cannot stand alone. They need instead to be bound on to the stem of the content words. What this reveals is that words may be broken down into still smaller units: ‘unhappy’, for example, may be subdivided into ‘un-’ and ‘happy’. The two segments which make up this word are referred to as morphemes. This technical term, with its predictable Greek etymology, derives from morphe meaning ‘form’, with the current linguistic term morphology now standing for the ‘study of forms’. Returning to the two sets of examples, it can now be proposed that all of the items in the second set of examples contain two morphemes, though each, of course, still constitutes only a single word. The particles ‘un-’, ‘-ed’, ‘-s’ and ‘-able’ we can label bound morphemes in so far as they must be bound to some other unit. The remainder of each word is called the root morpheme. These root 35
FROM SHAPES TO WORDS
morphemes are principally derived from content words: they are thus responsible for carrying the bulk of the meaning of the word in which they occur. Although bound morphemes can operate only in conjunction with other elements, root morphemes may stand alone. Indeed, this capacity is illustrated by the first list of words, which is comprised solely of ‘free-standing’ root morphemes. These morphologically simple words, which contain only a single root morpheme, may be compared to morphologically complex words which contain at least one free morpheme and any number of bound morphemes. Thus, a word like ‘desire’ may be defined as a root morpheme constituting a single word. ‘Desirable’, by contrast, is complex, combining a root morpheme with the bound morpheme ‘-able’. More complex again is ‘undesirability’ which comprises one root and three bound morphemes: un+desire+able+ity. Notice also how, in complex words of this sort, the spelling of the root may be altered to conform to the bound morphemes around it. Thus, ‘desire’ becomes ‘desir-’, while ‘beauty’ will be transformed into ‘beauti-’ in the formation of ‘beautiful’ and of the increasingly common complex ‘beautician’. Bound morphemes may be subdivided in terms of the way they conjoin with other morphemes. Operating on the general principle that it is a particle which ‘fixes’ on to another element, the general term reserved for any type of bound morpheme is affix. The position of a bound morpheme in relation to the root leads to finer distinctions. Bound morphemes placed in front of the root are referred to as prefixes’, those placed after the root as suffixes. Take, for instance, the root ‘moral’. Although this may stand alone, it also permits extensive affixation through prefixes and suffixes. The prefixes which are available are ‘a-’ and ‘im-’ deriving ‘amoral’ and ‘immoral’. Suffixes include ‘-ity’, ‘-ly’ and ‘-ist’ deriving respectively ‘morality’, ‘morally’ and ‘moralist’. The last of these actually permits further suffixation with ‘-ic’ as in ‘moralistic’. Other common prefixes in English are: ‘dis-’ as in ‘disrespect’, ‘un-’ as in ‘unreal’, ‘bi-’ as in ‘bifocal’ and ‘pre-’ as in the word ‘prefix’ itself. Other common suffixes are: ‘-ness’ (‘kindness’), ‘-ment’ (‘judgment’), ‘-s’ (‘looks’) and ‘-est’ (‘fastest’). Brief mention needs to be made of a third type of affix. In some languages of the world, though not in English, bound morphemes may be inserted into root morphemes. Such affixes are referred to as infixes. In Bantoc, a language spoken in the Philippines, the word for 36
FROM SHAPES TO WORDS
‘strong’ is ‘fikas’. If Bantoc speakers want to say ‘to be strong’ they simply insert the particle ‘-um-’ into the stem of the word deriving ‘fumikas’. Similar conversions can be performed on ‘kilad’ (‘red’) to yield ‘kumilad’ (‘to be red’) and ‘fusul’ (‘enemy’) to produce ‘fumusul’ (‘to be an enemy’). Latin also permits a degree of infixing. The root of the word for ‘break’ is ‘rup-’, the antecedent of modern English ‘rupture’. The formation of the present tense ‘I break’ requires the infixation of a nasal consonant to form ‘rumpo’. Although not a feature of English morphology, infixing none the less provides some potential for innovation in language. Traugott and Pratt (1980:90) offer the colloquial ‘fan-damm-tastic’ as an example of creative infixing. They also cite e e cummings’s ‘manunkind’, which can be interpreted as the infixing of ‘mankind’ with ‘-un-’. The stylistic effect created by this complex, they suggest, is one by which the concept ‘man’, in a generic sense, is portrayed simultaneously as not only unkind but also without true kin. Classifying affixes into their respective subcategories is one means of dealing with bound morphemes. There is another method, by which they may be classified not in terms of their positions in a word but in terms of the operations they perform on that word. Some morphemes alter the meaning of the word in which they occur, some change it to a new word-class while others are simply needed to main tain the grammaticality of the sentence in which the word occurs. To return to some of the earlier examples, the ‘un-’ morpheme when prefixed on to ‘happy’ clearly derives the new meaning of ‘not happy’, while ‘a-’ in front of ‘moral’ produces ‘without morals’. The primary function of other types of bound morpheme is to alter the word-class of the root on to which they are fixed. The addition of ‘ness’ to ‘kind’, for example, will produce a noun from an adjective. By the same token, ‘-ment’ when added to ‘judge’ will convert a verb into a noun, while ‘-fui’ when joined to ‘beauty’ will derive an adjective from a noun. Morphemes which perform either function— that is to say, alter the meaning or the word-class of the root to which they are attached—are called derivational morphemes. This category does not however account for all of the bound morphemes that have been discussed in this section. The addition of the suffix ‘-s’ to the verb ‘look’ alters neither its word-class nor its meaning. ‘Looks’ is still a verb and it doesn’t really differ in sense from ‘look’. All that the ‘-s’ particle does is to ensure grammatical agreement with another element in the sentence. More specifically, it signals third person 37
FROM SHAPES TO WORDS
singular agreement with an antecedent noun, as in ‘She looks’ or ‘John looks’. The same principle applies to the plural morpheme ‘-s’ (‘cats and dogs’), the bound morpheme ‘-ed’ which indicates past tense (‘John walked’), and the suffix ‘-’s’ which signals possession (‘Mary’s tale’; ‘John’s car’). These morphemes operate as grammatical signposts, providing information on how the unit to which they are attached relates to other units in the same grammatical environment. Such morphemes are called inflectional morphemes, or simply inflections. Unlike derivational morphemes, inflections do not alter the meaning or the part of speech of the word in which they occur. They are requirements of syntax, relating units to each other and indicating structural relations within sentences. Derivational morphemes on the other hand have the capacity to alter the meaning potential of a word, in spite of the fact that they cannot stand on their own. Here is a short summary of the principal differences between inflectional and derivational morphemes: Derivational morphemes 1 They change either the word-class or the meaning of the root to which they are attached. 2 They may be prefixes or suffixes, e.g. dis+respect+ful. 3 They typically occur before any inflections in a sequence, e.g. moral+ist+s. Inflectional morphemes 1 They do not change the meaning or word-class of a word, e.g. arrive, arrives and arrived are all verbs. 2 They typically indicate syntactic relations between different words in a sentence. 3 They typically occur at the end of a word, after any derivational morphemes, e.g. moral+ist+s. 4 They are only ever suffixes. This might now be an appropriate stage at which to pause and review all of the material covered so far in this section. In order to give a clearer picture of how the numerous categories and sub-categories interlock, Figure 2.1 is a simple schema which plots the relationship of morphemes to words, and distinguishes the different types of affix found in English. 38
FROM SHAPES TO WORDS
FIGURE 2.1 The system of morphology In order to illustrate how this model may be applied to actual stretches of language, we may perform a short analysis of the following made-up sentence: 4
Mary’s tale of unhappiness disheartened her friends
The first step would be to isolate the root morphemes in the sentence. These tend to be either free-standing content words or items which have formed morphological complexes with bound morphemes. Here is the sentence again, this time with root morphemes highlighted:
39
FROM SHAPES TO WORDS
4(a) Mary’s tale of unhappiness disheartened her friends This is what it looks like with grammatical words indicated: 4(b) Mary’s tale of unhappiness disheartened her friends and now with derivational morphemes highlighted: 4(c) Mary’s tale of unhappiness disheartened her friends and finally, with inflectional morphemes showing: 4(d) Mary’s tale of unhappiness disheartened her friends In this way the morphological elements which make up the sentence can be sifted out and classified systematically. One of the things that should have emerged from the foregoing survey of the lexicon is the degree of sophistication and productivity of the morphological system of English. There exists in English a variety of devices for the formation and derivation of new words. This offers users of the language numerous resources for innovation and creativity. The system is thus continually enriched and modified in order to meet the needs of speakers and writers, and, while some elements become obsolescent, so other new ones are created. An intriguing consequence of this process is the way in which asymmetries and gaps develop in the lexicon. Occasionally morphological complexes survive into Modern English, while, bizarrely, the root morphemes from which they are derived do not. Take for instance Modern English’s ‘uncouth’ and ‘unkempt’. The positive terms upon which these words are built are ‘couth’ and ‘kempt’ and both are well attested in earlier forms of English. The first, a past tense form of Old English cunnan, originally meant ‘to be able to, to know how to’. Yet only the negative derivation survives. Similarly, ‘kempt’ appears in Chaucer as in ‘He kempte his lokkes’4 (i.e. ‘he combed his locks’) but again only the prefixed version has been retained. Part of the explanation for this is that the marked forms achieve special prominence: ‘uncouth’ and ‘unkempt’ represent departures from a norm, something which seems worthy of special mention, whereas the root forms which constitute that norm represent the commonplace. Even though the unmarked root forms 40
FROM SHAPES TO WORDS
have since disappeared, speakers are sensitive to the asymmetry which results: hence, an undergraduate’s recent jocular remark ‘He’s very couth, you know.’ The same humorous turn is also evident in the Mock Turtle’s coinage of ‘uglification’ in Lewis Carroll’s Alice’s Adventures in Wonderland. When pressed by Alice on the validity of the word, the Gryphon supports the Mock Turtle’s lexical innovation with inexorable linguistic reasoning: ‘Never heard of uglifying!’ it exclaimed. ‘You know what to beautify is, I suppose?’ ‘Yes,’ said Alice doubtfully: ‘it means—to make— anything—prettier.’ ‘Well then,’ the Gryphon went on, ‘if you don’t know what to uglify is you are a simpleton.’5 Morphological asymmetry creates other types of gap in the lexicon. It is a feature of many words that they exist only in their root forms, resisting morphological development. Thus, where ‘clean’ and ‘happy’ may be prefixed by ‘un-’, the roots ‘dirty’ and ‘sad’ cannot. This restriction is what permits the experimentation in the following line from e e cummings which is discussed by Traugott and Pratt (1980:33). The last word in the sequence is relevant here: [love] is most mad and moonly and less it shall unbe When prefixed on to verbs like ‘do’, ‘lock’ and ‘screw’, the morpheme ‘un-’ has a reversative function. It encourages a reading along the lines of ‘reverse the action of the process expressed by the root’. Yet here cummings forms a complex with the verb ‘to be’, a verb which resists prefixing of this sort, creating a new type of sense. Although ‘unbe’ shares some of the quality of parallel expressions like ‘die’ or ‘not be’, it also suggests a process of reversal from a state of being to a state of not being which the other items do not. Gaps in the lexicon are also responsible for the process of innovation known as back-formation. The standard procedure for deriving new words is for the root to precede the complex: ‘write’, for example, will convert naturally to ‘writer’ just as ‘read’ yields ‘reader’. In back-formation, however, the complex form actually precedes the more simple form. The verb ‘beg’ is a prime illustration 41
FROM SHAPES TO WORDS
of this process. Historically, it did not, as might be imagined, produce the root for ‘beggar’; rather, it was actually derived from ‘beggar’. What appears to have happened is that ‘beggar’ has been interpreted as ‘agentive’ in the sense of ‘someone who begs’, whereas the genuine etymological origin of the word is ‘beghard’, a name for a medieval brotherhood. This re-interpretation or misanalysis— whatever one chooses to call it—results in the alignment of ‘beg’ with other regular verb-noun paradigms like ‘read/reader’ and ‘write/ writer’. Back-formation in English is also responsible for the derivation of the verb ‘burgle’ from the noun ‘burglar’ as well as the verb ‘diagnose’ from the noun ‘diagnosis’. In fact, new verbs enter the lexicon all the time through this process: witness recent additions like ‘giftwrap’, ‘peddle’, ‘enthuse’, ‘gatecrash’—all of which have been derived from noun complexes. Back-formation is also widely employed in the formation of nonce-words (words made up ‘for the nonce’, for use at a specific moment) which do not take hold in the lexicon but none the less provide an outlet for humour and wit in everyday interaction. It can thus account for a recent remark by a therapist friend who said she was going out to ‘therap’ someone, or a former colleague who, when informed about a particular student’s vulnerability, replied ‘Yes, you could easily vul him’! Another productive area of the English lexicon is its capacity to form new words by combining two or more root morphemes. The words derived from this highly flexible process are known as compounds. This capacity for compounding can be attributed in part to the Germanic origins of English. Indeed, some words which survive from Old English were originally compounds even though the roots which formed the compound have been lost. Our modern word ‘nightingale’ is historically derived from a compound of niht ‘night’ and galan ‘to sing’ producing the rather charming ‘night-singer’. All sorts of permutations are possible in the formation of compounds, with the major word classes able to enter into combinations with a wide range of other items. Table 2.1 is a matrix which illustrates the compounding possibilities of Modern English. For example, the first element of the word ‘hovercraft’ is a verb and the second a noun. In the compound ‘bittersweet’ both elements are adjectives. The principle of compounding lends itself well to creativity in language. Just look at the way it is employed in the following paragraph from the ‘Proteus’ episode of James Joyce’s Ulysses (1922):6 42
FROM SHAPES TO WORDS
Airs romped around him, nipping and eager airs. They are coming, waves. The whitemaned seahorses, champing, brightwindbridled, the steeds of Mananaan. TABLE 2.1 Matrix of compounding possibilities
Two roots make up ‘whitemaned’, while there are three in ‘brightwindbridled’. Compounding of this sort offers the potential to create rich densities of meaning, with individual content words compressed into multiple conceptual clusters. Notice also how in the environment of references to ‘steeds’, ‘-bridled’ ‘-maned’ and ‘Mananaan’ (a Celtic god of the sea), even an established English compound like ‘seahorse’ can acquire new significance. Before closing the more theoretical part of this chapter, it is worth considering a few more types of word formation. Leaving aside the straightforward borrowing of foreign words, here is a checklist of a few of the other common word-formation processes in English, along with definitions and illustrations: Type
Definition
blend A fusion of two root morphemes, where only parts of the words are joined and the remainder deleted
Examples ‘breathalyser’ (breath +analyser); ‘Chunnel’ (Channel+tunnel); ‘smog’ (smoke+fog)
‘pram’ (perambulator); clipping Where a word of two or more syllables is shortened, without a ‘flu’ (influenza); change in meaning or function ‘exam’ (examination); ‘fridge’ (refrigerator) 43
FROM SHAPES TO WORDS
acronym
¶
‘nylon’ (New York+ Words formed with the initial London); ‘Waaf’ letters of a phrase (Women’s Auxiliary Air Force); ‘AIDS’ (Acquired ImmunoDeficiency Syndrome)
This survey of word-formation concludes the introduction to the lexicon of English. The task now will be to investigate the ways in which the material assembled in this and the previous section can be used in textual analysis.
2.4.1 Using linguistic models for stylistic analysis This is the first part of the practical component of the chapter. In this subsection, an analysis will be provided of a poem by the American modernist poet e e cummings; in the next, the analysis is developed further into a workshop activity. A few short examples of cummings’s poetry have already been used as illustrations in the previous section. This should have offered a foretaste of how productive the language of this writer is for graphological and morphological analysis. In this subsection, a more sustained analysis of a complete text will be undertaken. One topic that will not receive a great deal of attention here is the general question of the literary ‘merit’ of cummings’s work. This is simply not the place to provide a detailed evaluation of cummings’s contribution to American literature, or to offer a case for his inclusion within or exclusion from the canon of ‘great’ literature. These issues are best left to the students of cummings, who, it must be added, are thoroughly divided both on the quality of his poetry and its relationship to the literary canon. Friedman, for example, while acknowledging the ‘aura of inadequacy’ which attaches to cummings’s reputation, none the less maintains that he is ‘one of the great creators of our time’ (Friedman 1972:xi). Comparably ambivalent statements have been offered on the literary genre to which the cummings oeuvre should belong: classifications range from ‘nihilist’ to ‘romantic anarchist’ (Wegner 1965:63). Whatever position one takes on canonical or generic issues, the linguistic experimentation which much of cummings’s work displays is well suited to the type of linguistic frameworks assembled over the last two sections. 44
FROM SHAPES TO WORDS
The text chosen for analysis is one of the many short untitled poems in the volume 95 Poems which cummings published in 1958. In the volume’s consecutive sequence of ninety-five poems, it is numbered 55 and its first line is ‘you no’.7 There is no motivation behind this choice other than that it appears to lend itself to the type of stylistic exercise in which we are engaged. It certainly hasn’t received much attention from the critical establishment: my scan of the major studies of cummings revealed not a single reference to this poem. Although no single text is likely to display the full range of graphological and morphological devices identified in sections 2.2 and 2.3, this one should none the less offer a locus for exploring many such techniques. Set in rules overleaf, then, is poem number ‘55’. A cursory inspection of this text should prompt connections with the material presented in the previous two sections. In linguistic terms, the poem exploits the interaction between the written system of language, principles of word-formation and psycholinguistic perceptual strategies. Deployed throughout the text are devices which manipulate the visual medium of language, devices which are supplemented with deviations from standard word-formation. This combination of visual tricks and striking morphological structure can be viewed as a product of the complex intersection of the graphological and morphological systems. Beginning with graphology, we can now investigate more systematically the ways in which these effects are created. The poem bears all the characteristics of the cummings stylistic blueprint. Typical of his work is the ‘deformation’ of standard orthography, layout and punctuation. Perhaps the best known of these techniques is the use of lower case for items which would normally be printed in upper case, such as the first word of a line or the personal pronoun ‘I’—not to mention the spelling of the poet’s own name. By contrast, when upper case is used, it is used in the most unlikely of environments: in this text it is restricted exclusively to comparative and superlative terms (‘Less’; ‘Most’; ‘More’). These terms thus become foregrounded in a text where lower case is the norm. Other significant graphological devices include the logogram ‘&’. This feature was discussed early in section 2.2, where it was pointed out how this and related symbols are situated at the extreme of the written system of language. These symbols have no direct spoken counterpart; rather they need to be ‘translated’ into words. In this instance, then, the logogram will ‘stand for’ the word ‘and’. 45
FROM SHAPES TO WORDS
you no tice nobod y wants Less(not to men tion least)&i ob serve no
5
body wants Most (not putting it mildly much)
10
may be be cause ever
15
ybody wants more (& more & still More) what the hell are we all morticians?
46
20
FROM SHAPES TO WORDS
The use of the logogram is supplemented by the extensive use of parenthesis, often with both types of symbol tied together so closely that they are adjoining. This creates patterns like that at line 6: tion least)&i Where logograms and parenthesis are used to close up elements within a line, the line layout itself serves to break up linguistic elements. Each single line of text is followed by a two-line space, which is in turn followed by a cluster of three lines of text. This cluster is followed by another two-line space before the whole pattern begins again. The shortness of each individual line makes the text almost columnar in format. In fact, apart from the final line (which will be discussed later), no line contains more than five syllables, and the average for the text as a whole comes out at around two syllables per line. All of these graphological features intersect subtly with patterns at the level of morphology, so it is impossible to explore the stylistic effects of any one level in isolation. Operating in tandem with graphological design, for instance, is a morphological technique which exploits the principles of word-formation. Line endings are used to produce morphological breaks, and these breaks subvert the reading process by forcing a series of perceptual ‘double-takes’. The specific strategies which engender this linguistic ‘backtracking’ can be grouped into four types of morphological device. The first of these is where a word is simply broken into two parts, with neither of these parts resembling an item which could make sense on its own. Put another way, the two resulting segments do not resemble free morphemes. Here is an example of this type of breakage, with the word ‘nobody’ becoming:
nobod y wants This type of breakage is the most linguistically unproductive of the four in that neither ‘nobod’ nor ‘y’ resembles any other meaningful units in the lexicon. The same cannot be said, however, of the remaining three types of morphological experimentation in the poem. All three of these strategies are based on the principle that parts of some words happen purely by chance to resemble other root morphemes. 47
FROM SHAPES TO WORDS
The first strategy involves making a split in a word so that the first element looks like a free morpheme. Through this process, ‘notice’ is carved up into ‘no’ and ‘tice’ (lines 1–2), with only the first element forming a ‘real’ word. In this instance, it produces the negative marker ‘no’, a grammatical word. Similarly, ‘mention’ becomes ‘men’ and ‘tion’ (lines 5–6) and while ‘men’ resembles an actual content word in the lexicon, ‘tion’ has no such counterpart. The crucial point is that none of the words so broken up is actually a morphological complex in the first place. None of them contains more than one morpheme. The particle ‘-tion’ is not a suffix of ‘men’ and ‘men-’ is certainly not a prefix of ‘tion’. Rather, the cummings technique is built on analogy: words are fragmented as if they were genuine morphological complexes. Moreover, the distance which is placed between these fragments contributes further to the illusion that they can be interpreted as independent units. Consider, for example, the manner in which ‘notice’ and ‘mention’ are split: you no
tice nobod y wants Less(not to men
tion least… Unlike the strategy employed for the break-up of ‘nobody’, a twoline space is used here for both of these splits. Placing a substantial gap between these segments helps to trigger the type of linguistic trompes l’oeil and perceptual tricks which were discussed in section 2.2. Because the reading process is predicated upon the retrieval of coherence and grammaticality, there is a tendency to search out chunks which look like complete units of meaning. The appearance of ‘men’ at the end of the fifth line looks precisely like one such unit. However, the later addition of ‘tion’ requires a revision of the original hypothesis, and the insertion of a relatively substantial gap between the two elements simply ensures that the cognitive volte face is all 48
FROM SHAPES TO WORDS
the easier to bring off. Interestingly, the same tactic is employed on the remaining example from this category of morphological breakage. The interpretation of ‘ever’ at line 16 needs to be reassimilated to ‘everybody’ when ‘ybody’ appears after a two-line separation: be be cause ever
ybody The third type of morphological breakage is effectively a reversal of the second. In this category it is the second and not the first part of the split which resembles a free morpheme. The only exponent of this type in the text occurs at lines 7–8: …&i ob serve no Here the second part is the one which has the same shape as another item in the lexicon. But again the break is ‘false’ in that ‘ob’ is not a bound morpheme which can be prefixed on to root morphemes. Like all the other types, this type of break creates only the illusion of a genuine morphological split. The fourth and final type is somewhat more sophisticated. This is where a split is placed so that the resulting components both resemble complete words in the lexicon. A sustained pattern of this type begins at line 13: may be be cause The grammatical words, ‘maybe’ and ‘because’, are each broken up into segments. These segments are not constituent morphemes 49
FROM SHAPES TO WORDS
because both words comprise only a single morpheme. Rather, the split is manipulated so that each part resembles another morpheme. Thus, ‘maybe’ yields the modal auxiliary ‘may’ and the verb ‘be’, while ‘because’ produces ‘be’ and the content word ‘cause’. Even a word like ‘nobody’, which had previously been subjected to the first type of breakage at lines 3–4, can be broken in a way that produces other morphemes. So, at lines 8–9, it now produces ‘no’ and ‘body’. The different types of morphological device employed throughout the poem can be captured quite neatly by means of a matrix like Table 2.2. This demonstrates the field of possible lexical permutations, ranging from the breaking of a word into two meaningless chunks to the production of two elements which are the same as other free morphemes in the lexicon. TABLE 2.2 Morphological devices in poem ‘55’
Of course, as has been pointed out, all this morphological experimentation is built on an analogy with genuine morphological complexes. All of the examples covered so far—with the possible exception ‘nobody’—have been divided into ‘false’ constituents. For instance, whereas a word like ‘becalm’ genuinely comprises the bound morpheme ‘be’ and the root ‘calm’, the word ‘because’ exhibits no such parallel structure. Similarly, while the root ‘observe’ may be prefixed and suffixed to create ‘unobserved’, the root itself cannot be subdivided into more minimal forms. The same principle extends, for that matter, to the root ‘notice’. It is this type of ‘breakage by analogy’ which forms the stylistic cornerstone of the text and creates the perceptual re-orientations which permeate the poem. Before assessing the implications of these techniques, a brief word is necessary on the last line of the poem. In many respects, this line carries a great deal of stylistic weight as it breaks many of the textual patterns established prior to this point. Its relative length, for one thing, 50
FROM SHAPES TO WORDS
sets it apart from the other lines of the poem. It also displays a degree of grammatical wellformedness not shared by the remainder of the text: indeed, its very coherence foregrounds it in a text where linguistic fragmentation has become the norm. Moreover, where the grammatical mood of the entire poem up to this point has been declarative (the form standardly used for making statements), it switches twice in the space of the final seven words. The sequence ‘what the hell’ is exclamative, while the final clause ‘are we all morticians?’ is interrogative. The poem thereby concludes with a grammatical structure used to ask a (perhaps rhetorical) question. The final word of this sequence—itself the longest word of the entire poem—produces what might be best described as a ‘morphological pun’. In American English the word ‘mortician’ performs the same function as British English’s ‘undertaker’. In terms of morphology, the American English version is built from the Romance-language root ‘mort-’ and the derivational morpheme ‘-ician’, to form a complex which means something along the lines of ‘practitioner of death’. In fact, the word ‘beautician’ discussed in section 2.3 is constructed on a similar principle. However, given the environment in which it occurs in the text, ‘mortician’ acquires new significance. In the three-line cluster which precedes the last line, the comparator ‘more’ is repeated three times, ultimately receiving the large case which is reserved for only a very few words in the poem. This linguistic context projects a second reading for ‘mortician’: one whereby it may be interpreted as a complex derived from the items ‘more’ and ‘tician’. The literal meaning of this complex would be someone who practises the concept ‘more’, or to push the pun a little further, a ‘tician’ of ‘moreness’. As to the thematic significance of this complex, it plays presumably on the opposition between ‘more’ and ‘less’ upon which the bulk of the meaning of the text rests. The constant play on these opposing values, heightened through the use of upper case and culminating in the pun of the final line, constitutes a fairly direct critique of acquisitiveness, consumerism and avarice. Readers interested in extending the scope of this critique may wish to argue, on the basis of their analysis, that there are still greater significances embedded within the text. It may perhaps be interpreted as a covert attack on western capitalism, a disguised call for responsible citizens to reduce their living standards, or, simply, as an impassioned plea to everyone to cease to be ‘ticians’ of ‘more’. So far in the analysis, quite a lot of attention has been paid to the formal properties of the morphological and graphological devices of 51
FROM SHAPES TO WORDS
the poem. But what can be said about the function of these devices and how dependent upon them is the meaning which is projected by text? One important consequence of the intersection between these unusual patterns of graphology and morphology is the effect it has on the reading process. A type of ‘game’ developed, in which two reading strategies are brought into competition with one another. The first of these is the familiar line-by-line reading strategy, where the eye makes gradual progress vertically down the page. However, this strategy is challenged by the morphological fragmentation noted above where vertical processing is disrupted, and where instead a kind of ‘horizontal’ processing is encouraged. On this horizontal axis, lines may be interpreted individually as little isolated statements. Because of the boundaries strategically placed around them, sequences like ‘not to men’ at line 5 look at first glance like self-contained units of meaning. Further progression down the vertical axis will lead to the reconstruction of ‘mention’ from ‘men’, and so the horizontal reading will eventually be overridden. Of course, this reading game is completely invalidated if any attempt is made to read the poem aloud— such linguistic experimentation is almost entirely dependent on the visual system of language. To that extent, it is ‘poetry of the eye’ and not ‘poetry of the voice’, with cummings himself reputedly remarking that his work was meant to be seen and not heard (Dumas 1974:72). In order to give some idea of how dependent the text is on its graphology, readers may care to try to read the poem aloud for themselves. You will, I imagine, be faced with two mutually exclusive options. The first of these would be to attempt it by preserving the horizontal axis at the expense of the vertical. Line endings could be interpreted as pauses and so each line would be read in isolation, with no attempt to reconstruct the text on the vertical axis. This would sound extremely odd and be virtually unintelligible to any audience who did not have the written version in front of them. Moreover problems of pronunciation are likely to be encountered with sequences like ‘tion least)&i’ and ‘ybody’. In adopting the second reading option, the text could be reassembled so that the disruptions were removed and be read as if none of the graphological and morphological devices existed. But effectively (re)punctuating the poem defeats entirely the purpose and function of its original stylistic structure. More importantly, this (re)punctuation of the text would bring it close to standard punctuation and orthography, producing the following less than exhilarating aphorism: 52
FROM SHAPES TO WORDS
You notice nobody wants less (not to mention least), and I observe nobody wants most (not, putting it mildly, much). Maybe because everybody wants more and more and still more. What the hell! Are we all morticians?
A less than profound critique perhaps of the social mores which underpin western culture. If anything, such a reconstruction reveals the amount of stylistic currency that can be gained by the exploitation of language’s visual medium. Now stripped of its key linguistic devices, the text becomes, literally, prosaic, and the message which it projects is rendered bland and unimaginative. Once the tricks of the eye are removed, the poem’s dependency on the stylistics of the written mode—perhaps at the expense of everything else—is revealed. The rendering down of a ‘deformed’ text into a more prosaic version is a practice which can be extended productively in a number of different directions. It facilitates, for example, a kind of stylistic reversal, enabling the conversion of straightforward texts into defamiliarised ones. In other words, if a list of suitable morphological and graphological operations can be collated, then this list can form the stylistic blueprint for creative and amusing ‘deformations’ of language’s written mode. The next subsection will outline a practical language-based activity which is based precisely on this principle and which is designed to enrich knowledge about the morphology and graphology of English further.
2.4.2 Extending the analysis: a workshop It might be useful to begin this subsection by summarising the stylistic strategies which were teased out in the course of the analysis in the previous subsection. Isolating an inventory of key linguistic devices allows us to explore the ways in which these operations may be performed on other texts. Here then is the basic stylistic ‘blueprint’ for the cummings text, set out in the form of a list of morphological and graphological axioms: 53
FROM SHAPES TO WORDS
Strategy 1 Disrupt predictable patterns of large and small case, placing heavy emphasis on the use of small case in unusual environments. Strategy 2 Deploy extensively features that are exclusive to language’s written mode, making particular use of logograms and parenthetical symbols. Strategy 3 Break up lines of connected prose to form shorter, grammatically incomplete segments; arrange these segments in a columnar format. Strategy 4 Split up words into smaller particles. These particles may be anything from two ‘meaningless’ segments to two particles which resemble genuine words in the lexicon. Deviation in language remains deviant for only a limited period of time, and when disruptive patterns become established in the text they begin to assume a kind of norm of their own. Once this ‘norm of oddity’ is established, the way is prepared for a further type of stylistic exploitation. Strategy 5 subverts those very patterns which were striking in the first place: Strategy 5 Break any of the patterns established through the use of Strategies 1– 4. For example: (a) use upper case where the text has established a dominant pattern of small case; (b) increase line length or word length in the environment of shorter lines or words; (c) change the grammatical mood if a particular type has become the textual norm (e.g. declarative to interrogative).
¶
54
Once a stylistic blueprint has been abstracted from the analysis in this manner, it can be re-shaped into a set of creative and practical exercises. Here, for example, is a simple workshop activity that I designed for use in my undergraduate English language classes. The activity permits experimentation within the parameters of the blueprint derived from the stylistic analysis. This one is designed to supplement the analysis of cummings’s poem 55, although it can easily be adjusted to account for virtually any other text in the cummings oeuvre.
FROM SHAPES TO WORDS
1 Compose a short text of between twenty and thirty words on any topic you want. This may be a little narrative, an aphorism or even something as banal as a comment on the weather. 2 Working in groups, try to ‘defamiliarise’ this text, using the stylistic blueprint derived from the analysis of the cummings poem. 3 Assess the effects of your transposition. What impact does it have on the reading process and how much stylistic ‘mileage’ can be obtained from it? Is the transposition more ‘literary’ in character than the original text? Readers may be interested to know what to expect in terms of the outcome of such an exercise, so here are the results of a particularly successful session with a small group of second-year British undergraduates taking a language-based option in 1991. Of the eight short texts elicited through Activity 1, the following was selected for graphological and morphological embellishment. It is the musing of one participant on the important question of leather jackets and gravity: My leather flying jacket cannot fly. Neither can pianos fly and they are quite heavy. So how the hell does a jumbo jet fly? The group then embarked on Activity 2, applying the set of strategies 1 to 5 identified above. The result is given in the ruled text overleaf. Hopefully the broad principles which underpin the transformation of the text should be clear by now, so a detailed explanation of how the ‘poem’ was produced will not be necessary. Instead, a few general comments on the manner by which the blueprint has been implemented should help. Predictably, extensive use is made of lower case where upper case is expected. However, this pattern is itself subverted when upper case is deployed in unexpected environments (see ‘Fly’ and ‘Jack’ in lines 1 and 2). Features of the written mode of language, such as logograms and parenthetical symbols, proliferate in the text (e.g. lines 6 and 10). Sentences are broken up so that a columnar format is produced, while words are split so that a range of new forms emerges. Some lexical splits produce 55
FROM SHAPES TO WORDS
my leather Fly ing Jack et can not Fly n (either can Pia? no! s Fly &the y’are… quit. eh? eavy So Ho W.T. he hell! do es a jumb? o, jet! Fly…
¶
56
5
10
15
acceptable units, such as the break-up of ‘piano’ at lines 7–8 into the proper name ‘Pia’ and the negative particle ‘no’. Other splits produce meaningless particles, like ‘ing’, ‘et’, and ‘n’ at lines 2, 3 and 5 respectively. These lexical breaks often create linguistic trompes l’oeil, where the initial interpretations have to be revised in the light of later text. The seemingly independent unit ‘quit’ at line 12, for instance, needs to be assimilated to ‘quite’ in the light of subsequent segments. By the same token, the first line (‘my leather Fly’), which has the appearance of a conceptually complete unit, needs to be revised progressively in the light of the next two lines. Perceptual strategies are thus manipulated in a manner ¶ similar to the technique exploited in the original cummings poem. By way of a footnote, it is worth commenting on the fate of the ‘deformed’ text which was produced in the course of the present workshop. The student who was responsible for the original composition made a note of the group’s embellished version and then
FROM SHAPES TO WORDS
typed it up later. So taken was he with the product that he actually offered it to a university literary magazine whereupon it was promptly published! To my knowledge (and perhaps thankfully, given its origins) the true provenance of the ‘poem’ has never been divulged.
2.5 Summary As is consistent with the general rationale of the book, this chapter has sought to integrate linguistic theory and practice, offering a selection of critical tools which can be employed in textual analysis. In sections 2.2 and 2.3, a model for the analysis of graphology and morphology was developed. This framework was then applied to a short e e cummings poem in section 2.4. Appended to this analysis was a proposed workshop in stylistics, the object of which was to develop a fuller understanding of the theoretical model. When designing the linguistic model, it seemed good sense to make it as powerful as possible so that it would be able to handle graphological and morphological devices beyond those realised in a single short poem. Certainly, no one text will be likely to realise every possible permutation, so it is best to be prepared for other types of innovation. Furthermore, there are a number of types of morphological and lexical deviation which, although not exhibited by poem 55, occur repeatedly in the cummings ‘corpus’. A glance through collections of his work reveals a number of recurring stylistic formulae, all of which can be comfortably accommodated within the present model. An especially common strategy is the idiosyncratic use of derivational morphemes to derive unusual, ‘nonce’ complexes. The prefix ‘un’, for example, is regularly combined with normally incompatible roots to form complexes like ‘unstrength’ and ‘unbe’ (see section 2.3). Suffixes, if anything, are even more widely exploited by cummings. Commonly occurring bound morphemes like ‘ish’, ‘ness’ and ‘ly’ are tacked on to roots with which they would never normally conjoin, yielding constructions like ‘neverish’, ‘muchness’ and ‘happeningly’. This process may even be duplicated within single items: ‘silverlyness’ incorporates not one but two suffixes (‘ly’ and ‘ness’). An extension of this tactic is the use of grading suffixes along with roots which would not normally permit gradation. Thus, non-gradable terms like ‘last’ and ‘chief receive the superlative ‘est’ suffix to produce ‘lastest’ and ‘chiefest’. Even if the root is gradable, the ‘er’ suffix is often preferred 57
FROM SHAPES TO WORDS
when the stricture of grammaticality demands the intensifier ‘more’. On this pattern, the adjective ‘beautiful’ will not become ‘more beautiful’ but rather ‘beautifuler’. The compounding of one or more root morphemes is also common, resulting in constructions like ‘yellowgreen’ and ‘mostpeople’. Finally, new forms are also created by analogy with the process of back-formation (see section 2.3), with new verbal forms being derived from related nouns. This explains the conversion of ‘September’ into a verbal particle in the sequence ‘septembering arms of year extend’. The central aim of the chapter has been to use textual analysis as a means of obtaining insights into a complex feature of English language. Of necessity, literary-critical responses to aspects of cummings’s work have been largely overlooked. Now that cummings’s style has been examined in detail and other texts have been generated on the basis of his stylistic profile, readers may be interested in what the critics have made of cummings’s experimentation with language. The many essays on the style of cummings offer little agreement on the function of the linguistic techniques which have become his stock-in-trade. With particular regard to the devices identified in the course of our own analysis, critical commentary has been diverse to the point of absurdity. The use of parenthesis has probably received more attention than all the other written mode markers which cummings employs, and has also been subject to most diffuse and inconsistent interpretations. Kidder (1979:13–14), while presenting a set of rules on how to ‘read’ cummings, offers the following straightforward piece of advice: ‘Treat parenthesis carefully… On some occasions…we must ignore the parentheses.’ No less obscure is Baum’s (1972:112) claim that parenthesis suggests ‘the simultaneousness of imagery’ and highlights cummings’s ‘extreme honesty as a poet’. The extensive deployment of lower case has also become the subject of much comment, with a great deal of interest in the small case personal pronoun ‘i’. Watson (1972:40) points out that ‘the little naked swaggering “i” asserts its reference to one real and single being’, while Baum (1972:115) contends that ‘By rejecting the pronoun “I” Cummings [sic] assumes a casual humility’ which ‘dissociates the author from the speaker of the poem’. Much discussion has centred on cummings’s prolific use of logograms, which are often referred to by the critics as typographs or, erroneously, as ideograms. To Wegner (1965:157) they are simply ‘a moment of coalesced awareness’, but to Dumas (1974:72) they constitute ‘a means by which nature’s dynamic process may be 58
FROM SHAPES TO WORDS
glimpsed, and the glimpse be passed on to the reader, who may thereby achieve transcendental vision’. Readers may assess for themselves the validity of these interpretations, but on this evidence, it would appear that cummings’s experimentation with language generates a great deal of interpretative ‘space’. The diversity of critical commentary shows just how variable interpretations can be. It also problematises the methods by which critics evaluate innovation and creativity in language. While the remit of this chapter has simply been to provide a systematic introduction to graphology and morphology, at least foundations will have been laid for a method of critical interpretation—a method which is, moreover, predicated upon a thorough working knowledge of how linguistic systems operate in the context of modernist poetry.
Suggestions for further reading Graphology and morphology Readers who wish to investigate further the linguistic topics covered in this chapter should find the following material useful. Fromkin and Rodman (1978:138–59) offer a useful and accessible introduction to basic morphology, while a more technical and advanced study of the subject can be found in Matthews (1974). Two books specifically on writing systems are Sampson (1985) and Coulmas (1989). A comprehensive survey of word-formation in English is Adams (1973).
Stylistic applications An extremely good stylistic application of both morphology and graphology is provided by Traugott and Pratt (1980: Chapters 2 and 3). In addition to the specific references cited over the course of the chapter, stylistic explorations of this aspect of the language system include: Leech (1969:39–44), on ‘deviation’ in poetic language; Carter (1989), on stylistics and ‘concrete poetry’; Van Peer (1993), on graphology and concrete poetry. Van Peer has also written a stylistic analysis of e e cummings’s poem ‘yes is a pleasant country’ (1987). He uses a cognitive model of language to investigate the relationship between reading strategies and literary comprehension. 59
Words and meanings: an introduction to lexical semantics
• 3.1
Introduction
62
• 3.2
Words and meanings
64
• 3.3
Words and combinations
77
• 3.4
Techniques for stylistic analysis
84
• 3.4.1 Cloze procedure and stylistic analysis
85
• 3.4.2 Multiple choice text and stylistic analysis 93 • 3.5
Summary
• Suggestions for further reading98
96
Chapter 3
Chapter 3
61
3.1 Introduction That famous philosopher of language, Humpty Dumpty, has this to say on the meaning of words: ‘“When I use a word,” Humpty Dumpty said in a rather scornful tone, “it means just what I choose it to mean—neither more nor less.”’1 Were Dumpty’s principle to gain universal acceptance amongst language-users, the outcome would be linguistic anarchy. As speakers would cease to be accountable for anything they said, so a potentially infinite number of meanings could be mapped on to any single word they uttered. The linguistic ‘contract’ that governs everyday interaction would simply break down and any recognisable system of verbal communication instantly evaporate. For example, readers of this book may have happily assumed that they were reading an introduction to English language. Yet were its author to invoke the Dumpty principle, all sorts of meanings could be ascribed to it on a whim: it could become, variously, a discussion of bee-keeping, a treatise on the life cycle of the mackerel, a discourse on medieval hygiene and so on. Outside the interactive world of Alice’s Wonderland, the Dumpty principle does not, of course, obtain, and interlocutors can rest assured that there is at least some interactive consensus about what a word means. Although never absolute or immutable, words routinely designate specific objects, processes and concepts, and these designations form part of the very bedrock of coherent human communication. Explaining precisely what, and how, a word means will be the main aim of this chapter. There will follow an outline of the principal types of lexical meaning in English, as well as an account of the ways in which word-meanings interact with one another. Poetry will, again, form the nucleus for much of the discussion of this aspect of the language system, although it will be stressed throughout that creative or striking patterns of meaning are a feature of many types of discourse. 62
WORDS AND MEANINGS
It has never been easy to attach an exhaustive, self-contained meaning to a word which will account sufficiently for its use in all contexts. Decisions about where exactly to place a boundary on meaning have exercised linguists and philosophers a great deal; indeed, even the meaning of the concept ‘meaning’ has itself come under scrutiny in the relevant literature. The task of specifying what a word means is influenced by at least four key factors. First of all, meanings shift over time, which is why dictionary-makers often have difficulty keeping up with the meaning of a word in its current usage. A case in point is the word chauvinist. Whereas its ‘official’ meaning designates someone who is distrustful of foreigners, it rarely means anything other than ‘sexist’ when applied in popular usage. A second factor is that the meaning of a particular word may vary in relation to the context in which it is used. Take, for example, the word terrific. Like chauvinist, this term has undergone a change over time: in this instance, the original meaning of ‘instilling terror’ has been replaced by its current function as an expression of high approbation. Yet even in current use, the term displays considerable diversity of meaning. Where it is wholly positive under one set of conditions (‘That film was absolutely terrific!’) the very opposite will be intended by its use in a context like: ‘Terrific! You’ve just erased all my data from the computer!’ A third factor affecting word-meaning arises from the cross-cultural differences that may pertain between users of the same language. For instance, if you are deemed politically liberal in the United Kingdom you can at least consider yourself a tolerant individual; yet to be called so in North America carries with it no such guarantee of integrity. The fourth and final factor concerns the many pairs of words in English which display seemingly identical meanings. For instance, while brotherly and fraternal purportedly ‘mean’ the same thing, their privileges of occurrence—as well as the varying degrees of formality which each conveys—are in no way identical. It is factors such as these which have polarised debate about wordmeaning into two main schools of thought. One school argues for a strictly context-free theory of meaning which is divorced from actual usage; the other for a contextually informed model which examines words in their natural environments of use. In view of our stylistic emphasis, both approaches will need to be represented if our basic schema is to provide a workable and coherent account of words and 63
WORDS AND MEANINGS
their meanings. This chapter therefore tries to establish a model that accounts for core meanings but which still grants usage and context a say. The first step towards this double goal is taken in the next section, where a survey of the principal types of word meaning in English is undertaken. The perspective widens in section 3.3 as attention turns to the types of meaning patterns that are created when words combine with other words. Section 3.4 proposes a set of techniques for stylistic analysis and applies a selection of these techniques to literary texts. The concluding section, section 3.5, suggests a set of extensions to the analyses and assesses the theoretical implications of semantically orientated studies of wordmeaning.
3.2 Words and meanings Semantics is the study of meaning. The study of meaning encompasses units of language which vary in size, from parts of words and sounds right up to whole clauses and sentences. The basic unit of analysis in semantics is the seme and this term is often appended as a suffix to other structural terms to indicate which precise unit is being described. Morpheme, grapheme and phoneme, as we saw in the previous chapter, are units which are thus derived. The term lexical semantics is specifically reserved for the study of word-meaning, while the term for a unit of meaning in lexical semantics is lexeme. Before embarking on a survey of lexical semantics, an important caveat needs to be heeded and it is this: there is not always a one-toone correspondence between lexemes and words. This unfortunate asymmetry arises largely because of the special case of idioms and related constructions in English.2 An idiom may comprise many words, but it realises just one lexeme. Take, for instance, the following colloquial expression: 1 he kicked the bucket Even though bucket and kick would constitute two lexemes in most contexts, their individual meanings are nullified within the restricted semantic domain of this idiom. In other words, it is not the individual component words of (1) but the entire sequence ‘kicked the bucket’ 64
WORDS AND MEANINGS
which corresponds roughly to the single unit of meaning ‘died’. The same principle is at work in this next set of idioms: 2
a red herring
3
it’s raining cats and dogs
4
to go for a song
5
as sick as a parrot
6
to give up the ghost
These constructions (and many others like them in English) express meanings which cannot be directly accessed from the sum of their constituent words. Rather, interpretation rests on the understanding that idioms are fixed expressions and that, moreover, rigid constraints govern the manner by which their component words are selected and combined. Idioms cannot therefore be rearranged grammatically or embellished with other compatible items. If they are, the results are distinctly odd: 3(a) it’s raining dogs and cats 3(b) it’s pouring cats and dogs 3(c)
it’s raining poodles and ginger toms
Furthermore, idioms are normally language-specific and will collapse if translated literally into another language. As a brief test of this thesis, you could attempt to translate the examples above into any other language with which you are familiar. Here, for instance, are some over-literal conversions into French: 1(a) II a donné un coup de pied dans le seau 2(a) un hareng rouge 3(d) il pleut les chats et les chiens 4(a) aller pour une chanson 5(a) malade comme un perroquet 6(a) abandonner le phantom 65
WORDS AND MEANINGS
This ‘franglais’ illustrates well the patent nonsense—if not the humour!—that results from the literal translation of idioms from one language into another. Most traditional accounts of lexical semantics distinguish between the sense of a word and the referent of a word. Sense is the essential property or ‘core’ meaning which a word exhibits irrespective of its context of use. For example, the word dog will encompass all members of this class of canine animal. Furthermore, the sense expressed by the term will carry enough essence of ‘dogness’ to differentiate it from, say, words for non-canine animals like cat, horse or pig. By contrast, the referent of a word is the actual entity picked out or identified in a context of use. Thus, when the owner of a pet says 7 I think I’ll take the dog for a walk it is clearly not an abstract, general class of animals, but a particular, unique and individual dog to which she refers. Therefore, while sense captures the ways meanings are organised and structured within a language; reference accounts for the ‘real-world’ objects, concepts and processes that exist outside the language system. One of the primary tasks of traditional dictionaries3 is to capture the generic, all-encompassing properties of words of different senses. A dictionary definition will therefore seek to explain what a word denotes regardless of the context in which it occurs. While this concept of denotation takes us part of the way, clearly it still does not provide an exhaustive description of what a word ‘means’. As we noted in the introduction to this chapter, the sense of a word may shift over time, so dictionaries can never be looked upon as repositories of stable, immutable word meanings. Equally importantly, there is often attached to a word a set of resonances which are very difficult to accommodate into a description of its sense. These resonances form a kind of connotational aura around a word, giving it a layer of affective, associative meaning which is quite distinct from the layer of core meaning which we have called denotation. This second, associative layer is normally termed connotation. The distinction between connotation and denotation can perhaps be best appreciated when it is examined in tandem with another feature of lexical semantics. This feature is the principle of synonymy in language. Synonyms are simply words which denote the same 66
WORDS AND MEANINGS
thing. For instance, mercury and quicksilver clearly denote the same substance, movie and film the same object, hide and conceal the same type of activity. The appearance of pairs of synonyms in a particular language is normally a result of lexical borrowing from other languages. As a Germanic language which has borrowed extensively from Romance languages, it is no surprise then that English is rich in synonyms. Here is a selection of pairs of English synonyms with the respective languages of origin indicated:
Germanic sweat buy climb friendly get whore
Romance perspiration purchase ascend amicable obtain prostitute
Even triple synonyms can be found:
Germanie kingly
Latin regal
French royal
The two Romance terms in this triplet reflect different periods in the historical development of English, with regal borrowed from the Latin of the first millennium and royal from the French of the postNorman-conquest period. Now, the significant thing about synonymy, when superimposed on to our earlier distinction between denotation and connotation, is that while the words may ‘mean the same thing’ in the strictest sense, they may carry markedly different affective resonances. In other words, synonyms may share denotation but exhibit different connotations. Moreover, the choice of a particular term is often subject to the constraints of register (see section 1.3), so the switch from one member of a pair of synonyms to another can mark a concomitant and significant shift in tone. The rule of thumb is that the Romance-derived items convey greater formality and the native Germanic items greater informality. In discourse, whereas Latinate derivations may help engender a more lofty or elevated tone, Germanic terms tend to have a more prosaic or ‘down to-earth’ feel. It is no coincidence, for example, that everyday words describing parts of the human body are Germanic in origin; just as it is no 67
WORDS AND MEANINGS
coincidence that in the substantially more formal register of a medical encyclopaedia these terms are replaced by their Latinate counterparts. In this set of examples it should be clear which members of each pair are Germanic and which Latinate: skull thigh belly kneecap
cranium femur stomach patella
The convergence of register with lexical choice means that terms are selected in accordance with discourse context. Consequently, speakers and writers always run the risk of drawing on an informal item when a more specialist or auspicious term is expected, or conversely, of using an overly formal item when a less ‘wordy’ form is required. Consider for a moment the following triplet of synonyms, ranked in order of descending formality: expectorate, spit, gob. Now in the context of an encounter with your dentist, any instructions regarding oral evacuation would presumably be couched in the medial term: 8
Please spit into the basin.
The mismatches accruing from the use of the other two terms in the same context would either be absurdly pompous: 9
Please expectorate into the basin,
or irreverently familiar: 10 Please gob into the basin. For an attested example of a mismatch between register and lexical choice, consider example (11). It is a notice from an Austrian hotel catering specially for skiers. Notice how the Latinate selections are particularly inappropriate, signalling a formality way in excess of that demanded by the context: 11 Not to perambulate the corridor in the hours of repose in the boots of ascension. 68
WORDS AND MEANINGS
In summary, it can be said that synonyms are words which share the same sense and which denote the same thing. However, connotational differences mean that most synonyms are not mutually interchangeable and that their use is closely associated with register and discourse context. Pure synonyms, which exhibit perfect congruence of denotation and connotation, are therefore rare. Nevertheless, the principle of synonymy offers a useful blueprint for explaining the general proximity of meaning that exists between many words in a language.4 The next stage of our survey of lexical semantics focuses on words which have more than one meaning. Two types of process are responsible for producing words with multiple senses. The first type is easily illustrated when we look at a word like bank. In current usage, this word has two principal senses: it can denote either the side of a river or a financial institution. The historical provenance of this double meaning can be explained in the following way. Two separate words found their way into English from different source languages. The first of these was bank, of a river, from a probable Scandinavian origin; the second bank, in the financial sense, from Old French. This process, where different words converge into the same form while preserving their original meanings, is known as homonymy and the words involved are referred to as homonyms. Here are some more English homonyms: ear mail sage seal grave
(i) (i) (i) (i) (i)
of corn; (ii) of a person armour; (ii) postal service wise; (ii) culinary herb marine mammal; (ii) insulation or fastening burial area; (ii) serious or weighty
As it results from the distillation of different words into the same shape, homonymy is a rather chance process. None the less, dictionaries will still attempt to provide separate entries for homonyms, highlighting their status as genuinely distinct words. The second process responsible for producing multiple meaning is when the semantic scope of a single word is extended. Unlike homonymy, the meaning of an individual word radiates outwards from a single source, with its sense undergoing a kind of metaphorical extension. In spite of the semantic divergence which results from this process, the relationship between the senses can 69
WORDS AND MEANINGS
still often be inferred. This principle of extension is known as polysemy and a word whose semantic range has widened in this way is referred to as polysemous. Here is a selection of examples:
flight neck score eye drip
(i) of bird; (ii) of a dart or arrow; (iii) act of escaping (i) of bottle; (ii) of body; (iii) of guitar (i) of music; (ii) of contest (i) of body; (ii) of needle (i) of water; (ii) weedy person
When the connections between the different senses of a word are more opaque, it is sometimes difficult to differentiate between homonymy or polysemy, and so solving the semantic riddle often requires a good etymological dictionary. Consider, for instance, the two principal senses of the word sole, where one designates the bottom of a shoe and the other a species of fish. On the face of it, the word could either be homonymie or polysemous. It is actually the latter, because the word stems from the Latin source word solea meaning a sandal or flat shoe, and has since been extended metaphorically beyond its original application to footwear. (A moment’s reflection on the anatomical structure of the fish in question should reveal the connection!) A trickier task again would be sorting out the pair port (harbour) and port (fortified wine). On first glance, these two look distinct enough in sense to be homonyms; that is to say, two separate words which happen to have the same shape. However, the second term has developed because it refers to a drink from Oporto, a city so named because it is itself a harbour. So a brief etymological search reveals that what looked initially like distinct homonyms are, in fact, extensions of a single, polysemous word. Figure 3.1 is a summary of the semantic pathways which lead to polysemy and homonymy. Polysemy and homonymy facilitate creativity in many types of discourse. The figure of speech with which they are most commonly associated, given their potential for word-play, is the pun. Literary discourse, which places great emphasis on controlled and motivated ambiguity, is an obvious site for such linguistic experimentation. Nevertheless, puns are a feature of many other everyday discourses, perhaps the most notable of which is advertising language. In designing messages about products, advertising copywriters draw heavily on puns to produce the linguistic ‘jingles’ which are at once 70
WORDS AND MEANINGS
FIGURE 3.1
striking and memorable. Here are two short illustrations of how this technique is employed. The first is taken from an advertisement for a ‘power’ shower-unit, and the second from an advertisement for a particular brand of bathroom sealant. The second text also carries the logo of a seal (of the animate variety!) balancing a ball on its head: 12 You’ll feel a drip under anything else! 13 See how well our seals perform! The words which activate the puns here are ‘drip’ and ‘seals’, creating plays on meaning which are polysemous and homonymie respectively. In anticipation of these examples, both items were included in the lists of homonyms and polysemous words provided earlier so they can be checked there. Puns may be intentional or accidental. In the next example, a headline from an Irish newspaper,5 it is not clear whether the homonymy exhibited by ‘grave’ is consciously exploited or whether it is just one of those semantic coincidences which people are apt to call ‘Freudian slips’: 71
WORDS AND MEANINGS
14 DUBLIN HAS GRAVE CEMETERY PROBLEM The next pair of rather hapless examples6 will conclude our survey of polysemy and homonymy. The first is a notice to the clientele of a Mexican hotel, while the second advertises donkey rides in Thailand. That any semantic play is intentional in either case is unlikely in the extreme: 15 The manager has personally passed all the water served here. 16 Would you like to ride on your own ass? Next up in our survey of lexical semantics is the important principle of ‘oppositeness’ in language. Semantic oppositeness is called antonymy, and pairs of words that have opposite senses are referred to as antonyms. However, antonymy is a rather broad, catch-all category which embraces at least three distinct subtypes of semantic opposition. The first, complementary opposition, is the most strictly ‘logical’ of all lexical antonymy. If one member of a pair of complementary opposites is true, then the other must, by imputation, be false. Cancelling one term will, in other words, automatically guarantee the truth of the other. For example, in the complementary pair: present
absent
only one of the two terms can be a valid representation of a particular state of affairs. So, if you are present, you cannot be absent; conversely, if you are absent, then you cannot be present. The same principle underpins all of this next batch of complementary opposites: asleep married legal alive
awake unmarried illegal dead
Gradable antonyms, which evince a different type of antonymy, are particularly common in English. These opposites are so named because they can be plotted along a graded scale according to varying degrees of the property or quality they specify. For example: 72
WORDS AND MEANINGS
old big wide good
young small narrow bad
A useful test for gradable antonyms is to see if they can be modified with adverbs like very, quite or rather, as in ‘very old’, ‘rather wide’, ‘quite good’ and so on. As the qualities designated by gradable antonyms are relative rather than absolute, it is sometimes possible to embed pairs of gradable antonyms inside other pairs. Thus: hot
cold warm
cool
In keeping with other sorts of lexical meaning, context will be an important determinant on the interpretation of gradable antonyms. Given that size, quality and proportion are all relative, no permanent or fixed value can be attached to any term on the graded scale. In other words, the denotation of concepts like hot and cold, big and small, may shift depending on the nature of the non-linguistic referents which they describe. A cool oven, for instance, will be much warmer than a hot day; just as the biggest mouse will still be very much smaller than the smallest elephant. The third subtype of lexical antonymy is relational opposition. Relational opposition is perhaps best conceived of as an alliance of ‘converseness’ rather than of true ‘oppositeness’. Here are some examples: give buy teacher parent below
receive sell pupil child above
The distribution of relational opposites is often determined by the type of sentence in which they occur. For example, if give appears in a sequence like ‘Fatima gave chocolates to Ibrahim’, then its opposite can be used when the sequence is reversed: ‘Ibrahim received chocolates from Fatima.’ Similarly, if Maria is Juan’s teacher, then Juan is Maria’s pupil, or, if the painting is above the fireplace, then the fireplace is below the painting. 73
WORDS AND MEANINGS
Earlier in this section, mention was made of the elusiveness of ‘pure’ synonymy. In many respects, those same reservations can be expressed about antonymy. Contextual factors exercise a powerful influence on the associative and affective values that attach to each member of a pair of antonyms. One especially sensitive aspect of lexical antonymy, which has come under the scrutiny of feminist linguists, concerns pairs of gender-differentiated opposites such as: bachelor governor master patron
spinster governess mistress matron
It is significant that the female member of each pair displays a degree of negativity which is not carried by the more positive male term. This gender-determined asymmetry is known as semantic derogation and it arises when a word acquires pejorative or inimical connotations simply by dint of its exclusively female reference. Deborah Cameron elaborates: ‘Whereas the male terms connote power, status, freedom and independence, the female, which in many cases used to be parallel, now connote triviality, dependence, negativity and sex’ (1985:77). Clearly, semantic change doesn’t happen in a cultural vacuum. Feminist linguists argue that the semantic derogation of words for women reveals just as much about the history of patriarchy as it does about the history of language. The final type of word meaning to be examined in this section concerns classes of items which are semantically compatible. This type of sense relation is known as hyponymy and differs principally from synonymy in that it invokes the concept of ‘inclusion’ rather than ‘sameness’ of meaning. Hyponymy explains how the sense of one item can sometimes be included in the sense of another. For instance, the item flower includes other classes of items such as tulip, rose and daffodil. However, the term flower can itself be included within the sense of a higher, more general term like plant. The semantic categories in hyponymy are defined as follows: the included, more specific terms are hyponyms (literally, terms that are ‘under’ something else), while the encompassing, higher terms are superordinates (literally, terms that are ‘over’ or ‘above’ something else). Hyponyms and superordinates can be set out systematically on a semantic scale. Figure 3.2 illustrates how a term like dog enters into 74
WORDS AND MEANINGS
a semantic relationship with other related terms. Any lower term on the scale is a hyponym and any upper term a superordinate. This means that dog fulfils the two functions simultaneously, being both a hyponym of canine and a superordinate of poodle and spaniel. Lower terms are always instances of terms higher up; indeed, part of the semantic specification of a term depends upon whichever higher terms can enclose it. The scale also indicates how a lower term entails a higher term, so that the relationship between poodle and dog can be expressed with the proposition: all poodles are dogs. However, as dog is a superordinate of poodle, then the reverse proposition does not hold. Consequently, it is not true that all dogs are poodles. ¶
FIGURE 3.2 As always, it is important to stress how context shapes the way language is used. Semantic scales offer an illuminating explanation of how and why speakers make strategic use of particular lexical items in particular circumstances. Their semantically inclusive nature means that hyponyms and superordinates can be used for a variety of interactive purposes. Clearly, appropriateness to discourse context will be a determining factor. For instance, if someone asks ‘What’s that?’ at a dog show, the reply ‘A dog’ will not suffice. In order to match the degree of detail which the context demands, much greater lexical specificity7 is required. The reply would therefore need to be something of the order ‘An Irish Water Spaniel’ or ‘A King Charles V Spaniel’, both of which contain hyponyms significantly lower on 75
WORDS AND MEANINGS
the scale for dog. Other factors influence the way hyponyms and superordinates are chosen from semantic scales. Lexical specificity is an important strategy-framing device and often shapes the pragmatic force of an utterance. In the following exchange, notice how speaker B is deliberately under-specific: 17 A: What have you got in that case? B: A musical instrument. B’s reply clearly incorporates a term which is markedly less specific than is demanded in the context. And even though it still answers the question truthfully, this under-specification reveals much about the speaker’s attitude to the speech situation itself. It suggests in this instance that some secrecy is to be attached to the contents of the case, or alternatively, that this is a topic of conversation which should not be pursued. Here are two more examples where superordinates are exploited in a similar way: 18 A: Where are you going? B: Out. 19 A: B:
What did you buy in town? A garment.
Again, by offering less than is anticipated by the question, the replies convey a certain ‘standoffishness’ or interactive reticence.8 Here now are some examples which, for the sake of contrast, make explicit use of lexical over-specification: 20 Siobhán let me drive her new 7-series BMW. 21 I play a ‘59 Stratocaster with Seymour Duncan humbuckers. Where ‘car’ and ‘guitar’ would probably suffice for most contexts of (20) and (21) respectively, the utterances here draw on markedly over-specific hyponyms. This serves to emphasise any included feature or extra quality that is imported by the hyponym. The semantic content of (21) may therefore be designed to convey expertise or specialist knowledge, while (20) may be seeking to foreground the special quality of ‘BMWness’ which differentiates it 76
WORDS AND MEANINGS
from ordinary cars. Whatever the intended goals of these particular utterances, they illustrate clearly how extra ‘pragmatic’ meanings can be attached to words simply through their natural contexts of use. This section has focused on the principal categories and typologies of meaning in lexical semantics. By concentrating on the meanings that inhere in words, we have to some extent neglected the types of meaning that derive when words are combined with one another. As a counterbalance to this section, then, the next part of our survey will suggest a framework for the analysis of the more complex meanings that are produced when words are combined with other words.
3.3 Words and combinations One of the traditional ways of explaining how words combine into patterns is by invoking the concepts of ‘choice’ and ‘chain’. These concepts are normally represented diagrammatically as two intersecting axes, as in Figure 3.3. The syntagmatic axis shows how words combine and which sort of structural relationships link them. The paradigmatic axis explains how certain words in the lexicon can be selected at particular points along the chain and why others would be inappropriate. The syntagmatic axis thus forms a structural frame which informs the selectional possibilities offered by the paradigmatic axis. By the same token, the paradigmatic axis offers a pool of potential lexical entries for each slot along the syntagmatic axis. At a more conceptual level, syntagmatic relations between words are of the order: a and b and c and d; and paradigmatic relations of the order: a or b or c or d. The following simple phrase, from which the final item has been removed, will provide our first illustration of how the two axes function: 22 my sunburnt ____ First of all, a structural frame is established by the syntagmatic combination of ‘my’ and ‘sunburnt’. This configuration sets up strong predictions about the nature of the next lexical item that can be selected on the paradigmatic axis. In this instance, the frame 77
WORDS AND MEANINGS
FIGURE 3.3 Choice and chain anticipates an anatomical noun. More specifically, it requires a noun which denotes a part of the body that can be exposed to the sun. The pool of selections available on the paradigmatic axis is therefore likely to be restricted to the following set of entries: 22(a) my sunburnt face arms nose legs There is an oft-quoted adage in language study which goes: ‘You know a word by the company it keeps.’9 Translated into semantic description, this axiom suggests that a significant chunk of the meaning of a word will be derived from the syntagmatic relationships into which it conventionally enters. As example (22) illustrates, a full semantic profile of ‘sunburnt’ will clearly require some account of the types of lexemes with which it regularly combines. This principle of lexical combination is known as collocation. Collocation refers broadly to the grammatical combination of lexemes, while the term collocate is used to describe any word which exhibits a standard pattern of co-occurrence with another word. Given that the syntagmatic axis sets up strong structural constraints, collocates are often easily predicted. You may verify this for yourself by completing the following sequences: 78
WORDS AND MEANINGS
23 These ____ are addled. 24 Here comes a bay-coloured ____. 25 Throw out that rancid ____. 26 It’s a particularly ferocious pride of ____. Suitable collocates for ‘addled’, ‘bay-coloured’, ‘rancid’ and ‘pride’ would be ‘eggs’, ‘horse’, ‘butter/bacon’ and ‘lions’ respectively. These potential selections from the paradigmatic axis underline the predictive power of the structural frame, showing how items in a syntagmatic chain generate strong expectations about the sorts of items that will follow. The principle of collocation helps explain why words occur in the sequences they do. Yet equally importantly, this principle has the capacity to deal with what happens when words do not occur in familiar combinations and when lexical predictions are not fulfilled. To illustrate this, we will need to re-introduce example (22). However, the phrase is now reproduced within its original context, which is the ninth line of the first sonnet of Sir Philip Sidney’s poem sequence Astrophil and Stella (circa 1582):10 22(b) Some fresh and fruitful showers upon my sunburnt brain None of the standard collocates suggested above are realised here; instead, an unlikely collocation is produced which does not meet the predictions of the structural frame. The combination of the semantically disparate items ‘sunburnt’ and ‘brain’ creates what is known as a collocational clash. If ever there was a profoundly simple linguistic technique which has enormous and far-reaching stylistic potential, it is the collocational clash. This type of semantic mismatch underpins many of the creative turns of language which are traditionally referred to as metaphors, images or figures of speech. Moreover, as it is built on the juxtaposition of lexemes with incompatible or contradictory senses, the collocational clash also forms the semantic core of the classical rhetorical trope called oxymoron. Clearly, poetry offers a rich locus for the exploitation of collocational clashes. In (22b), for example, there are two within the space of a single line: ‘sunburnt brain’ and ‘fruitful showers’. Yet clashes figure in genres other than poetry and are, perhaps surprisingly, a 79
WORDS AND MEANINGS
commonplace of many types of everyday discourse. In order to demonstrate this, I will introduce and develop a short stylistic exercise. This exercise, called ‘shunting’ (for reasons that will become apparent soon), offers a useful tool for studying language patterns on a number of different levels. (See also section 5.2 below.) The first stage in the exercise is to devise a set of straightforward syntagmatic combinations. These combinations should contain two elements each: a+b. Here is a set of nine unremarkable syntagms: stone masons sad dream shredding machine tangerine pips purple roses sick joke killing fields soft rain unhygienic café Some readers may have already spotted the deliberate ‘plants’ in this collection, but we shall proceed with the exercise none the less. The next stage involves taking the b elements of each pair and shunting them randomly around the set. Whatever new a+b pairings result from the shunt, there are likely to be collocational clashes. Here are six new possible configurations exhibiting unlikely or improbable pairings: stone roses purple rain
¶
80
sad café killing joke
soft machine tangerine dream
In fact, these six clashes all turn out to be well-attested constructions in English. They are all either names of popular rock and pop groups (e.g. ‘Stone Roses’, ‘Tangerine Dream’) or the names of album or song titles (e.g. ‘Purple Rain’ by Prince). What the shunt has unveiled is a formula that is used—perhaps subconsciously—by many musicians when they attempt to create striking and memorable monickers for their bands. Given the ephemeral and transient nature of the rock and pop genre, the need to make a lasting psychological impact is all the more pressing. Startling combinations of lexemes clearly possess greater mnemonic (memory-making) potential than typical collocations, so it is not surprising that the collocational clash has become the preferred stylistic strategy of many rock and pop artists. As our collocational shunt has not yet been completed, here are three suggested pairings using the items which remain unaccounted for. Any student readers, with musical aspirations and whose band has still to be christened, are welcome to use them:
WORDS AND MEANINGS
The Shredding Fields The Unhygienic Pips The Sick Masons A similar shunting exercise can easily be performed on the set (23–6) above, yielding some further intriguing clashes. The crucial point to remember is that while the individual component words would rarely pass muster on their own, when they are forced to keep company with unfamiliar partners the results are often notably bizarre. Again, aspiring rockers may draw on this collection as they feel fit: 23(a) [the] rancid lions 24(a) [the] bay-coloured bacon 25(a) [the] addled horse 26(a) [the] pride of eggs We shall look shortly at the ways in which this essentially simple principle informs linguistic creativity in poetry. Before doing so, we shall close this section by examining one of the principal methods used for recognising and accounting for collocational clashes. The question that will be addressed is: what is there in a word that tells us about the company it keeps? One of the first attempts by linguists to explain semantic compatibility (and by imputation, semantic incompatibility) between lexemes is known as componential analysis.11 This type of analysis involves the sifting out of the subcomponents of meaning that make up the overall sense of a lexical item. For example, a componential analysis of a lexeme like boar will consist of a series of smaller, particularistic descriptions about what sort of entity it is. Boar supports the following descriptions: it is a pig; it is an adult; it is male. The formal method for breaking down lexemes into constituent components is to lay them out systematically using plus and minus symbols for each attribute. Hence: boar:[+ animate]; [+ porcine]; [+ adult]; [- immature]; [+ male]; [- female] The more components there are in the description, then the greater the likelihood that there will be some redundancy between them. If 81
WORDS AND MEANINGS
the attribute [+ porcine] is invoked, the quality of animacy will be presupposed and so the component [+ animate] need no longer be included in the list of attributes. Similarly, as the component [+ male] presupposes the component [- female], then the second piece of information can be deleted. One of the main benefits of the componential method for stylistic purposes is that it can isolate the specific semantic features that activate collocational clashes. For example, in the combination pregnant boar it is the incompatibility between the component [+ male] for boar and the component [+ female] for pregnant that produces the anomaly. Other components, such as [+ adult] or [+ porcine], are not responsible for this particular mismatch. However, in a sequence like pregnant piglet it will indeed be the criterion [+ adult] / [- adult] which triggers the clash. In dealing with a combination like pregnant table isolation of the higher-order components [+ animate] and [- animate] should be enough to explain the anomaly. Here now are some fuller sequences containing collocational clashes. The significant semantic components of the mismatched items are displayed next to each example: 27 The concept crushed the boulder concept [+ abstract]
boulder [+ concrete] 28 I drank my computer
drank [+ liquid] computer [- liquid]
29 This bachelor is a good husband bachelor [- married]
husband [+ married] 30 That woman is the father of two boys 82
woman [+ female]
father [- female]
WORDS AND MEANINGS
As a handy means of explaining how parts of words exercise constraints on collocation, componential analysis is certainly useful as far as it goes. However, sceptics (and there are many) have argued that the theory simply doesn’t go far enough. Since its initial development, a substantial amount of criticism has been directed towards different aspects of the model. For a start, its early practitioners had an analytic habit which did little to win favour with feminist linguists. This was their predilection for describing lexemes like woman, wife or girl with the component ‘[- male]’. A typical breakdown of example (30), for instance, would have been:
father. [+ male] woman: [- male] Concepts like ‘female’ and ‘feminine’ were consistently downgraded to the attribute ‘absence of maleness’ rather than ascribed any positive quality of their own. However, in my opinion this sexist practice is largely a fault of the shortsightedness of the analysts, rather than, as has been suggested by some feminists, an inbuilt sexist principle within the system of language itself.12 Another problem concerns the highly selective manner by which sample items are chosen in componential analysis. The types of lexical item which lend themselves well to this sort of semantic breakdown tend to exhibit stable, generic properties within clearly defined semantic scales (see section 3.2 above). Items denoting members of the animal kingdom have proved especially popular because they can often be differentiated through convenient superordinates, such as bovine, porcine or human, which pick out the genus to which the terms belong. Lexemes denoting abstractions, such as beauty, truth or fate are largely avoided. It would indeed be difficult to account for the opposition between antonyms like truth and falsity within the componential framework. All that really can be done is to prefix each term with a minus symbol and use it as a semantic component of its antonym:
truth: [- false] falsity. [- true] and this is certainly an arid and circular breakdown.
83
WORDS AND MEANINGS
Finally, components describing relative dimensions and degrees, which would be needed to distinguish between many semantically proximate lexical items, are difficult to accommodate satisfactorily into the model. To differentiate between, say, stone and boulder— both of which are [+ solid] and [+ mineral]—it would have to be specified that boulder is [+ large] and stone [– large]. Yet if pebble were to be brought into the reckoning, it would now be stone that would assume the [+ large] component in order to distinguish it from the new term. Measurements of dimension and degree are relative and are calculated along varying parameters, which simply makes them difficult to assimilate into a semantic model of this sort. Although componential analysis remains useful as a general explanatory tool, the reasons given above suggest that it is nowhere near a fail-safe method for predicting collocations. Collocation is a probabilistic phenomenon: it is a question of whether this or that item is more likely to occur than another. It is no surprise, then, that a rigid semantic model will have difficulty formalising a principle that is more a tendency in language-use than an unflinching structural rule of language. However, it’s not that we don’t have very strong expectations about what word follows what in language. We do. And as the next section demonstrates, there are a number of literarystylistic techniques available for exploring our intuitions and expectations about the company a word keeps.
3.4 Techniques for stylistic analysis An important feature of modern stylistics is the way in which it employs pedagogical techniques commonly used in the fields of applied linguistics and English language teaching. It has imported a variety of such techniques in recent years, including two popular methods for examining collocation and lexical semantics. These are cloze [sic] procedure and multiple choice text. The two practices involve the blanking-out of items in a syntagmatic chain and asking for students’ (or informants’) predictions about which sorts of entries could fill that particular grammatical context. Although there are localised methodological differences between them, cloze procedure and multiple choice text are both designed to explore the rich interface between what actually occurs in a text and the expectations that people have about what should occur in a text. 84
WORDS AND MEANINGS
3.4.1 Cloze procedure and stylistic analysis Cloze procedure (or cloze test) is the deletion of particular words in a sequence with a view to eliciting predictions about appropriate paradigmatic entries. This relatively straightforward technique has in fact already been used in this chapter. In the previous section, examples 22 to 26 were all presented initially in a cloze format, which helped to focus and develop subsequent discussion of collocational patterns. There are three principal ways of using cloze test. The first, and perhaps most obvious, way is as a language teaching tool. Cloze has proved especially popular in the EFL context where it serves as a useful vocabulary testing device. Because it attempts to impart knowledge about which lexical items are appropriate to which grammatical context, this, of all its applications, is the one that comes closest to getting students to fill in the ‘right’ term for the structural slot. To this extent, cloze test is able to shed a great deal of light on the concept of the lexical set. A lexical set is a bundle of semantically compatible items which are closely linked to a specific topic or register. Although synonyms (and near synonyms) are obvious candidates for inclusion in lexical sets, the concept extends much more widely to encompass clusters of key words which correlate generally with a particular field of discourse. The words Byzantium, flora, floppy disk and popcorn do not, for example, form a lexical set. The word goalkeeper, header, off-side, ball and defender clearly do, as they all relate to the discourse of soccer. A cloze test sequence like 31 The defender headed the———. can therefore consolidate awareness of this register. Here, the blanked-out term belongs to a lexical set already clearly signalled by a syntagmatic chain comprising other members of the same set. Cloze procedure can also be used as a tool for linguistic research. By asking informants to predict lexical collocations and then collating their patterns of response, important information can be gleaned about the grammar and vocabulary of a language. Crucially, this information is not the sole product of the analyst’s own (often unreliable) intuitions, but is instead derived from an attested corpus of ‘real’ language. In this type of research, no prior assumptions are made about ‘right’ and ‘wrong’ answers to the cloze test; rather, it is 85
WORDS AND MEANINGS
the answers themselves which form the quantitative basis for explaining which items collocate with which. Third, and most importantly for our present purposes, there is the stylistic application of cloze procedure. In this context, cloze is a productive way of exploring the territory between what we expect to see in a text and what a writer does in a text. It also creates a new focus for interpretation, because it commits students to providing a partial analysis of the text in advance of actually seeing the complete version. In stylistic applications, cloze is most certainly not used as a test of language skills. On the contrary, it simply asks informants for their intuitions about a text—intuitions which can’t realistically be considered ‘right’ or ‘wrong’. One final comment on the applications of cloze test is necessary before we embark on the stylistic analysis that will occupy the remainder of this subsection. In common with other stylistic techniques, cloze is not entirely the exclusive preserve of the language class or of academic linguistics. It enjoys a wider nonacademic circulation than might at first appear. The popular satirical programme called Have I Got News For You enjoys a regular primetime slot on BBC television. The programme is organised loosely along the lines of a quiz, one stage of which involves the completion of what amounts to a cloze test. Topical newspaper headlines, with an item blanked out, are presented to a panel of participants. A flurry of guesses follows, with answers ranging from the surreal to the irreverent. For the most part, anything but the correct answer is offered. Here is a selection of examples from the programme whose topicality has not been entirely lost. A selection of ‘guesses’ is provided below each example with the actual deletion, which is eventually provided by the quiz’s host, underlined: 32 QUEEN MOTHER COMBINES ROYAL DUTIES WITH LOVE OF———. (i) CHAMPAGNE AND RAFFLES FAGS (ii) SUPERNINTENDO (iii) RENAISSANCE SPANISH CHURCH MUSIC (iv) HORSES AND ENTERTAINING 33 YARD’S NEW CHIEF URGES POLICE TO ABANDON ————. (i) HOPE 86
WORDS AND MEANINGS
(ii) RULE BOOK (iii) SOCIAL ENGINEERING 34 NUMBER OF SEXUAL PARTNERS LINKED TO———-. (i) EACH OTHER (ii) MOBILITY AND CASH 35 BBC LOSES 60 MILLION IN———-. (i) ALL NIGHT POKER GAME (ii) CASH BUNGLE 36 PRIME MINISTER PROMISES FRESH————-. (i) AIR (ii) HALIBUT (iii) ULSTER INITIATIVE Although points are notionally awarded for guessing the right collocate, the real contest is to see who can offer the wittiest, funniest or most bizarre suggestion. Now to our stylistic application of cloze procedure. I intend to concentrate on a short poem entitled ‘Dolour’ which was first published in 1948 by the American poet Theodore Roethke.13 Here is the opening of the poem, and, in keeping with the principle of cloze test, a lexeme has been deleted from the first line:
¶
Dolour I have known the inexorable sadness of _____, The title ‘Dolour’ establishes a semantic framework which is reinforced by its synonym, ‘sadness’, in the opening line. One would assume therefore that this structural frame will exercise strong constraints on the sorts of words that could complete the line. And judging by the consistency of response in seminars where I have elicited suggestions, this assumption does indeed appear to be valid. Suggested entries for the blanked-out item routinely converge on the following types of terms: death; life; love; parting; dying; leaving. These items all refer to aspects of human existence which may be associated (in obviously varying degrees) with feelings of sadness or unhappiness. All the more effective, then, is the cognitive jolt delivered when the actual entry in the first line is revealed: 87
WORDS AND MEANINGS
I have known the inexorable sadness of pencils, The mismatch between ‘sadness’ and ‘pencils’ constitutes a collocational clash on any semantic parameter. Ascribed to inanimate objects (‘pencils’) is a quality which is normally attributed to human emotional experience. A componential analysis of the two lexemes ‘sadness’ and ‘pencils’ would uncover a bundle of semantic features that could explain the clash systematically. In fact, any one of the following set of contrasts would be enough to activate it: [+/– animate]; [+/– human]; [+/– concrete]; [+/– abstract] or [+/– sensate]. The clash is accentuated further at the level of vocabulary and register. The tone established by the Latinate lexis of ‘Dolour’ and ‘inexorable’ is portentous and grandiose, but when the lexeme ‘pencils’ suddenly intrudes, it switches abruptly to the prosaic and commonplace. After having reinstated the missing term from the first, here is the second line with its deletion: Neat in their boxes, _____ of pad and paperweight, Unlike those offered for line 1, the predictions offered here display much less consensus. Some participants admit to experiencing difficulty in finding selections that are compatible with both ‘pad’ and ‘paperweight’. Consequently, the sorts of entries suggested are diverse and often lack a common semantic link: desk (sic); piles; box; workstation; cousins. However, the responses of other participants to the task are rather intriguing. Many seem to be influenced by the jolt delivered by the earlier line, and so predict the following types of terms: pain; loneliness; misery; depression; boredom. It is as if the cloze test has been ‘infected’ by the first line and the clash has now become part of a new cognitive model that informs our reading of subsequent lines. This reading is further impelled by the parallel grammatical structure that links the two lines. The first line sets up a syntagmatic structure of the order:
I have known the inexorable x of y In this syntagm, x represents a term from the paradigm of words describing human emotion. The y element is a member of the paradigm of words referring to, well, office equipment. And as the second line exhibits the same ‘x of y’ pattern, then there is every 88
WORDS AND MEANINGS
good reason that the clash be sustained, a hunch which is indeed borne out when the original item is reinstated: Neat in their boxes, dolour of pad and paperweight, By the third line, this ‘framework of incompatibility’ is firmly rooted in the cloze test, to the extent that now some participants actually claim to find the task straightforward: All the ______ of manila folders and mucilage, The deleted item, predicted with uncanny precision in students’ responses, is ‘misery’. Here, it is not just the general lexical set of terms denoting unhappiness that is picked out, but this very lexeme which is often identified. It is easy to see why. For a start, the ‘x of y’ syntagm is sustained, suggesting that the clash between paradigmatic entries be sustained also. Another significant factor is the sound structure of the line, which provides further cohesion along the syntagmatic axis. The first consonants of the words ‘manila’ and ‘mucilage’ form an alliterative /m/ sequence, which appears to influence selections from the paradigmatic axis. Even though the level of awareness at which this sound pattern is perceived may vary, the preferred entry by many is still ‘misery’. It is noticeable also that, in entries suggested for the previous line, the lexeme pain was suggested by two participants on the grounds that it ‘just sounded better’. Although no further explanation was forthcoming in this session, the prediction may have been influenced by the consonantal /p/ pattern evident in ‘pad and paperweight’. In fact, the realised term turns out to be a repeat of the title, so the /p/ pattern is not sustained. Nevertheless, the ‘wrong’ entries show just how sophisticated is our awareness of different layers of linguistic patterning in a text. In this case, the alliterative lexeme offered by participants proved perhaps to be a little more ‘poetic’ than the lexeme actually used by the poet! Here are the next three lines of text, presented in cloze format: _____ in immaculate public places. Lonely reception room, lavatory, switchboard The unalterable ______ of basin and pitcher, The framework of incompatibility generated by the clashes in the 89
WORDS AND MEANINGS
previous three lines now governs firmly the processing of the text. Predictions seldom deviate from a narrow set of synonyms or proximate synonyms describing negative or unpleasant human experience. And again they prove to be accurate: the first term is ‘desolation’ and the second, which exhibits another phonological correspondence, ‘pathos’. The primary semantic strategy employed in the poem has now been identified and, hopefully, the basic principle of the exercise will have been grasped. Suffice it to say, the clash engendered by the collocation of semantically incompatible lexemes is carried through the remainder of poem. As there is no real need to demonstrate the cloze test further, the rest of the poem is produced in pristine form: Ritual of multigraph, paper-clip, comma, Endless duplication of lives and objects. And I have seen dust from the walls of institutions, Finer than flour, alive, more dangerous than silica, Sift, almost invisible, through long afternoons of tedium, Dropping a fine film on nails and delicate eyebrows, Glazing the pale hair, the duplicate grey standard faces. Before we move on to consider the general significance of the exercise, it is worth making a couple of observations about this last section of the poem. First of all, the ‘x of y’ formula becomes progressively more condensed over the last few lines of the poem until ultimately the ‘incompatibles’ are distilled into the single phrase ‘the duplicate grey standard faces’. Second, and as this final phrase suggests, there is a gradual shift from abstract facets of human experience to more palpable aspects of humanity. This shift is encoded in terms like ‘nails’, ‘eyebrows’, ‘hair’ and ‘faces’. However, it is only a fragmentary humanity which is depicted: words for parts of the body eclipse holistic terms such as ‘people’, ‘women’ or ‘men’. And there is, moreover, a frailty and powerlessness in the physiological minutiae conveyed by words like ‘nails’, ‘hair’ and ‘eyebrows’. This is not the first time that we have witnessed a deviant or striking linguistic structure becoming established as a kind of ‘norm’ in a text. In the previous chapter, it was shown how repeated graphological and morphological breakage in an e e cummings poem created its own regular design for that poem. In this section, it is a sustained pattern of collocational clashes which establishes the internal stylistic fabric of 90
WORDS AND MEANINGS
the text. Roethke incorporates two semantically disparate lexical sets into a single syntagmatic combination. One set comprises items typically found in an office environment, such as pad, paper-weight, mucilage and paper-clip, whereas the other consists of synonyms (or near synonyms) for the experience of unhappiness, and includes misery, sadness, desolation and tedium. However, having identified the technique employed, the next step is to assess how it influences and shapes the reading and interpretation of the text. Collocational clashes allow semantic disparates to be brought into grammatical proximity with one another, so the projection of the human on to the inanimate can be crystallised into a single compact image in the poem. The stylistic merger of the living and the lifeless serves presumably as an indictment of the dehumanisation engendered by institutional bureaucracy. However, the merger also functions as a conceptual shortcut. It bypasses the intermediary connections which would be needed to link two such disparate phenomena: namely, that working in institutions is stultifying and depressing, and that this condition is made manifest in the very paraphernalia that characterises the workplace. Instead, the clash permits humanity and insensate objects to be distilled together into an image which is both compressed and minimal. The question of precisely how readers tease out coherence from such semantically anomalous language will be addressed later in section 3.5. This application of cloze test lends itself well to follow-up activities. Because it unveils the basic stylistic formulae which underpin a text, the exercise can be readily transferred to creative and compositional work. As was pointed out in subsection 2.4.2, stylistic analysis often yields a replicable ‘blueprint’ which allows for the development of experimental material within those same linguistic parameters. Reduced to its most basic ingredients, the formula of ‘Dolour’ could be characterised as ‘structural frame+lexical incompatibility’. This formula can be replicated in follow-up exercises which are not only entertaining in themselves but which can also offer valuable extra insights into synonymy, lexical sets and the principle of collocation. Here, for example, is a short ‘recipe’ for collocational experimentation centred on the poem we have just studied. It is built on the structural frame of ‘Dolour’, retaining the ‘x of y’ syntagm, but uses a different set of lexical entries from those identified during the cloze test. 91
WORDS AND MEANINGS
1 Alter the lexical set of terms referring to human experience in the following way. Replace the words describing negative emotions with their positive antonyms. Or to put it another way, develop a lexical set containing synonyms of happiness. Here are some suggestions: joy, pleasure, bliss, ecstasy, delight, elation, rapture, gusto, relish, enjoyment, satisfaction, fun 2 Draw up a set of terms from a technical register. This set may include the vocabulary associated with any hobby or pursuit which demands some degree of specialist terminology. It may therefore encompass anything from motor car maintenance to tropical fish keeping. Here, for example, is a selection of terms associated with the register of cookery: jug, oven, microwave, kettle, pot, pan, knife, fork, spoon, ladle, skillet, fish-slice, toaster, blender, worktop 3 Select an item from each lexical set and, using the structural frame of ‘Dolour’, combine them grammatically. If possible, create some sound correspondence between the terms. This can be achieved simply by matching up alliterative consonants thus: ‘bliss of blender’; ‘joy of jugs’; ‘fun of forks’. To close this subsection, here is a brief sample of the sort of text which results from the practical activity outlined above:
Pleasure I have known the inexorable fun of forks; Neat in their drawers, pleasure of pot and pan, All the satisfaction of saucepan and spoon, Elation on immaculate kitchen worktops.
92
WORDS AND MEANINGS
3.4.2 Multiple choice text and stylistic analysis In the contexts of language learning and teaching about language, multiple choice text involves the drawing up of a restricted set of terms for each blanked out slot in a syntagm. As it has much in common with cloze procedure, all that is needed here is a brief sketch of its important features, together with an explanation of when it is likely to offer a productive alternative to cloze. Multiple choice text is of particular benefit when there is an abundance of available paradigmatic entries for a given slot. It closes down the range of interpretative possibilities, allowing attention to be centred on the specific aspects of the lexicon which are the focus of study. Like cloze test, multiple choice text can be used to investigate virtually all of the topics in lexical semantics introduced in the course of this chapter. Not only is it a handy tool for investigating synonymy, collocation and lexical specificity, it is also a useful apparatus for exploring the relationship between lexis and register. In this latter application, groups of semantically related terms can be drawn up for each structural slot along the chain. Example 37 is a syntagm from which an item has been deleted and below it are listed four choices. Notice how fixed expressions such as idioms and euphemisms may be included in the list of lexical entries. 37 We regret to inform you that your great aunt has just ____. Choices: died passed away kicked the bucket expired While all four entries denote the same thing, there are clearly significant connotational and associative differences between them. Consequently, some terms are more appropriate than others to this type of communicative context. Whereas the first two choices on the list would be likely candidates for inclusion, the second two most certainly would not as they are manifestly at odds with the register evoked by the syntagm. Knowing when to use one term and not the other is an indispensable communicative skill which, while often taken for granted by native speakers, is the bane of many non-native 93
WORDS AND MEANINGS
¶
speakers. This type of exercise is one way of assessing the appropriateness of semantically compatible items to a particular grammatical context and of differentiating between synonyms according to the constraints of register. Multiple choice text (as well as cloze test) can also be used for investigating a particular feature of vocabulary which has not been covered thus far in this chapter. It can help to identify core vocabulary. Core words in a language are those which have especially wide distribution and usage. They have an extensive collocational range, being able to combine syntagmatically with many other words. Take for example the following lexical set which comprises four closely related lexemes: fat, obese, overweight, plump. Although broadly synonymous, in no way are these words distributed equally in the lexicon. However, a common error in language-learning is the assumption that they are distributed equally. Because it places them against different syntagms, multiple choice text can highlight the varying collocational possibilities of these synonyms, thereby testing for their respective degrees of ‘coreness’ in the language. In the four examples that follow, a symbol beside a term indicates collocational acceptability; an × collocational unacceptability: 38 ————person fat Ö plump Ö obese Ö overweight Ö 39 ————chicken fat Ö plump Ö obese × overweight × 40—————salary fat Ö plump × (?) obese × overweight × 41—————chance fat Ö
94
WORDS AND MEANINGS
plump × obese × overweight × With its extensive collocational range, fat is clearly the most ‘core’ of the four terms, whereas obese and overweight, with their more restricted range, are the least ‘core’ of the set. There is not the space to undertake a detailed demonstration of the stylistic potential of multiple choice text, so to end this subsection a brief illustration only will be offered. Like cloze test, this semantic framework foregrounds the use of lexis as a stylistic strategy. A poet whose use of language generally, and whose use of vocabulary in particular, has been an important focus of attention in modern stylistics is W.H.Auden. Here is the opening stanza of his well-known elegy ‘In Memory of W.B.Yeats’ (1939):14 He disappeared in the dead of winter: The brooks were frozen, the airports almost deserted. And snow disfigured the public statues; The mercury sank in the mouth of the dying day. What instruments we have agree The day of his death was a dark cold day. Even though title of the poem signals a sombre, funereal theme that will shape and influence subsequent reading of the text, the manner by which this elegiac content is embodied in lexical choices is striking. A lexical paradigm built around the second structural slot in the first line, which might contain the core word ‘died’ and its euphemistic synonym ‘pass away’, would test the expectations established by the discourse context. The realised term, ‘disappeared’, is startling, although, significantly, explicit semantic reference to death is transposed to a later position along the syntagmatic axis (‘dead of winter’). Similarly, whereas ‘brooks’ in the first half of line 2 arguably sets up a pastoral framework of interpretation, the appearance of ‘airports’ in the second half of the line is incongruous and unexpectedly prosaic. Noticeably, where commonplace or formulaic expressions might be anticipated in the discourse context, subtle twists of perception are created when an unexpected term is realised. For instance, in line 5 it is not, as one might expect, ‘sources’ or ‘authorities’ who ‘agree’, but ‘instruments’. Auden’s lexical strategies, here as in much of his work, 95
WORDS AND MEANINGS
¶
do not consist in the deployment of the sort of wholesale semantic violations that were seen earlier in this chapter; rather, they operate more as understated deflections from probabilistic collocational patterns. Carter (1982b: 41–2) talks of Auden’s capacity to develop a kind of lexical ‘code’ in his poetry: a distinctive poetic style comprising words which are conventional in themselves but which are foregrounded through deviations from internal contextual norms. This vocabulary, Carter continues, draws attention to itself as lexis and poetic convention, so much so that the ‘lexis itself becomes an important part of the subject matter’.
3.5 Summary For the purposes of studying lexical semantics, the practical activities outlined in the previous section are quite flexible and can be tailored to suit specific teaching and learning goals. For example, we conducted our cloze procedure analysis of ‘Dolour’ on a ‘line by line’ basis. This seemed an appropriate method for assessing how a sequence of discourse informs and affects our reading of subsequent sequences. However, the cloze test format can be much more expansive: whole stanzas or even whole texts, with their blanked-out slots, can be examined in one go. This larger-scale format can also be used in the application of multiple choice text to highlight how strands of meaning operate across larger chunks of text. Whatever the precise style of their implementation, both techniques help to scotch the myth that there is such a thing as a ‘literary lexis’ (see section 1.3, above), emphasising instead that the real stylistic issue is to do with the company a word keeps. It is often the semantic space that lies between words, rather than the properties that inhere in the words themselves, which forms the core of verbal imagery in poetry. The analysis carried out in section 3.4 also raises important issues to do with grammatical theory. It has particularly significant repercussions for those theories of grammar which have at their core the distinction between semantically acceptable and semantically anomalous constructions in a language. The Transformational-Generative model, for instance, is designed to explain, amongst other things, how speakers know the difference between a permissible ‘string’ of words and an impermissible one. Yet in the course of our study of ‘Dolour’, it quickly became clear 96
WORDS AND MEANINGS
that not only were participants in seminars good at assimilating and understanding anomalous strings, but that they were even able to predict anomalous strings. How is it then that we can process such apparently opaque language? One solution to the problem has been proposed by the linguist J.P.Thorne in a well-known article entitled ‘Generative Grammar and Stylistic Analysis’ (1981). Thorne’s article is worth reading for several reasons. He carries out a stylistic analysis of ‘Dolour’, although the TransformationalGenerative model he uses is theoretically very much at odds with my own approach. This offers the opportunity for some productive debate between our two positions. Thorne’s main argument is that ‘deviant’ structures in poetry, such as the collocational clashes evident in ‘Dolour’, can be explained by developing a ‘grammar’ for the text. This new ‘grammar’ will accommodate the deviant structures in the poem, so the reader’s task of processing semantically anomalous combinations becomes akin to learning a new language. Thorne presents his ‘grammar-of-the-poem’ theory thus: Behind the idea of constructing what is in effect a grammar for the poem lies the idea that what the poet has done is to create a new language (or dialect) and that the task that faces the reader is in some ways like that of learning a new language (or dialect)…For these poets the point of creating a new language, therefore, seems to be that it enables them to say not only things that can be said in Standard English, but in a different way, but also things that cannot be said in Standard English at all—though they can be understood only by someone who understands Standard English. (1981:50–1) The ‘grammar-of-the-poem’ theory is certainly a challenging one, but in my opinion it is flawed in several respects. For one thing, to say that collocational clashes are a ‘new’ language is to nullify their stylistic effect totally. Deviant patterns function at an intralinguistic level: their impact results from their interplay with other elements inside a particular language. If indeed they constituted a ‘new’ language, then describing them would be much the same as accounting for the formal differences between, say, French and English. Other parts of Thorne’s analogy with language learning are problematic. Learners new to a language obviously try to acquire strings of words that are ‘normal’ in that particular language. It is 97
WORDS AND MEANINGS
only the experienced users of that language who can be expected to make sense out of anomalous or irregular patterns. Furthermore, if deviant combinations create a new language in which the striking becomes the norm, then presumably this new language is itself susceptible to the sort of defamiliarising operation that generated it in the first place. Would the result of this operation be yet another new language or would it simply constitute a return to the ‘normal’ code from which we started? Clearly a lot more investigation needs to be done before we have an adequate model of how readers make sense of the complex and anomalous language that characterises much poetry. And such a model would be just as much a theory of reading literature as it would a theory of grammar.
Suggestions for further reading Lexical semantics Useful introductions to lexical semantics (which also cover the debates between different approaches to the topic) are: Carter (1987); Hurford and Heasley (1983); Palmer (1982); Crystal (1987:102–7). These books are generally accessible and offer fuller accounts than is possible here. For extra reading on lexical semantics see: Fromkin and Rodman (1988:205–50); Jackson (1988); Cruse (1986). More advanced surveys of broader issues in semantics can be found in the following books: Lyons (1977); Leech (1981); Kempson (1977).
Lexical semantics and stylistic analysis There are a number of useful books and articles which use lexical semantics for stylistic analysis. Leech (1969) is the first standard textbook to be written on the stylistics of poetry. A more recent contribution to the field is Verdonk (1993), which is a collection of essays aimed at the non-specialist. The general stylistics of naming in rock and pop is covered in Wales (1992). Nash (1985) and Chiaro (1992) are two books on the language of humour which contain entertaining sections on puns, clashes and plays on word-meaning. Ronald Carter has written a number of studies of W.H.Auden’s poetry 98
WORDS AND MEANINGS
from a stylistic perspective. For his discussion of Auden’s poem ‘Mundus et Infans’ using cloze procedure, see Carter (1981 and 1987:193– 207). An analysis of the poem ‘Oxford’, using similar semantic techniques can be found in Carter (1982b). For a general discussion of cloze procedure and multiple choice text in a stylistic context see Van Peer (1989). Van Peer has also written a more specialist monograph on stylistics and psychology in which many of the techniques suggested in this chapter are used (Van Peer 1986).
¶
99
Exploring narrative style: patterns of cohesion in a short story
•
4.1
Introduction
•
4.2
Practical activity: reconstructing a short story 103
•
4.3
Results and discussion
105
•
4.3.1 General patterns of response
106
•
4.3.2 An idealised narrative
110
•
4.3.3 The original version
112
•
4.4
115
•
4.4.1 Cohesion
115
•
4.4.2 Natural narrative
116
•
4.5
Practical activity: creative writing
120
•
4.6
Summary
126
Extending the analysis
• Suggestions for further reading
102
Chapter 4
Chapter 4
126
101
4.1 Introduction In keeping with the general orientation of the book, this chapter illustrates how literature can be used as a point of entry into the study of English language. Here the focus of attention will be on prose fiction and the features of language that will be introduced and explored are cohesion and narrative. The structure of this chapter, however, differs from that of the previous two in one key respect. Whereas the practice up to now has been to discuss topics in language before the development of practical exercises, this chapter by contrast takes a direct route to the text and expands upon topics in language after the main practical activities have been undertaken. The chapter is structured as follows. In the next section, readers will be invited to take part in a practical workshop activity which is based upon a short story by American writer Ernest Hemingway. A jumbled version of the story is presented from which readers will be asked to assemble a coherent and well-formed narrative. In section 4.3, the responses of groups of students who have carried out the same activity will be reviewed and their (re)constructions will eventually be compared with the original Hemingway version. In a more technical way, section 4.3 will also be concerned with exploring the patterns of cohesion in the narratives developed by the students. From their collected versions of the story, an idealised narrative will be constructed which will then form the basis of comparison with the original Hemingway version. In the fourth section of the chapter, the analysis of cohesion will be supplemented with a discussion of natural narrative. A number of extensions to the main study are developed in section 4.5, including some suggestions as to how the kind of activity outlined here might be integrated with creative writing. Finally, some concluding remarks are offered on the ways in which additional linguistic frameworks may be brought into play in the study of narrative. 102
EXPLORING NARRATIVE STYLE
4.2 Practical activity: reconstructing a short story The Hemingway text selected for reconstruction is one of the many short vignettes peppered throughout the In Our Time collection. These vignettes are untitled, are seldom over one hundred and fifty words long and are often simply wedged between the more substantial short stories in the collection. They are distinct from the longer stories not only in terms of content, but in graphology also: most are italicised, which separates them visually from the other material around them. Given that the object of the exercise is to foreground the linguistic devices which bind stories together, I began by dismantling the text entirely. Both the size and overall structure of the text helped enormously here, in that the story comprises only eleven sentences totalling 130 words. I (literally) cut up a photocopy of the story into its constituent sentences and shuffled the eleven pieces of paper. The pieces of paper (containing a single sentence each) were then drawn in random sequence and every sentence was assigned a letter from the consecutive sequence a, b, c,…up to k. The sequence of sentences is now completely haphazard; even if any particular combination happened to match the original it will have been entirely due to chance. The sequence which emerged from this dismantling process is reproduced within the ruled text. Before that, there is a set of instructions on how to ‘re-assemble’ the text which readers might care to follow. In fact, if time and facilities permit, it would be a good idea to photocopy the next page, cut it up and then experiment by laying out and rearranging the eleven sentences on a worktop. Better still if you can confer with colleagues or peers about which sequence is best. For the moment, the original Hemingway version will be withheld until we have had time to explore fully the sorts of reconstructions that are produced under such circumstances. Here now are your instructions: 1 You will find a complete (untitled) short story by Ernest Hemingway. The story comprises in total just eleven sentences. However, these sentences have been jumbled so that now they do not occur in the sequence of the original story. Each sentence has been assigned a letter.
103
EXPLORING NARRATIVE STYLE
(a) All the shutters of the hospital were nailed shut.
(b) When they fired the first volley he was sitting down in the water with his head on his knees.
(c) There were pools of water in the courtyard.
(d) They tried to hold him up against the wall but he sat down in a puddle of water.
(e) One of the ministers was sick with typhoid.
(f)
Two soldiers carried him downstairs and out into the rain.
(g) There were wet dead leaves on the paving of the courtyard.
(h) Finally the officer told the soldiers it was no good trying to make him stand up.
(i)
They shot the six cabinet ministers at half past six in the morning against the wall of a hospital.
(j)
It rained hard.
(k) The other five stood very quietly against the wall.
104
EXPLORING NARRATIVE STYLE
2 Try to reconstruct the story. Try to build up what you consider a normal narrative sequence. This should not become an attempt to reassemble the original Hemingway text but rather an attempt to produce a sequence of the sentences that results in a coherent narrative for you. 3 Make a note of the sequence you have selected (e.g. a b c d e f etc.) and try to think about why you chose this particular sequence. It may help to isolate any features of language which signal that a particular sentence should precede rather than follow another. Underline any such features.
¶
4.3 Results and discussion Readers by now should have developed their own versions of the narrative and have formed some opinion as to why they prefer the sequence they do. Unfortunately, the communicative dynamics of print media are such that I am unable to see what you have done or to discuss with you your particular versions! Perhaps the next best thing, then, is to look at what others have done using the same materials. What follow next are the results of an undergraduate seminar based on the same activity. This was a largish seminar group, comprising twenty-eight students, so rather than being asked to produce individual responses, participants were asked to work in six groups comprising four or five members apiece. The results obtained from the seminar activity are summarised in Table 4.1. Readers can now check across the table to see if their own versions generally square with (or are even identical to) any of the sequences proposed by the six groups. One of the advantages of a quantitative summary of this sort is that it is possible to view at a glance a collected set of responses. These responses are based specifically upon native speakers’ intuitions about what constitutes a well-formed, coherent narrative. We can gain insights into not only the sorts of patterns which recur but also the variety of responses which is reflected across the six groups. The summary also creates a kind of ‘Consensus’ response against which readers can check their own versions. And, of course, all of this will provide an illuminating point of contrast with the original Hemingway version when it
¶
105
EXPLORING NARRATIVE STYLE
TABLE 4.1 Results of workshop activity sentence
appears later in the section. For the moment, however, it will be useful to reflect upon and discuss the general patterns that have emerged from the students’ responses.
4.3.1 General patterns of response It is noticeable that only two of the available eleven sentences were selected as viable openings to the story. Of these two, sentence (i) was favoured: They shot the six cabinet ministers at half past six in the morning against the wall of a hospital. The justification offered by the groups who chose (i) as the crucial opening sentence were interesting and merit some consideration. The underlying motivation behind this selection was that of all the sentences this was the one which provided the neatest summary of the entire story. No other sentence encapsulated the point of the story in as ‘snappy’ (sic) a way. The precise linguistic justification offered in support of this centred primarily on two features of the sentence. First, the noun phrase ‘the six cabinet ministers’ is the most elaborated of all the references to the ministers, and would naturally be expected to precede subsequent, more reduced forms. For instance, the appearance of the partitive expression ‘one of the ministers’ later in the narrative presupposes an earlier reference to the group of ministers as a whole. Second, and more significantly, this is one of only two sentences in the story which contain an 106
EXPLORING NARRATIVE STYLE
indefinite article: ‘a hospital’. This is a linguistic feature which is often used to introduce a new entity into the text, with subsequent references adopting the definite article ‘the’. (See, for example, the reference to ‘the hospital’ in sentence (a).) In spite of this justification, however, some groups still expressed misgivings about the selection of (i) as an opening. One feature which perturbed a number of participants was the use of ‘They’ without an antecedent. Now, in terms of the cohesion model mentioned in section 4.1, the pronoun ‘they’ is normally used anaphorically (Halliday and Hasan 1976:14). That is to say, it usually refers ‘backwards’, anaphora being the type of cohesive relation which points to some earlier, fuller reference in the text. By contrast, in the present example, the pronoun is used cataphorically, it points forward to a reference or references in a subsequent portion of text. In this respect, the ‘They’ can be said to form a cataphoric link with the references to ‘the soldiers’ and ‘the officer’ later in the narrative. On balance, however, it was generally agreed that this feature did not warrant the de-selection of (i) as the best possible opening to the story. Indeed, a number of students suggested that this type of cataphoric reference was a common enough characteristic of the openings of literary texts. The two groups who chose sentence (j) for their opening took a somewhat different approach. They argued that a short ‘scenesetting’ sentence like ‘It rained hard’ was a more appropriate beginning to the story. Moreover, they had no problems in explaining away any unusual cataphoric references as the ‘It’ in the sentence is simply a dummy subject, an item which has no reference value and is purely a formal requirement of the grammar of English. One of the more striking aspects of the table of results is the way in which it highlights certain configurations of sentences which are almost invariable across all groups. Particularly noteworthy is the following sequence of sentences proposed by five of the six groups: (e) One of the ministers was sick with typhoid. (f) Two soldiers carried him downstairs and out into the rain. The justification for this configuration rested on both the formal aspects of the language system and on the background assumptions of the participants themselves. First of all, the pronoun ‘him’ in (f) provides a clear anaphoric link to the antecedent phrase ‘one of the ministers’. 107
EXPLORING NARRATIVE STYLE
Second, further coherence is created through the suggestion of causality: the minister is ill so he has to be carried. This ‘commonsense’ interpretation prompted a number of illuminating comments on the general lack of causality in Hemingway’s style. Formal links between propositions, such as if, so and therefore, tend to be uncommon and, consequently, ‘cause and effect’ and ‘logical’ relationships often have to be constructed by readers. The agreement reached over the sequence of sentences (h) and (b) also warrants some comment. Four groups proposed the following pattern, remarkably, at identical points in the narrative: (h) Finally the officer told the soldiers it was no good trying to make him stand up. (b) When they fired the first volley he was sitting down in the water with his head on his knees. The reasons put forward in support of this pattern were interesting. In addition to the cohesion created through the use of the anaphoric pronouns ‘they’, ‘he’ and ‘his’, all groups pointed out that extra cohesion was created by the juxtaposition of the opposites ‘stand up’ and ‘sitting down’. This cohesion through antonyms (see section 3.2) is identified by Halliday and Hasan as one of the main types of lexical cohesion in English. Another area of cross-group similarity is the sequencing of sentences (i), (e) and (k). Although the exact position of these sentences in the narrative varied, and the sentences were interrupted by other elements, the precise sequence was, with only one exception, invariable. Only Group 5 did not propose the following pattern: (i) (e) (k)
They shot the six cabinet ministers at half past six in the morning against the wall of a hospital. (…) One of the ministers was sick with typhoid (…) The other five stood very quietly against the wall (…)
The relative rigidity of this pattern can be explained by reference to the cohesive chains which develop through it. As was pointed out above, the partitive expression ‘One of the ministers’ is tied to the initial reference in (i). The cohesion between (e) and (k) is even more rigid, with the expression ‘the other five’ undergoing ellipsis. This cohesive device is a type of ‘substitution by zero’, where one or 108
EXPLORING NARRATIVE STYLE
more items may be deleted but may still be inferred from the linguistic context. For instance, in the second clause of the following sentence both the verb ‘scored’ and the noun ‘goals’ have been ellipted, yet are presupposed through reference to the first clause: Everton scored three goals and Liverpool two. Returning to the text, it is clear that the ‘other five’ is only meaningful through reference to sentences (e) and (i), thus developing strong anaphoric links between this particular group of sentences. The cohesive chain which emerges might be formulated as in Figure 4.1. The chain is further strengthened by the presence of other cohesive links such as repetition; in this instance, the phrase ‘the wall’ occurs in both (i) and (k).
FIGURE 4.1 The one group who placed (i) after (e) and (k) (see Group 4 above) had specific reasons for doing so, although they felt unhappy about the overall structure of their narrative. Their decision for placing (i) towards the end of the narrative will be dealt with later, in subsection 4.4.2. 109
EXPLORING NARRATIVE STYLE
It is interesting also that only one group chose to conclude the narrative with a sentence which contains a clear ‘signpost’ in the form of the initial adverb ‘finally’. Group 6, then, took ‘finally’ to mean precisely that, yet the other five groups were dissatisfied with the sense of incompleteness which final (h) would generate.
4.3.2 An idealised narrative It was anticipated at the outset that the results of the workshop would produce an ‘idealised’ narrative based on the majority preferences displayed in each of the eleven sentence columns. This would produce a sample narrative derived from the collected responses which could then be contrasted with the original version. Unfortunately, given the set of results obtained genuine abstraction of this sort is difficult: some preferences are evenly distributed over a number of columns whilst other columns yield no statistically significant set of preferences. Nevertheless, it is still possible to select a sample narrative representing the most ‘typical’ narrative produced by all the groups. In this respect, the narrative of group 2 is appropriate as their choices are most widely shared by other groups. The complete narrative proposed by this group is as follows, with letters retained for ease of reference, but with the gaps between the sentences closed up in order to render the sequence ‘storylike’.
(i) They shot the six cabinet ministers at half past six in the morning against the wall of a hospital. (j) It rained hard. (e) One of the ministers was sick with typhoid. (f) Two soldiers carried him downstairs and out into the rain. (d) They tried to hold him up against the wall but he sat down in a puddle of water. (g) There were wet dead leaves on the paving of the courtyard. (k) The other five stood very quietly against the wall. (h) Finally the officer told the soldiers it was no good trying to make him stand up. (b) When they fired the first volley he was sitting down in the water with his head on his knees. (c) There were pools of water in the courtyard. (a) All the shutters of the hospital were nailed shut.
110
EXPLORING NARRATIVE STYLE
A significant feature of this narrative as a whole is the division between what might be loosely termed descriptive and actional frames. The group who produced this version attempted to create a balance between, on the one hand, the central action of the shooting, and, on the other, the more ‘static’ descriptions of the rain, the hospital and the courtyard. To this end, the descriptive sentences such as (j), (g) and (c) were quite consciously woven around the central narrative events realised chiefly by sentences (f), (d), (k), (h) and (b). Like many other groups, this group felt that the narrative which resulted from this strategy was ‘smoother’, more balanced and more neatly structured. Another important feature of this narrative which also reached a wide measure of agreement is the narrowing of focus created through the placement of sentences (j), (g) and (c). Almost invariably, sentence (j) precedes both (c) and (g), the most common configuration across all groups being: (j) (c) (g)
It rained hard (…) There were pools of water in the courtyard (…) There were wet dead leaves on the paving of the courtyard (…)
The justification offered for the sequence is illuminating. Group 2 argued, for instance, that this configuration is not only cohesively linked by a broad lexical set including the items ‘rained’, ‘water’ and ‘wet’, but it is also linked in terms of an implicit cause and effect relationship between the propositions expressed by each sentence. On the basis of largely commonsense reasoning, the following argument could be proposed, where the second and third sentences are predicated by the first: It rained and so there were pools of water… and so there were wet dead leaves… Again, it should be stressed that the relationship between the propositions is implicit, the absence of formal connectives forcing readers to infer cause-and-effect relations. Whilst a number of groups had (j) adjacent to either (c) or (g), there was a tendency to seek maximum separation between (c) and 111
EXPLORING NARRATIVE STYLE
(g). Group 2’s justification for this was echoed unanimously: (c) and (g) were felt to be too repetitive to be placed side by side. For one thing, both are existential sentences beginning with the dummy subject ‘there’; for another, both contain virtually identical sentencefinal prepositional phrases (‘in the courtyard’, ‘of the courtyard’). In short, the very similarity in structure of the two sentences makes their proximity to one another feel ‘unnatural’. Another common pattern reflected in the Group 2 narrative is the tendency to make sentence (a) the ending of the story. Again, the justification for this is interesting as it draws on both structural and thematic evidence. First of all, sentence (a) is part of the descriptive frame which is woven around the actional frame, providing a particularly apposite ending with its echo of ‘the hospital’ referred to in the opening sentence (a cohesive tie of repetition, if one recalls). The thematic evidence concerns the ‘allusion’ created by the phrase ‘nailed shut’. This group, and indeed the others who chose an (a) ending, argued that this developed an image of a coffin being nailed shut, thereby providing a symbolic analogue to the death of the cabinet ministers. This decision to adopt an ‘emblematic’ reading is especially telling, as this particular sentence never provoked such an interpretation in tutorials which used the unaltered, original version. Comparable identifications of an image of death in other parts of the descriptive frame (for example, in the reference to the ‘dead’ leaves in sentence (g)), which were not made in previous tutorials, were also made in the course of the present exercise. Of course the absence of any such response to this sentence in the original version might have been due in part to its location in the story. However, if this and other related points are to be developed more fully, it is to the original we must now turn. In the following subsection, therefore, the ‘real’ version will be produced and discussed.
4.3.3 The original version Here, at last, is the original Hemingway vignette:
(i) They shot the six cabinet ministers at half past six in the morning against the wall of a hospital. (c) There were pools of water in the courtyard. (g) There were wet dead leaves on the 112
EXPLORING NARRATIVE STYLE
paving of the courtyard. (j) It rained hard. (a) All the shutters of the hospital were nailed shut. (e) One of the ministers was sick with typhoid. (f) Two soldiers carried him downstairs and out into the rain. (d) They tried to hold him up against the wall but he sat down in a puddle of water. (k) The other five stood very quietly against the wall. (h) Finally the officer told the soldiers it was no good trying to make him stand up. (b) When they fired the first volley he was sitting down in the water with his head on his knees.1
Dealing firstly with the similarities between this and the student reconstructions, it is noticeable that the preferred opening is vindicated: sentence (i) is indeed the first sentence of the story. (The reasons why (i) was considered most appropriate need not be re-introduced here.) The agreement reached over the (h)–(b) and (e)–(f) pairs is also borne out by the original. Noticeably, however, both sequences are part of the actional frame, and there is a strong sense of a temporal progression from the first element of the pair to the second. It would be difficult to produce a coherent pattern by reversing these sentences. Beyond this, there are few similarities between the original and the reconstructions. There are a number of substantial differences between the texts, many of which are highly illuminating. There is broad disagreement on the distribution of the actional and descriptive frames, for example. In the Hemingway version, these frames are separated into two blocks which, if we leave aside the opening sentence, comprise five sentences of description followed by five sentences of action. These are not at all interwoven in the style of most of the reconstructions. Moreover, although the (h)–(b) configuration is widespread in the reconstructions, no group even contemplated this as an appropriate ending to the story. When confronted with the actual ending, all the groups were surprised and many felt that (b) rendered the story incomplete, or, at least, indeterminate. Even more surprise was expressed at the sequencing of sentences in the descriptive frame of the story. Contrary to the intuitions of the participants, the anticipated (j), (c), (g) is dispreferred in favour of the following: (c)
There were pools of water in the courtyard. 113
EXPLORING NARRATIVE STYLE
(g) (j)
There were wet dead leaves on the paving of the courtyard, It rained hard.
The focalisation created through this pattern is almost the reverse of that expected, with a narrowing to the detail of the ‘leaves’ in the second sentence, followed by an abrupt transition outwards to a general statement about the rain in the last sentence. Furthermore, the cause-and-effect relationship which forms the basis of most of the students’ predictions is actually reversed to an effect-and-cause relationship. Another consequence of this pattern is that (c) and (g) are now side by side—a configuration which was ruled out by every group. Many felt that combining (c) and (g) in this way was repetitive and clumsy, almost becoming, to use Leech and Short’s term, overcohesive (1981:252). More will be said on the discrepancies between the workshop and Hemingway versions in the following section, where some further explanations will be offered as to why such dissimilarities occur. But one thing that the introduction of the original version should have highlighted is how a break-up and subsequent reconstruction of the text is possible in the first place. There is a marked lack of formal connectives between the sentences of the text, allowing considerable manipulation of their sequencing to take place without much disruption to the overall meaning of the story. This, in turn, leads to a clearer picture of the technique—some would say lack of technique—of the author. There is little, if any, authorial commentary on the events of the story, just as there is little signposting as to how one event precipitates another. It is not made explicit either how one element in the descriptive frame is supposed to interconnect with another. One general outcome of the comparison between the original and reconstructions is that it brought out the very blandness and the almost self-conscious ‘flatness’ of Hemingway’s style. Moreover, the linguistic analysis undertaken during the comparison provided concrete support for these observations. Indeed, many of the groups actually felt that their own efforts were ‘neater’, more ‘polished’ and had more all-round ‘literary merit’. This in itself provides some considerable insight into what students of literature regard as ‘literary’ at least in terms of the kinds of structures which they expect literary texts to exhibit. 114
EXPLORING NARRATIVE STYLE
4.4 Extending the analysis 4.4.1 Cohesion In the course of the analysis and discussion of the previous section, a number of key terms from Halliday and Hasan’s model of cohesion were introduced. Although this was done in an ad hoc and somewhat piecemeal fashion, it was intended to mirror the direction taken by the workshop itself. In other words, linguistic terminology was not introduced prior to the comparative analysis, but was instead allowed to unfold during the course of the analysis. Among the specific terms introduced were the notions of anaphora and cataphora. Anaphoric references, which ‘point back’ to earlier references in the text, provide one means of developing coherence in a narrative. In relation to this, a number of cohesive chains have been identified in the analysis, the most significant of which are the progressions of phrases relating to the cabinet ministers and the soldiers. By contrast, a more striking pattern of cataphora is established early on with the pronominally reduced form ‘they’ used to refer to the soldiers. The pronoun here therefore ‘points forwards’ to subsequent elaboration in the form of full noun phrases such as ‘two soldiers’, ‘the officer’ and ‘the soldiers’. Cohesion through antonymy is also created in the text through the juxtaposition of the verb phrases ‘stand up’ and ‘sitting down’ in adjacent sentences (h) and (b). The first of these phrases is also a repetition of ‘stood’ in the sentence immediately preceding (k). However, these largely straightforward patterns of cohesion were offset by a number of ‘anomalies’ in the original. The ‘overcohesiveness’ in adjacent sentences (c) and (g) was commented on particularly, with workshop participants expressing dissatisfaction at the repetitiveness of the ‘there were…’ construction and the similarity of the phrases ‘in the courtyard’ and ‘of the courtyard’. Also noted was the general absence of connectivity between the sentences in the story. A wide variety of devices are available in English for developing this type of cohesion, including additive (‘and’), adversative (‘but’) and causal (‘so’) conjunctions. However, only one explicit signal is provided in the form of ‘finally’ (sentence (h)) and yet, ironically, the sentence which it governs does not constitute the last of the story. Indeed, much of the benefit of this type of comparative stylistics is that 115
EXPLORING NARRATIVE STYLE
¶
it foregrounds the discrepancies between the students’ and the original versions, thereby highlighting the peculiar characteristics of the latter.
4.4.2 Natural narrative Part of the point of an exercise such as this is that it can be supplemented with a variety of additional linguistic models which enable the analysis to be developed in a number of different directions. One such model, which would not only provide an insightful analysis of the story in its own terms but would also help clarify some of the reconstructions discussed earlier, is the framework of natural narrative proposed by the sociolinguist William Labov. Although this is not the place to undertake a detailed introduction to the model, a brief sketch will none the less prove useful. Labov defines a natural narrative as a method of recapitulating past experience by matching a sequence of clauses to the sequence of events which (it is implied) actually occurred. A minimal narrative can be defined as a sequence of two clauses which are temporally ordered: that is, a change in their order will result in a change in the temporal sequence of the original semantic interpretation (Labov 1972:359–60). From the dozens of stories collected, Labov proposes the following narrative categories:
116
EXPLORING NARRATIVE STYLE
With the exception of Evaluation, the categories listed above are arranged in the sequence in which they would occur in a typical oral narrative. Evaluation is situated outside the central pattern and can be inserted at virtually any stage during a narrative. A fully formed narrative realises all six categories, although many narratives may lack one or more components. The Labovian model can be used to explore further the similarities and differences between the idealised and original versions. In this regard, perhaps the most significant point of agreement between the two versions is the correct selection of sentence (i) by the students as an opening: They shot the six cabinet ministers at half past six in the morning against the wall of a hospital.
117
EXPLORING NARRATIVE STYLE
Interestingly, this sentence, in the context of the story as a whole, satisfies the criteria for a well-formed Abstract. It provides a short summary of the main event of the story, although in itself does not constitute a narrative. There was also substantial cross-group agreement on the selection of this feature, with most groups arguing that the ‘summarising’ nature of this sentence made it a strong candidate for an opening to the story. The one group who placed (i) towards the end of their story did so for reasons which will be discussed shortly. A major pattern of disagreement centres on the distribution of what were loosely termed the descriptive and actional frames of the story. Hopefully, it should be clear from the foregoing that these two informal categories may be now aligned with the natural narrative categories of Orientation and Complicating action respectively. Orientation provides the all-important background detail which helps the listener/reader identify the ‘who, what, when and where’ of the story. It is realised by sentences such as the following: (c) (g) (e)
There were pools of water in the courtyard. There were wet dead leaves on the paving of the courtyard. One of the ministers was sick with typhoid.
By contrast, the Complicating action of the story is realised primarily through those sentences which have an actional verb in the simple past. These core narrative clauses are represented by sentences (f) and (d) in the following sequence from the original version: (e) (f) (d)
One of the ministers was sick with typhoid. Two soldiers carried him downstairs and out into the rain, They tried to hold him up against the wall but he sat down in a puddle of water.
Much wider agreement was reached on such patterns, where the sentences are temporally ordered and where a change in their sequence would result in a change in the interpretation of the chronology of the story. Furthermore, a reversal of (f) and (d) above would be difficult to justify on the grounds of coherence. Where the idealised and original versions diverge most noticeably is in the way in which the descriptive (Orientation) frame is deployed. Whereas it is scattered throughout the idealised version, in the 118
EXPLORING NARRATIVE STYLE
original version it is confined to the first half of the story. In fact, the original version is virtually split evenly into two blocks of Orientation and Complicating action. The idealised version, by contrast, is peppered with sentences of Orientation which break up what would otherwise be almost continuous Complicating action. (Consider, for example, the (d)–(g)–(k) and (h)–(b)–(c) patterns.) Students’ reactions to the pattern of the original version were interesting: they argued that it felt too ‘top-heavy’, too ‘symmetrical’ and was on the whole rather unsubtle. In support of their own mixed pattern, they contended that a more balanced narrative emerged, with the transitions between Complicating action and Orientation creating a more ‘engaging’ and ‘dialogic’ (sic) story. This last point leads us directly to the question of the problematic ending to the story. It may be recalled that no group contemplated placing the following sentence at the end: (b)
When they fired the first volley he was sitting down in the water with his head on his knees.
From an analytic point of view, this sentence raises a number of issues. For one thing, it tends to cut across the distinctions between narrative categories established in this section. It satisfies to some extent the criteria for the recognition of Complicating Action: it has an actional verb in the simple past (‘fired’) and generally has a ‘what happened’ rather than ‘who, what, where’ feel to it. Yet a couple of features in the sentence push it towards Orientation: it contains an explicit temporal signal in the form of ‘when’ and, more significantly, contains a past continuous verb phrase (‘he was sitting’). Both features are more characteristic of Orientation sections of natural narratives. This categorical indeterminacy is what many groups felt renders the story itself indeterminate. It possesses no clear Coda, no signal that the events of the narrative are over and the narrator is, so to speak, giving up the floor. Group 5 felt, for example, that sentence (i), the opening of the original and idealised versions, accomplished this much more effectively—which is why they placed it near the end in their version. Certainly, all of the groups were unhappy with (b) as an ending and no group placed it more than three sentences from the end. Instead, more ‘symbolic’ Codas were preferred, such as (j) (‘It rained hard.’), (g) (‘There were wet dead leaves on the paving of the courtyard’) and, most popularly, (a) (‘All the shutters of the hospital were nailed shut’). 119
EXPLORING NARRATIVE STYLE
The first of these was chosen because it is by far the shortest sentence of the story, whilst the remaining two contain, respectively, explicit and symbolic allusions to death. All of these choices, it was felt, not only avoided the sense of incompleteness generated by (b) but also rounded the story off more neatly with a short descriptive statement devoid of any actional context. Much more could be said of the implications of a natural narrative analysis here. The lack of Evaluation in the story, for example, has not been touched upon. To do so would in turn lead to questions of authorial modality, mind-style and point of view—and such questions really deserve fuller treatment elsewhere.2 One consequence of Hemingway’s non-evaluative technique is the notable absence in the story of any rhetorical ‘colour’. There are no stylistic flourishes, such as departures from the basic syntax through the use of modals, questions and negatives. Indeed, one might wonder how the same basic story line would look if supplemented with those evaluative devices. Would rewrites, rather than reconstructions, reveal more about what participants informally termed the ‘flatness’ of Hemingway’s style? It is to issues such as this that the following short section will be addressed.
4.5 Practical activity: creative writing Up to now, information on the biographical and literary-critical context of the story under analysis has been quite consciously withheld from the study. This has not been done in the interests of achieving a restrictively formalist or text-immanent reading of the story, nor is it suggested that contextual detail does not have a part to play in a multidimensional, interpretative process. On the contrary, it has been done simply because, as a language-orientated activity, the concern with linguistic form and function has been uppermost. As a preparatory stage in the development of the ‘sub-exercise’ which follows, however, we shall find it necessary to introduce precisely such contextual detail. The actual historical event portrayed in the vignette is the execution of six Greek cabinet ministers, including the ex-premier, in Athens in 1922. The executions took place after Greece’s unsuccessful campaign against Turkey. Hemingway had previously covered the war for the Toronto Star and, although based in Paris at the time, would have 120
EXPLORING NARRATIVE STYLE
followed these events closely. Three months after the executions, he wrote his vignette, probably basing it on a newspaper account he read in the Paris edition of the New York Tribune (Baker 1969:108; Reynolds 1972:82). The full text of the story, including a short preliminary editorial, is reproduced in full within the ruled text.
ATROCITIES MARKED GREEK EXECUTIONS OF FORMER LEADERS
Uncensored Account Brought From Athens—Dead Man Was Propped Up in Line GOUNARIS NERVED BY DRUG Ex-Premier, Dying From Illness, Was Artificially Stimulated to Stand LONDON, Dec. 20—The Daily Express published the first detailed account of the recent executions of the Greek exMinisters supplied by its correspondent who was lately in Athens. . . . . . . . . . . . . . . . . . . . . . . M.Gounaris, an ex-Premier, was in a hospital in a very critical condition. About 11 a.m. he was taken out on a stretcher, placed in a motor van and driven to a place about one and a half miles outside of the city. He was left lying on his stretcher in a dying condition while the car went back to fetch five others from the prison where they had all been confined in a single room. To begin the horrors of that morning it was discovered by the guards that one of the five had died in the van on the way out from heart failure. On the arrival of the van Gounaris was lifted out of the stretcher to stand up and face a firing party. It was then found that this wretched man, who, after all, had been a figure in the recent history of Europe, was unable to stand at all. He was thereupon given sufficient injections of strychnine to strengthen the action of his heart to enable him to stand up in front of the firing party. The man who had died on the way out was propped up beside him—a ghastly line of four live men, one half alive and one dead man. They were then asked—Gounaris, the dead man and all—if they had anything to say, an appalling instance of mockery. No reply was 121
EXPLORING NARRATIVE STYLE
made, but M.Baltazzis took out his monocle, polished it and put it back again. General Hadjanestis calmly lit a cigarette. The order to fire was given. The moment the prisoners fell the firing party rushed forward and emptied their revolvers into the corpses. Including that of the man who had died on the way from the prison. The bodies were then thrown into a lorry and taken to a public cemetery just outside of the city and were thrown out casually in a heap in the mud which covered the ground.3
¶
122
The compositional differences between Hemingway’s short story and its source material needed only be sketched here. Most noticeable perhaps is that the dead minister of the newspaper account is removed completely, thereby channelling attention towards the sick minister in the vignette. Indeed, it would be difficult to cover both a sick man and a dying man in so short a space. The time of the execution is also altered, being brought forward to a time traditionally associated with ‘dawn’ executions. Extrapolating from a reference to ‘a heap in the mud’ in the last sentence, Hemingway foregrounds the inclement weather through a series of references to rain and wet conditions. The location is also altered with the executions conveniently taking place just outside the hospital, as opposed to one and a half miles away. Again, this allows for further economy of description. Interesting also are the allusions by Hemingway to the ‘nailed shut’ windows of the hospital which, whilst provoking particular comment in the workshop discussion, are on the basis of this evidence entirely fictitious. Although the juxtaposition of short story and source material would provide in itself a useful framework for an informal discussion of composition and technique, the activity proposed here involves developing a story based on the newspaper account. I designed the following protocol for use as a variation on the central workshop plan (introduced in section 4.2) and have found that it works particularly effectively with participants who want to develop their compositional and creative skills in English language. It basically involves writing a short story using only the newspaper account, without any prior access to Hemingway’s story. (Of course, the stories collected in this way can later be compared with the Hemingway vignette.) Reading through the protocol, you might care to imagine what sort of story you would produce on the basis of these directives.
EXPLORING NARRATIVE STYLE
1 Read the accompanying newspaper report which describes certain events that took place in 1922, The report appeared in the New York Tribune and details the execution of six Greek cabinet ministers. 2 Now compose a short story based on the events portrayed in the newspaper text. Do not exceed two hundred words. Feel free to:
¶
(i) omit events that you feel are less central to the main story, (ii) amplify certain details which you feel are thematically significant. In short, you may use your imagination. 3 Now explain how you reached your particular story. Have you felt it necessary, for example, to delete certain events during the transition from news report to short story? On the other hand, have you embellished aspects of the news report for thematic or ‘literary’ reasons? 4 Now examine a short story written in 1922 by the American writer Ernest Hemingway. This story is believed to have been based on the newspaper account you have just read. Are any features of Hemingway’s version particularly striking? Make a list of the features if you can. What are the principal ways in which your story differs from that of Hemingway?
Having paid close attention to Hemingway’s vignette, we are probably stylistically ‘disaffected’ by now, and will be difficult to read the newspaper account without being influenced by his narrative. However, on those occasions where I have elicited stories through the protocol and where the workshop participants have had no prior awareness of the Hemingway text, the results are both striking and extremely insightful. For a start, most of the elicited narratives are written in the first person. This technique often presents a highly personalised view of events, mediating them through the consciousness of a character who participates in the story. Here is one such version produced by a group of students who had not read 123
EXPLORING NARRATIVE STYLE
Hemingway’s story before they embarked on the creative writing protocol:
It was sometime last December—I can’t recall exactly. All I can remember is that the army called me just before lunchtime to pick someone up. I was to go to the hospital. It turned out to be someone important who was in bad shape. I wondered what the hell they wanted to do with him. My instructions were to pick up five other guys and to drive them one and a half miles outside the city. When I dropped them off, I realised one of them was dead. Maybe a heart attack. Fortunately, someone was there to give the guy on the stretcher a booster. Thank God it worked! He stood up. The guy who seemed to be in charge asked the six of them if they needed anything. But before they could answer they were shot dead.4
Many narrative devices are employed here which are thoroughly at odds with those used by Hemingway. For instance, a consequence of the first person narrative framework is that many events are reported through verbs which emphasise the narrator’s mental cognition and perception: ‘I can’t recall’; ‘I can remember’; ‘I wondered’; ‘I realised’. This type of verbal process, it was noted, was conspicuously absent from Hemingway’s third-person narrative framework. The students’ narrative is also embellished with explicit modality. Modality is language which expresses a speaker’s or writer’s opinion about the validity of what they say and which indicates whether they are certain or doubtful about the truth of the claims they make. While non-existent in Hemingway, modal devices are clearly present here: ‘I can’t recall exactly’; ‘Maybe a heart attack’; ‘the guy who seemed to be in charge’. A related narrative feature of the student text is the use of interpolations, asides and exclamations which signal the narrator’s reaction to the events of the story as they unfold: ‘what the hell’; ‘Fortunately’; ‘Thank God’. In terms of Labov’s natural narrative model introduced in subsection 4.4 2, the students’ story exhibits marked narrative Evaluation, 124
EXPLORING NARRATIVE STYLE
consisting mainly of explanations of how one event in the story precipitates another: ‘they called me…to pick someone up’; ‘I was to go to the hospital’; ‘My instructions were to pick up five other guys.’ This is supplemented by the use of a specific subtype of Evaluation known as an explicative, in which a post hoc explanation of a particular narrative event is offered when it becomes available later in the story: ‘It turned out to be someone important…’ All of these evaluative devices serve to establish a chain of cause-and-effect relationships which were so markedly absent from the Hemingway story. With reference to another category of the natural narrative model, Orientation, the students’ version extrapolates from details in the newspaper account in order to locate the beginning of the narrative within a specific temporal framework: ‘last December’; ‘before lunchtime’. Whereas these references to time are removed completely by Hemingway, references to physical location and the weather are, conversely, embellished substantially in his vignette. Finally, the last clause of the students’ version (‘they were shot dead’), which is in the simple past tense, is much more in keeping with the criterion for a narrative Coda than the equivalent clause in Hemingway (‘he was sitting down in the water…’). The openendness engendered by the past continuous stands in sharp contrast to the sense of finality and closure produced by the simple past. This is not to say that there are no points of stylistic similarity between the two stories. In the students’ version, for example, no characters are named—a rare characteristic indeed of the narratives elicited through this exercise. Similarly, there is a certain indefiniteness about the students’ story, conveyed primarily through repeated references to ‘sometime’ and ‘someone’, which again resembles Hemingway’s technique. Nevertheless, irrespective of the precise degree of similarity between the Hemingway and workshop texts, this activity invariably generates enough material to form the basis for productive comparison and it can always be used to locate and identify different features of language in narrative texts, whether they be modality, cohesion or natural narrative. Finally, it is worth noting that the students who created the present workshop narrative were quite startled when they came to read the Hemingway text—a response perhaps not often evinced in seminars on this writer’s short stories. In fact, I have had the feeling ever since that they much preferred their own version to that of one of the reputed greats of American fiction! 125
EXPLORING NARRATIVE STYLE
4.6 Summary The frameworks of cohesion and natural narrative are only two of many linguistic models which may be brought into play in the analysis of texts. What makes these two frameworks particularly appropriate for our present purposes, however, is their shared emphasis on the ‘surface’ features of the text. In other words, the models of cohesion and natural narrative offer clear indices of stylistic formulae which can easily be ‘read off against explicit markers in a story. Not all interpretations are as readily accessed, however, and there are other, more abstract layers of narrative communication exhibited by stories. For instance, in the review of the main workshop activity (section 4.3), it was noted how students made comments on the covert and symbolic allusions to death in the story. These inferences are not predicated on overt ‘surface’ clues in the text; rather, they are meanings which need to be actively mapped on to the text. Indeed, readers may have made similar inferences in their own reconstructions of the story. The type of meaning fabric engendered by this interpretative process is often referred to as coherence. Although an emphasis on coherence rather than cohesion would have shifted the scope of this chapter into quite a different area, it underscores the fact that texts are organised simultaneously on a number of different linguistic levels. In fact, the analysis of narrative texts may be undertaken from a whole host of linguistic perspectives; it has been the concern here to suggest but two of the range of possible models available for this type of language study.
Suggestions for further reading Cohesion and natural narrative The model of cohesion presented in this chapter is developed largely from that of Halliday and Hasan (1976), although see also Halliday (1994). The model of natural narrative is adapted from William Labov’s work on urban vernacular. A particularly comprehensive version of the model can be found in Labov (1972). Toolan (1988: chapter 5) provides a clear account of the key stages in the development of Labov’s framework. 126
EXPLORING NARRATIVE STYLE
Stylistic applications Both cohesion and narrative have received extensive and detailed treatment by stylisticians over the years. From the pool of important work on narrative structure is Carter (1984) which proposes a framework for the analysis of narrative openings. Carter partly bases his model on sets of responses by informants to disarranged sentences, and thus must be acknowledged as a study which prefigures the type of activity undertaken here. Stylistic work which draws more centrally on the Labovian natural narrative model includes Simpson (1987) which concentrates on a section from Hemingway’s The Old Man and the Sea. Haynes (1989) illustrates passim the stylistic potential of the Halliday and Hasan model, whilst Fowler (1986:53–68) provides a compact summary of the cohesion model as well as an analysis of Kurt Vonnegut Jnr’s Cat’s Cradle from this perspective. Finally, Simpson and Montgomery (1995) develop a broad-based model of narrative which they use to explore a variety of stylistic features in the prose and film versions of Bernard MacLaverty’s novel Cal. The related study of coherence in narrative has been informed by relevant stylistic work on, for example, discourse structure (Hoey 1989) and plot structure (Stubbs 1982). The second of these offers a semantic analysis of plot, and constitutes a useful supplement to the present study in so far as it focuses on Cat in the Rain, another Hemingway short story. Because of its focus on the inferencing strategies used to derive readings from a text, a study of narrative coherence may also encompass research in text-linguistics and artificial intelligence. Again, the potential of these two approaches has been amply explored by stylisticians, samples of which are: de Beaugrande (1987); Van Peer (1987); Werth (1994).
127
Dialogue and drama: an introduction to discourse analysis
• 5.1
Introduction
130
• 5.2
Dialogue and discourse
131
• 5.3
Models for the analysis of discourse
143
• 5.3.1 The analysis of discourse structure
143
• 5.3.2 Discourse strategy: maxims and relevance
148
• 5.3.3 Discourse strategy: politeness phenomena
155
• 5.4
Discourse analysis and drama dialogue 164
• 5.5
Summary
174
Suggestions for further reading
176
•
Chapter 5
Chapter 5
129
5.1 Introduction This chapter explores the nature of dialogue. It investigates the structure of verbal interaction and assesses some of the strategies that speakers and hearers use in conversation. Because of this remit, the literary genre that will offer the primary focus of attention in this chapter will be drama. It will be argued, amongst other things, that dramatic dialogue provides excellent source material for explaining the basic patterns of everyday conversation. And, odd as it may seem, play dialogue that critics dub ‘absurdist’ is an especially productive tool for foregrounding the routine and commonplace in verbal interaction. In language study, the term reserved for the highest level of linguistic organisation is discourse. Discourse refers to the structure and function of language beyond the level of the sentence. The units in discourse analysis are therefore relatively large and are certainly much larger than any of the linguistic units covered so far in this book. One of the primary emphases in discourse analysis is on explaining how conversation works. Discourse analysts are interested in how spoken interaction is structured and how speakers’ conversational contributions are connected. Not only will this emphasis be sustained throughout this chapter, but it will be orientated towards two interrelated areas. First, does conversation have a structure and, if so, what sort of structure? Second, what do speakers hope to achieve in interaction and how do they set about achieving it? In order to answer these questions, the chapter will be arranged as follows. The next section will establish some ground rules for discourse analysis. It will set out the broad principles of discourse organisation in general terms. The focus narrows considerably in section 5.3, where a phalanx of influential discourse models are assembled and evaluated. These models cover both the structure of discourse and the strategies speakers use in discourse. The ‘strategic’ models are drawn from that branch of discourse 130
DIALOGUE AND DRAMA
analysis called pragmatics, which is primarily concerned with the meaning of language in context. In keeping with the other chapters of the book, use will be made only of models that represent both a significant theoretical development in language study and a viable and accessible method for stylistic analysis. Section 5.4 is devoted to the analysis of a passage of drama dialogue and it draws on the analytic models assembled in the previous two sections. The text used is Edward Albee’s play The Zoo Story. Literary critics habitually categorise this play as ‘absurd’, although curiously, Albee himself tends not to be classified as a dramatist of the absurd. The analysis will attempt to demonstrate how this sort of play dialogue can offer fruitful insights into aspects of naturally occurring conversation. The final section of the chapter draws together some of the strands of the analysis and provides an inventory of important interactive features which can be explored in dramatic dialogue.
5.2 Dialogue and discourse At first glance, conversational interaction appears to be more loosely organised than the other linguistic levels we have studied in this book. For one thing, it is peppered with dysfluencies, non-sequiturs, false starts and hesitations. For another, there is simply no telling what exactly someone engaged in talk is going to say. Nevertheless, although discourse seems more ‘fluid’ than other linguistic levels, this is not to say that it has no underlying structure. On the contrary, we have very strong expectations about what should occur in interaction: questions normally evince answers, requests anticipate reactions while remarks and comments require at least some acknowledgement from other interactants. Moreover, all these speech activities will be expected to be delivered with a degree of politeness that befits the situation or task in hand, and for a speaker or hearer to deviate from these conventions is to do so at her interactive peril. Conversation, though complex and multifaceted, is still structured. Good evidence of what happens when this structure collapses is provided by communicative breakdowns. Here, for example, is a piece of real conversation between two sisters (A and B). Speaker A is planning to make a telephone call to a relative living abroad: 1
A: What’s the code to Los Angeles? 131
DIALOGUE AND DRAMA
B: Are you hungry? A: What? B: What? This ‘breach’ in the discourse framework was created by speaker B’s misinterpretation of A’s utterance. Curiously, she thought that A had asked ‘Do you want some coleslaw sandwiches?’, and consequently believed her response to have been appropriate in the context. As far as speaker A was concerned, it clearly wasn’t a coherent response and her second utterance precipitates the communicative breakdown. Often taken as a benchmark in discourse analysis are the following basic principles of conversational coherence: we have intuitions about what constitutes well-formed discourse; we rely on speakers and hearers to be generally co-operative in interaction; we assume that what people say to us has some degree of relevance. Like many aspects of language organisation, these principles are best illustrated when we encounter aberrant or ill-formed interaction. The shunting exercise that was developed in Chapter 3 (see section 3.3) is a good way of teasing out our expectations about interactive wellformedness. Combinations in discourse operate on a similar principle to grammatical combinations, with the collocation in lexical semantics broadly analogous to the category of exchange in discourse. An exchange is a combination of conversational contributions on the same topic produced by different speakers. Exchanges commonly (though not necessarily) consist of two parts. Here are four attested exchanges, each comprising two structural components:1 A: Hi! How’s it going? B: Oh, not so bad, and yourself? A: Right now, have the bowels been working OK? B: Mmm…yeah… A: What now do you think should be our reaction to yesterday’s mortar attack in Sarajevo? B: Well I’m someone who’s said that unless we take firm action then these problems will escalate. I mean, the best thing you can say about Bosnia—and there aren’t many good things you can say—is that it is a crisis that came before Europe was ready. 132
DIALOGUE AND DRAMA
A: Come on now, eat up your Crunchy Nut Cornflakes. B: No want it! No want it! With only decontextualised utterances to go on, our communicative competence is still sophisticated enough to provide us with a lot of information about these exchanges. Not only can we deduce their probable sources, but we can even make predictions about the status, identity and power of the participants involved in the dialogue. As it happens, the first exchange is an informal opening to a conversation, the second is from a doctor-patient encounter, the third from a political interview and the last from a domestic encounter between an adult and a toddler. All four exchanges represent nothing out of the ordinary. However, by simply nudging the second utterances in each exchange into the equivalent positions in the exchange immediately following, the following rather surreal collection is produced: A: Hi! How’s it going? B: No want it! No want it! A: Right now, have the bowels been working OK? B: Oh, not so bad, and yourself? A: What now do you think should be our reaction to yesterday’s mortar attack in Sarajevo? B: Mmm…yeah… A: Come on now, eat up your Crunchy Nut Cornflakes. B: Well I’m someone who’s said that unless we take firm action then these problems will escalate. I mean, the best thing you can say about Bosnia—and there aren’t many good things you can say—is that it is a crisis that came before Europe was ready. It is through the very peculiarity of these shunted exchanges that the commonplace and prosaic in ‘normal’ discourse is foregrounded. Furthermore, the juxtaposition of the original set with the second highlights the subconscious and perhaps taken-for-granted assumptions that we have about coherence and wellformedness in discourse. It also underscores expectations about the appropriateness of topic in interaction. For instance, outside a medical context, human bodily emissions make an unsuitable or at least unlikely topic of discourse. 133
DIALOGUE AND DRAMA
Explanations for why the shunted exchanges should look so outlandish will unfold over the course of the next two sections. For the moment, readers can use their intuitions to tease out the bizarre features of each exchange. Your hunches about why these examples sound weird may be more profound than you think—in fact, you may be taking an important step towards defining the ‘absurd’ as a dramatic genre! How can we account systematically for the routines, patterns and structures that are the mainstay of conversation? The best way to start is by painting a picture of discourse organisation with the widest brush strokes possible. This picture can then be embellished by introducing some specific models for analysing discourse. To begin the ‘macro-survey’ of general patterns in discourse, it is worth reintroducing the two axes of language that were developed early in section 3.3 to account for patterns of collocation. These intersecting axes, the syntagmatic and the paradigmatic, also serve as a useful conceptual model for explaining how conversation works. The syntagmatic axis, it may be recalled, is the axis of ‘chain’. It forms the structural frame along which combinations of units are strung. These units are linked in the relationship: a and b and c. The paradigmatic axis is the axis of ‘choice’. It accounts for the pool of possible entries for a given slot in the structural frame and links elements in the relationship: a or b or c. When mapped on to discourse, these axes realise the twin poles of structure and strategy (Figure 5.1). The structural axis is a neat way of conceptualising how discourse units are combined and built up to form exchanges. Put another way, the structural axis accounts for the linear progression of discourse as
FIGURE 5.1 Structure and strategy 134
DIALOGUE AND DRAMA
a sequence of exchanges. We have already seen four exchanges, consisting of two structural elements apiece, in our shunting exercise. Exchanges often take the form of familiar pairings such as ‘questions and answers’ or ‘statements and acknowledgments’. However, exchanges can also encompass non-verbal behaviour. If, for example, a speaker issues a request, then a non-verbal reaction from the hearer is normally warranted. Even though its second part is not a verbal utterance in the strictest sense, this ‘request-reaction’ sequence none the less constitutes an exchange in discourse. Here now are the three principal types of discourse exchange—questioning, stating, requesting—arranged along the syntagmatic axis: 2 3 4
A: What’s the time? & B: Three o’clock. A: That was a great goal. & B: Yeah, you bet. A: Open the window. & B: [opens the window]
With its focus on strategy, the paradigmatic axis foregrounds the ‘tactical’ nature of discourse. This axis represents an utterance as a choice from a pool of available options. Consider again the ‘requestreaction’ sequence expressed by example (4). While speaker A’s utterance is unquestionably efficient and clear as a directive, its forthrightness will be interpreted as peremptory and rude in many contexts. There are other, more subtle ways of getting a hearer to perform the same operation. A stratagem like ‘Could you open the window please?’ is less direct, though, of course, it is rather more tangential to the task required of the hearer. Even more oblique is a gambit like ‘Goodness me, it’s hot in here!’, which, although subtle, is even further removed from the task required. In Figure 5.2 all three strategies are arranged along the paradigmatic axis. They are ordered according to the degree of directness which they exhibit to the service requested. The paradigmatic axis thus forms a strategic continuum ranging from ‘direct’ to ‘indirect’ along which different types of utterances can be plotted. There is however a crucial third dimension in discourse analysis. The contextual setting of an interaction is both a significant determinant of discourse structure and an important influence on discourse strategy. Setting is the non-linguistic context which envelops a piece of communication. This does not, however, just mean the physical environment of interaction; it also extends to the assumptions and beliefs that people bring to discourse. The setting, 135
DIALOGUE AND DRAMA
FIGURE 5.2 The strategy axis moreover, is constantly changing as a dialogue progresses. And this change need not necessarily occur because, say, a bull elephant happens to crash into the room during a conversation. It may simply be that a new framework of shared knowledge develops between interactants as discourse develops. For instance, if you suddenly blurt out to someone in the course of a conversation that you love them, then clearly a new discourse context is created which will affect profoundly subsequent patterns of interaction. The importance of setting as a third dimension in discourse can be illustrated with a brief example. Here is an actual exchange between two speakers which I witnessed a few years ago: 5 A: er…you have to go to head office…please B: [nods and rises to go] Even without any description of its setting, it is still possible to work out that this is a ‘request-reaction’ exchange of the sort we have just been discussing. However, this structural description, while straightforward enough, is not especially interesting in itself. Exploration of the strategy axis reveals more about the exchange. For example, the speaker employs two particles either side of the main content-bearing component of the utterance: ‘er’ and ‘please’. Both particles, in tandem with the hesitation used around them, mitigate the overall force of the request function. The impact of the central sequence of the utterance is also comparably softened. The ‘have to’ 136
DIALOGUE AND DRAMA
phrase, like its counterpart ‘must’, operates as a marker of obligation. However, unlike ‘must’, it expresses objective obligation. In other words, the requirement expressed by ‘have to’ is presented as a general rule and not something that can be attributed directly to the speaker. The use of ‘must’, on the other hand, makes the obligation encoded in the request more like a subjective decree that is passed from speaker to hearer. All of these features combine to push the utterance towards the indirect end of the strategy continuum, reducing the overtly ‘commandlike’ status of the request more to the status of a ‘suggestion’. Yet it is only with an account of the setting in which this exchange took place that the full picture begins to emerge. The exchange took place in a travel agent’s shop in Liverpool. Speaker A is a young black manageress who emerges from her own office into the main office. The addressee is a white middle-aged male employee who is positioned behind a desk and is dealing with customers. Descriptions of discourse setting often help explain the motivation behind linguistic strategy. Here, the woman is the ‘powerful’ interactant, at least in terms of institutionally sanctioned employment practices. Nevertheless, in spite of her status, she treads a cautious interactive path. In designing her utterance, she avoids the higher risk ploys from further up the strategy continuum, opting instead for an interactive gambit which appears less direct and less coercive. This more cautionary strategy may also have been prompted by the fact that she is entering the physical ‘territory’ of her addressee. She leaves her own space, and is moving towards a stationary interlocutor who is comfortably ensconced behind a desk. The outcome of her gambit is a successful one: she gets someone to do something without having appeared to have given them an order. However, the exchange as a whole raises a number of issues to do with the complex interaction between language and social roles. Whereas perceived power and status is one determinant of strategy, this is perhaps counterbalanced in this exchange by a web other social variables such as gender, age and ethnicity. This interrelatedness of discourse strategy and discourse setting will be spotlighted further in the next section when attention is focused on the linguistic routines of politeness. We are now in a position to replace our two dimensional discourse model with a model which comprises the three components: structure, strategy and setting. These three ‘Ss’ can be employed as a handy mnemonic for discourse analysis at its broadest level. Although clearly interrelated, the three parameters enable the analyst to sift out and concentrate upon different aspects of discourse 137
DIALOGUE AND DRAMA
organisation. Figure 5.3 is the revised model which depicts setting as a circle embracing the other two axes. Any piece of discourse can be viewed as an intersection of the three Ss. As we saw in the shunting exercise above, it is often possible to deduce information about the three Ss even with only decontextualised utterances to go on. Further experimentation along these lines can be carried out using setting as a contextual variable. By taking one utterance and offering a selection of different contexts, the manner by which setting acts as a constraint on discourse appropriateness can be highlighted. Here is a short two-part exercise in which readers may care to participate. It involves taking a decontextualised chunk of language and mapping it against three possible scenarios. The three scenarios are: Setting 1 A chance encounter between two middle-aged strangers at a bench in a public park Setting 2 An eight-year-old child in conversation with her father’s adult friend whom she has met for the first time Setting 3 A doctor engaged in a diagnosis of a patient’s illness The next stage involves taking the following utterance and assigning it to one of the given settings: 6
I’ve been to the zoo. I said, I’ve been to the zoo. MISTER, I’VE BEEN TO THE ZOO!
Make a note of your decision about which setting fits the utterance, and if possible, the reasons which led you to make this decision. You could also try to write brief utterances which would fit the remaining two settings. The second part of the exercise is as straightforward as the first. It involves matching up vocabulary, register and setting. The point that can be made here is that even a lexical set (see subsection 3.4.1), which is much less developed than a fully formed utterance, is often strongly associated with a particular discourse setting. Here then is a lexical set which contains some rather grisly terms: 138
DIALOGUE AND DRAMA
FIGURE 5.3 A revised discourse model: the three ‘Ss’ 7
prosthesis lung cancer cancer of the mouth doctor jaw
Again, the task is to assign this set to one of the three settings provided. If possible, draw up a lexical set which would be predicted by the remaining two settings. I imagine there will have been little difficulty in aligning the three settings with the two pieces of data provided. Given the straightforwardness of the exercise, readers may indeed be wondering where its heuristic value lies. For now, simply make a note of your reactions to the two protocols. We shall have good cause to return to them later. There remains one final matter that needs to be cleared up in this section before we move on to assess more specialised frameworks for discourse analysis. This is to serve partly as a note of caution and partly as a way of consolidating one or two earlier points. Although they have been touched upon earlier in the book (sections 1.3 and 2.4.1), nowhere in this discussion has there been any mention of the three important grammatical terms imperative, interrogative and declarative. When talking about discourse, we need to be very careful about how these terms are used. This is mainly because they describe the formal make-up of sentences and not the function of utterances in context. Although they are often erroneously used interchangeably with discourse terms, imperatives, interrogatives and declaratives are 139
DIALOGUE AND DRAMA
really types of grammatical mood. Why this confusion is a particular problem in discourse analysis will become clear soon. Before that, a basic description of these three key categories of mood is required (Figure 5.4, see opposite). Imperatives differ principally from the two indicative moods, declarative and interrogative, in that they express no grammatical subject.2 Furthermore, the verb which fronts imperative constructions is always in its base form and so cannot be marked for tense. (See what happens when you try to convert any of the four imperatives above into a past tense.) Indicatives, on the other hand, contain both subjects and verbs, although they may be further subdivided according to the position of these elements. In declaratives the subject simply comes first. In many interrogatives, either all or part of the verb phrase comes first which is then, in turn, followed by the subject. In the following interrogatives, the verb phrases are indicated by underlining and the subjects by italicisation: 12 Has Roisín closed the door? 13 Is the cat in the garden? 14 Can they sing well? Sometimes a special verb, do, needs to be imported in order for this grammatical operation to be carried out: 15 Did the snake eat it? On other occasions, a so-called ‘WH-word’, like where, why or how, can be incorporated into the interrogative: 16 Where is the cat? This enables the interrogative to be used to prompt a more open kind of response rather than the ‘yes’ or ‘no’ that would be elicited by examples (12–15). The three principal moods are commonly associated with particular discourse functions. An interrogative is the form standardly used for asking questions; a declarative is the form standardly used for making statements; an imperative is the form standardly used for commands or requests. However, if ever there was a crucial axiom about the interaction of grammar and discourse it is this: there is no necessary correspondence between the mood of a sentence and its 140
DIALOGUE AND DRAMA
FIGURE 5.4 Important types of mood in English function in discourse. For instance, you can do a lot more with an imperative than just give an order. Here are a few instances of imperatives which are clearly not commands. Notional descriptions of their respective discourse functions are provided beside each example: 21 22 23 24
Have a drink. Have a nice day. Blow me down! Look after yourself.
[offer] [ritualistic parting] [exclamation] [exhortation]
Interrogatives are similarly flexible. Here is a selection of interrogatives of which not one is a genuine question for information: 25 26 27 28
Are you still here? Could you pass the salt? Will you have a drink? Is the Pope a Catholic?
[command] [polite request] [offer] [rhetorical question; i.e. no reply elicited] Finally, the declarative form also has many discoursal possibilities. For instance, it can function as a command in an utterance like the following: 141
DIALOGUE AND DRAMA
29 That door is still open. In fact, the same declarative sentence can carry a number of discoursal functions depending on the intonation (rises or falls in pitch) which it exhibits. Consider the following declarative: 30 You’ve had enough cake. If uttered with a rising tone in the context of, say, a polite dinner party, it can function either as an offer or as a genuine request for information. Yet when uttered with a pronounced falling tone, by a mother to a toddler, then it is clearly a command and it signals to the youngster that she is out of luck! Sometimes, we don’t even need to have a recognisable mood at all to get things done in language. The following construction contains no verb so it is moodless. Nevertheless, its request function would be clear in any context: 31 The door, please. This ‘lack of fit’ between form and function is an important strategyframing device in discourse. The strategy continuum, it may be recalled, runs between the poles direct and indirect (see Figure 5.2, above). The form-function asymmetry identified here provides a valuable criterion for measuring this continuum systematically. The most direct strategy will be one which draws on the mood type that is standardly associated with a particular discourse function. For example, a direct request will be one which employs an imperative construction, as in ‘Open the window’. A less direct utterance will employ an oblique grammatical form, such as the use of an interrogative to make a request. This is precisely the tactic that is used in the medial example on the strategy continuum: ‘Could you open the window?’. The most indirect strategy of all is when an utterance employs not only an oblique form, but also has no overt semantic link with the ostensible purpose of the exchange. Notice how, in the most indirect request on the continuum, there is no actual mention of the service requested: ‘Goodness me, it’s hot in here!’ This is a kind of optimal indirectness which is so opaque in character that it tests the strategic nature of discourse to the limit. Explaining precisely how and why speakers should choose to behave in so linguistically enigmatic a fashion will be one of the main tasks of the next section. 142
DIALOGUE AND DRAMA
5.3 Models for the analysis of discourse This section offers a general overview of some of the important developments that have taken place in discourse analysis over the last two decades. One of the main purposes of this is to suggest ways in which discourse models can be productively applied to the analysis of talk. The overview that follows is organised in accordance with the broad principles of discourse organisation that were established in the previous section, thereby enabling several diverse strands of work to be brought together under a few basic overarching principles. The subsection that follows will focus specifically on discourse structure. The remainder of the section will be devoted to discourse strategy. The third parameter on our discourse model, setting, will be integrated where appropriate with those discourse models in which it plays a key role.
5.3.1 The analysis of discourse structure One of the most significant advances in the structural analysis of verbal interaction has been the development over the years of the ‘Birmingham model’ of spoken discourse. This model is so named because it has been pioneered by the English Language Research Unit based at the University of Birmingham. The remit of the early research undertaken by the unit was to design a model of discourse that could account for classroom discourse. The object was therefore to isolate the exchange patterns that typified interaction between teachers and pupils during lessons. It emerged in the course of the research that exchange patterns like the following were typical of classroom discourse: 32
Teacher: What’s the capital of France? Pupil: Paris. Teacher: Yes, Paris. That’s right.
This teaching exchange consists of three basic components: an initiation by the teacher, a response by the pupil and some feedback from the teacher. However, while clearly appropriate in a classroom setting, this three-part pattern is not especially common in much other naturally occurring language. Unless you happen to be running 143
DIALOGUE AND DRAMA
a quiz, you do not normally know in advance the answers to the questions you ask people. If you did, then exchanges like the following would be a perfectly normal part of everyday conversation: 33
A: What’s the time? B: Three o’clock. A: Yes, three o’clock. That’s right.
The need to account for interaction outside the classroom prompted extensions to the early model, the most substantial of which was Deirdre Burton’s study of discourse structure (1980). Burton revised the classroom model to enable it to deal with all sorts of talk and, interestingly, she chose to illustrate many of her revisions by applying them to drama dialogue. Here is a brief summary of some of the significant features of her model. First of all, in keeping with the structural emphasis of the Birmingham tradition, she places units of discourse in a hierarchical scale. This is the lower part of her scale illustrating how the exchange is situated above two smaller units of discourse:
The scale shows how a lower term is nested inside the structure of a higher term. In other words, exchanges comprise moves, while moves are in turn realised by acts. For their part, acts have no structure: if they had, then they simply wouldn’t be the bottom-most unit on the scale. Like all structurally orientated research into discourse, Burton’s model is built by analogy with grammatical structure. In fact, a good way of approaching her model is to see it as a parallel to the way words form structural relationships. If we imagine that moves are equivalent to words, then we can see that, just as words combine into collocations, so moves are able to join with other moves to form exchanges. Moreover, as we saw in Chapter 2, words can be sub-divided further into smaller particles called morphemes. In terms 144
DIALOGUE AND DRAMA
of discourse, the smaller particles which make up moves are acts. This parallel with word structure may be extended further by reintroducing root and bound morphemes (see section 2.3). Root morphemes are the main content-bearing component of a word and can often stand on their own; bound morphemes are clustered around the root and never stand on their own. Discourse acts can be similarly subdivided: some acts carry the main content of a move and are relatively freestanding whereas others can only operate in combination with another more substantial act. Here is a summary of the parallels that have been drawn between discourse and grammar:
discourse Exchange Move Act
grammar Collocation Word Morpheme
It is important to stress the centrality of the move as an interactive unit. Moves are often coincidental with a speaker’s turn in conversation and they also mark the transition points at which other speakers are drawn to respond and react. Of the seven principal moves identified by Burton, the three most important will suffice for our purposes. These are: Opening moves essential topic-carrying items which are recognisably ‘new’ in terms of the immediately preceding talk Supporting moves occurring after any other type of move and involving items that concur with the initiatory moves they are supporting Challenging moves functioning to hold up the progress of a topic or the introduction of a topic in some way As the discourse scale suggests, moves can combine with other moves to form exchanges. Here is a ‘question-answer’ exchange of the sort introduced in section 5.2, with the move types indicated beside each speaker’s contribution: 34
A: John! Have you got the time? (Opening move) B: Four o’clock. (Supporting move) 145
DIALOGUE AND DRAMA
This single conversational exchange is clearly made up of two moves: the opening move introduces a new topic and the supporting move maintains the discourse framework by providing the answer to the question posed. However, it is important to note that speaker A’s opening move comprises more than just a single act. It begins with the act summons (‘John!’) which is followed by the more substantial act question. (This progression would be mirrored in word grammar when a bound morpheme is tacked on to the front of a root morpheme.) The supporting move in (34) consists of just a single act, an answer. For a breakdown of an attested sequence of discourse, it is worth looking again at the travel agent exchange discussed in section 5.2: 5
A: er…you have to go to head office…please B: [nods and rises to go]
It was established earlier that this sequence is a ‘request-reaction’ exchange. It consists of two moves which are, as is commonly the case, coterminous with each interactant’s contribution. The manageress’s move consists of three discourse acts. The main content-bearing act is the one that is most central to the coherence of the exchange as a whole: ‘you have to go to head office’. However, studded around this act are two subordinate acts: ‘er’ and ‘please’. These ‘bound’ acts can be categorised as, respectively, a marker (a signal that a more substantial act is about to follow) and a prompt (a check to ensure that the request function is understood). In order to give some idea of what a structural analysis of discourse looks like, Table 5.1 is a breakdown of the exchange following the Birmingham framework. It will serve as an illustration of the type of visual layout favoured by this approach. As yet, we have not looked at any exchanges which contain challenging moves. As their label suggests, these moves hold up the onward progression of discourse in some way and they occur when an addressee withholds the anticipated second part of an exchange. For example, answering a question with another question, rather than providing the anticipated answer, clearly constitutes a challenging move in discourse. A good illustration of how challenging moves function in this way can be found in the communicative breakdown exhibited by example (1) which opened section 5.2. Where a response would have supported speaker A’s initial question, another 146
DIALOGUE AND DRAMA
TABLE 5.1 Requesting Exchange
question is offered in its place. This challenge creates a breach in the framework which ultimately leads to the collapse of the exchange as a coherent transfer of information between speakers. This is a structural breakdown of example (1), with an extra column for challenging moves added to the left-hand side of the diagram. Notice how the column for supporting moves is empty here (Table 5.2).
TABLE 5.2 Questioning Exchange
Although a key element in the overall organisation of discourse, discourse structure is not the only element. Structural models tend to explore the surface of discourse, rather than tease out the strategies which operate below the surface. So, while the material presented in this subsection will hopefully have given a flavour of this important branch of discourse analysis, a full-blown application of the Birmingham model to drama dialogue would not be possible in this context. For our purposes, the structural model is useful as a way of explaining how chunks of discourse are connected. It will not 147
DIALOGUE AND DRAMA
¶
however explain why people say the things they do. To answer that question we will need to look elsewhere.
5.3.2 Discourse strategy: maxims and relevance The philosopher H.Paul Grice published in 1975 an article which set the agenda for exploring the strategies of verbal interaction. His article had an especially important influence on the study of the way speakers use indirectness as a ploy in conversation. His entire thesis rests on a fundamental premise about interaction which he refers to as the cooperative principle (CP). The CP, which is mutually understood between interlocutors, is a basic assumption that speakers normally intend to accomplish purposeful and effective communication in conversation. Grice contends that if speakers wish to observe the CP faithfully they observe four conversational maxims. These are: The maxim of Quality Make your contribution one that is true and don’t say things that are false. The maxim of Quantity Make your contribution as informative as is required for the current purposes of the exchange—don’t say too much or too little. The maxim of Relation Say things that are relevant in the context. The maxim of Manner Be clear—avoid unnecessary obscurity and ambiguity. Readers may rightfully be sceptical about the way these highly normative maxims would be implemented in interaction. Conversation is simply not like this. If people behaved in this automaton-like way, conversation would become an over-literal, direct and unsophisticated routine. However, the fact that speakers do not always obey maxims in conversation is precisely the essence of Grice’s theory. He argues that many of the non-literal meanings that accrue from interaction are derived from deliberate departures from these maxims. Speakers may chose to ‘flout’ any of the maxims 148
DIALOGUE AND DRAMA
by blatantly departing from them. This places the CP in jeopardy because it leaves the listener to do some inferencing work to unravel the more covert meaning that the utterance conveys. These covert meanings are referred to as implicatures and they are typified by the sorts of utterances which are situated at the indirect end of our strategy continuum (see Figure 5.2 above). Here are some illustrations of how implicatures work. First of all, a speaker may choose to flout the maxim of quality by saying something that is manifestly untrue in the context: 35
A: You’ve failed your exam. B: Terrific!
As speaker B’s response is unlikely to be a true reflection of their state of mind, an implicature can be inferred whereby speaker B’s intended meaning is the opposite of what is literally asserted. This is another quality flout which I witnessed recently: 36 I like the way you let the kettle boil dry. Again, it is clear from the context that the intended meaning is the opposite of what the speaker actually asserts. Other common flouts of the quality maxim are metaphors, which are often strictly speaking ‘untrue’ descriptions of states of affairs. Grice’s own example here is ‘You are the cream in my coffee’ (1975:53), a remark which is literally untrue but which none the less conveys meaning at the level of implicature. Flouts of the quantity maxim normally involve deliberately saying too much or too little. Grice (1975:52) offers the following example of a reference written for a candidate who has applied for a post in philosophy: 37 Dear Sir, Mr. X’s command of English is excellent, and his attendance at tutorials has been regular, Yours, etc. Whereas (37) is clearly under-informative and would indicate tacit disapproval, the reply in the following exchange is clearly overinformative: 38 A: Did John go to the party last night? B: Yes, and don’t let anyone tell you anything different. 149
DIALOGUE AND DRAMA
The assumption behind the uptake of implicatures is that departures from maximal efficiency are strategically motivated and are not just aberrant or purposeless bits of discourse. Uptakes depend on the willingness of the addressee to do the inferencing work necessary, otherwise the utterance may be left looking like a disconnected remark. Here is an attempt at a quantity flout made by a television sports commentator.3 39 Well, there are entrances and there are entrances, and that was an entrance. Although this utterance is tautological, it is nevertheless meaningful at the level of implicature. However, when devoid of the original setting in which it occurred, it does look curious. There are two more types of implicature which are derived from flouts of the maxims of relation and manner. Flouts of the relation maxim often take the form of non-sequiturs which rely heavily for their uptake on the hearer assembling a meaningful connection between them and the previous utterance. The following exchange occurred during a recent telephone conversation between myself (speaker A) and a colleague (speaker B). We had been talking for a while about a rather tedious administrative matter in which we were both involved: 40 A: So should we do the transitivity stuff on day one? B: Mary’s just come into the room with a big plate of supper and switched on the telly. Speaker B’s utterance clearly flouts the maxim of relation. On the assumption that it was not just a disconnected remark and that he was still obeying the CP at some level, I inferred from this that it was a signal to terminate the conversation. It is in the nature of implicatures that they carry with them a certain amount of calculated risk. It would have been possible not to have accessed this implicature. So while indirectness is often useful as a politeness tactic (as in (40)), the likelihood of the success of the utterance diminishes in direct proportion to the degree of indirectness (see below, subsection 5.3.3). Finally, flouting the maxim of manner often involves saying something that is blatantly obscure in the context. Take the following example: 150
DIALOGUE AND DRAMA
41 Can you give him his B-O-T-T ‘cos I’m goin’ for a C-I-G. The speaker of this utterance (a smoker) is talking to another adult in the presence of an astute toddler. Explicit mention of both ‘bottle’ and ‘cigarette’ is disguised to avoid trouble. Grice’s model offers useful guidelines for accounting for the ways in which implications, suggestions and hints function in conversation. Perhaps its greatest benefit lies in its capacity to explain how ‘nonliteral’ meanings in language are passed between interlocutors. Although influential for these reasons, it has been criticised for being a rather informal model. Grice’s own illustrations are all carefully contrived to fit his analytic model; as a result there emerge from the theory too few explicit criteria to handle the complexity of naturally occurring language. Ironically, Grice is regarded as a philosopher of ‘natural language’, even though he never actually uses attested natural language. There is also some overlap between the maxims, especially quality and manner. This in turn throws up another issue about how many maxims there really are. Is four too few or too many? Grice himself remarks that there is an additional maxim, ‘Be polite’, which influences all the others (1975:47). We shall explore further the issue of politeness shortly. To round off this subsection, attention needs to be focused on another important extension to Grice’s theory. This is a model that expands the relation maxim, and although it is not the main analytic apparatus that will be used in section 5.4, it cannot be overlooked if our survey is to be properly informative. Dan Sperber and Deirdre Wilson have proposed a theory of communication which is based on the concept of ‘relevance’ (1986). Relevance theory, as it has come to be known, advances a cognitive model of interaction which takes as its point of departure Grice’s maxim of relevance. Like Grice, Sperber and Wilson are not so much interested in the formal or logical meanings of sentences but rather in the types of non-literal inferences and assumptions that speakers and hearers draw in communication. Sperber and Wilson contend that communication is successful not when hearers recognise the ‘linguistic meaning’ of an utterance but when hearers infer the speaker’s ‘meaning’ from it (1986:23). Sperber and Wilson describe communication in terms of two interrelated components: ostension, which is a speaker’s act of showing or making manifest through language; and inferencing, 151
DIALOGUE AND DRAMA
which is what hearers do when they attempt to decode acts of ostension in their search for meaning. Together these elements comprise ostensive-inferential communication. The next stage in the development of the model involves building in the underlying principles which shape and inform ostensiveinferential communication. Communication, Sperber and Wilson argue, is designed to bring about contextual effects. However, when they say contextual effects they do not mean adjustments to the physical context of interaction. They mean instead modifications in the store of knowledge between participants in interaction. The bringing about of a contextual effect may mean that a communicator’s cognitive domain is enriched by: the addition of new information, the strengthening of previously existing knowledge or even the contradiction of previously held assumptions. Whatever the precise nature of these contextual effects, they all warrant an adjustment of a communicator’s cognitive framework in some respect. Sperber and Wilson argue that an act of ostensive communication automatically communicates a presumption of relevance. To understand an utterance is to derive some contextual effect(s) from it; and to derive some contextual effect(s), you have to identify its relevance. The more overt the ostensive stimulus, then the less the processing effort needed to decode it. When this happens, the stimulus can be said to carry strong relevance. When the stimulus is less overt, more processing effort is required. Consequently, the ostensive signal is said to exhibit weak relevance. By now, readers may have quite justifiably formed the impression that relevance theory is not exactly an amenable, hands-on framework for the analysis of discourse. No doubt my attempt to explain the theory in four paragraphs will have strengthened this impression. It is time for an illustration. Imagine that you are driving along a country road in Ireland. You see, walking beside a river, a man who is carrying a fishing rod in one hand and a large salmon in the other. This ostensive stimulus interacts with other existing assumptions you have to lead you to the inference that he has caught the salmon in the river. The other existing assumptions that you bring to bear in your processing might be something like the following: salmon are fish which are found in rivers; fishing rods are implements for getting fish out of rivers and so on. As your inference enhances your store of knowledge with new information, it produces a contextual effect. In reaching this 152
DIALOGUE AND DRAMA
contextual effect, you have accessed only the most relevant stimuli. In other words, you search for relevance only as far as you need to and once relevance is proven, then inferencing is halted. For example, it is still possible that the inference that you have drawn here is not the right one. The man could be an actor, playing the part of a poacher, on his way to a film set; or perhaps a rather over-zealous partygoer on his way to the local fancy dress ball. However, in cost-benefit terms, the aim of interaction is to derive the greatest substantial contextual effect at the lowest processing cost and the most accessible conclusion in this scenario will be the one that is most relevant given the available ostensive stimuli. One of the most significant benefits of relevance theory is the way it handles the concept of ‘intention’ in communication. Crucially, it is not a speaker’s intention but a speaker’s intention to communicate which underpins communication. Making clear your intention to communicate is not the same thing as making clear your intention. From a speaker’s point of view, the object is to get your hearer to recognise your intention to communicate. From a hearer’s point of view, the intention of your interlocutor is simply not recoverable, so what you do is construct an assumption on the basis of the speaker’s behaviour (Sperber and Wilson 1986:65). Again we are coming up a wall of rather abstract theorising which can probably be unpacked properly only through an example. The nature of communicative intention (and indeed related aspects of relevance theory) can be demonstrated by examining an incident which recently ran in the newspapers in the Republic of Ireland.4 An eighty-year-old man was taken to court for allegedly driving while under the influence of drink. It transpired during the court proceedings that the arresting officer (a member of the ‘Gardai’, the Irish police force) had not breathalysed the suspect. The gard explained that he had not taken a breath sample on account of the following remark which was made to him by the suspect: 42 If an eighty-year-old man had pleurisy of the chest, would you make him blow into a breathalyser? The case against the man was eventually dropped, perhaps not surprisingly, for lack of sufficient evidence. However, it became clear during the proceedings that not only was the defendant hale and hearty, but that he had never had pleurisy of the chest in his life! His gambit in (42) would appear therefore to have been a complete 153
DIALOGUE AND DRAMA
success. The steps which link (42) to his ultimate acquittal can be presented in relevance theoretic terms: 1
In saying (42) the old man makes an ostensive stimulus.
2
The gard recognises the old man’s intention to communicate. But, crucially, the gard does not recognise the old man’s intention.
3
The gard seeks to establish the relevance of the assumptions in utterance (42).
4
The gard brings to the inferencing process other contextual assumptions (e.g. that the man is old; that pleurisy restricts your breathing etc.) which interact with the assumptions in what the man has said.
5
The outcome of this inferencing process is to derive a new assumption: that is, to produce a satisfactory contextual effect. The gard derives an inference which is most relevant for him in the context: namely, that his interlocutor has pleurisy. This contextual effect determines the gard’s subsequent actions.
6
Importantly, the gard draws no further inferences: he assumes his first hypothesis is best and that, in cost-benefit terms, he has derived maximum contextual effect from a minimum of processing effort.
Of course, the gard has recovered an inappropriate implicature, which ultimately leads to the man’s acquittal. He simply underestimates the communicative-cognitive skills of his suspect. What is more, the old man would appear to have been fully aware that the gard would underestimate his skills. Knowing that the gard has a false impression of his cognitive abilities and resources enables the old man to open up to his interlocutor an ostensibly neat inferencing path—a veritable garden path as it turns out! Although the above is no more than a superficial sketch of relevance theory, there should be enough material here to provide an important cognitive perspective on the concept of discourse strategy. The relevance theory model has, however, attracted criticism for being over-formal in approach, for not being developed on the basis of attested examples and for not paying enough attention to how the social characteristics of speakers and hearers determine strongly 154
DIALOGUE AND DRAMA
patterns of dialogue. Although a cognitive dimension in discourse analysis is invaluable, we will still want a framework that says something about the social and personal dimensions of verbal interaction. It is to such a model that we now turn.
¶
5.3.3 Discourse strategy: politeness phenomena It is perhaps fitting that the last major theoretical model to be introduced in this book is, in my opinion, one of the most elegant, powerful and convincing linguistic frameworks available to English language study. It is an elegant model in that it sets out a relatively simple rationale for explaining complex linguistic behaviour. It is powerful in that it develops a system of universal principles which underlie many different languages. And it is convincing in that it is based on strong empirical support derived from substantial crosscultural comparisons between languages. This model is Penelope Brown and Stephen Levinson’s theory of politeness phenomena (1978; 1987). Given its focus on discourse strategy, the politeness model quite naturally has points of overlap with the other pragmatic models introduced in subsection 5.3 2. In common with these other models, the politeness framework seeks to explain the linguistic strategies of verbal interaction. Where the models differ principally, however, is that while those in subsection 5.3.2 are primarily cognitive and philosophical in orientation, that of Brown and Levinson is social and anthropological. Furthermore, whereas Grice’s maxim of ‘relation’ provides the point of entry into relevance theory, it is his maxim ‘be polite’ which informs the Brown and Levinson model. The best place to begin our sketch of the Brown and Levinson model is by seeing how the authors themselves state their position. Here are the rudiments of their theory of politeness: certain precise parallels in language usage in many different languages can be shown to derive from certain assumptions about ‘face’—individuals’ self-esteem. We phrase this derivation in terms of three main strategies of politeness, ‘positive politeness’ (roughly, the expression of solidarity), ‘negative politeness’ (roughly, the expression of restraint) and ‘off-record (politeness)’ (roughly, the avoidance of unequivocal impositions), and claim 155
DIALOGUE AND DRAMA
that the uses of each are tied to social determinants, specifically the relationship between speaker and addressee and the potential offensiveness of the message content… In the case of linguistic pragmatics a great deal of the mismatch between what is ‘said’ and what is ‘implicated’ can be attributed to politeness. (1987:2) Although useful as a statement of their main thesis, their commentary does not in itself offer enough usable criteria for the analysis of politeness. Therefore, what follows is an account of the central tenets of their model. As is clear from the quotation, Brown and Levinson build their framework of politeness around the concept of face. Face is a fundamental aspect of the human psyche. It is primarily a sociological phenomenon which has two interrelated dimensions. These are: Positive face This is your wish to be liked by others; your desire to have your interests approved of; your desire to have what you like and want shared by others. In short, this is your positive self-image. Negative face This is your wish not to be coerced, ordered or forced into things; your desire not to be encroached upon or have your actions impeded by others. In short, this is your basic desire to be free from imposition. In interaction, there are a host of situations where someone’s face is likely to be threatened. Trying to get people to do things for you, trying to broach a subject that you know is touchy or trying to offer a bit of constructive advice: these are all circumstances which threaten some aspect of face. If carried through, these threats become encoded into verbal acts like requests, criticisms, offers, questions and complaints. Any such utterance, which constitutes a threat to face in some way, is called a Face Threatening Act (FTA). Doing FTAs is part and parcel of language usage. If it weren’t, verbal interaction would become extremely odd, as no one would ever complain, no one would ever ask a question and no one would ever ask anyone to do anything. Allowing then that FTAs are a necessary feature of interaction, the next question to be asked is what sort of strategies do people use to perform FTAs? To answer this, let us return for a moment to our discourse strategy schema 156
DIALOGUE AND DRAMA
(Figure 5.2, above) which provided a selection of different tactics for getting someone to open a window. The form ‘Goodness me, it’s hot in here!’ was categorised as maximally indirect, in that it contains no overt semantic link to what is requested and employs a grammatical mood that is tangential to its request function. Following the Brown and Levinson taxonomy, this strategy would be classified as offrecord. It is so named because it goes, so to speak, ‘off-the-record’ by avoiding any explicit mention of the goods or services requested. Basically, an off-record strategy involves dropping a hint. At the opposite extreme on our continuum lies the maximally direct utterance ‘Open the window.’ This strategy is clearly onrecord, in so far as it draws explicit attention to and contains a direct semantic link with the service requested. However, it is also extremely blunt. It contains no softeners or mitigating elements, offering no redress whatsoever to the hearer. Brown and Levinson describe this unmitigated, direct strategy as having been performed baldly, without redress. With the bald, non-redressive strategy the function of the utterance is clear, unambiguous and concise. In terms of Grice’s model (see subsection 5.3.2 above), it follows faithfully all four conversational maxims. In terms of Sperber and Wilson’s model, it exhibits strong relevance as it would require little processing effort to achieve a satisfactory contextual effect. To get a better understanding of how all of these pragmatic concepts are interconnected, it might be worth aligning Brown and Levinson’s two categories diagrammatically with those of Grice and Sperber and Wilson. Figure 5.5 is a modified version of Figure 5.2 with the key concepts of each pragmatic model brought into line with each other. It should be clear that the direct-indirect cline can be plotted against a parallel continuum which goes from impolite to polite. The strategic value of both the politeness strategies identified so far can be measured in cost-benefit terms. The impolite bald nonredressive strategy has an obvious payoff: it is direct and unambiguous, and its function is clear. However, it also carries a fair amount of risk: to be so blunt with someone indicates that you do not care about their face. In damaging the face of the interlocutor, the way is opened up for retaliatory face threats or at least a negative evaluation of your own behaviour. The polite off-record strategy also has payoffs and risks. Given that it involves dropping a hint, the immediate payoff gained from using it is that you are not seen to be coercing anyone or impeding them. But this very indirectness is 157
DIALOGUE AND DRAMA
FIGURE 5.5 Aligned pragmatic models itself a risk, as the function of the utterance is obscured. Consequently, there is a much greater likelihood that the strategy will fail as you may not be understood. And even if you are understood, the interlocutor can use your indirection as a ‘get out’ by pretending not to have understood the request function. We have yet to account for the third strategy which is situated in the middle of Figure 5.2 and which is represented by the utterance ‘Could you open the window?’ First of all, this FTA is on-record in that it contains clear lexical links (‘open’; ‘window’) to the specific service requested. Second, although it threatens negative face, it does so indirectly. This deliberate use of indirectness is a means of offering the hearer some redress. It is an example therefore of negative politeness, politeness which is specifically designed to preserve or protect the negative face of the interlocutor. It is motivated by a principle of avoidance: while doing a FTA, a speaker tries simultaneously to avoid offending that person. The strategy therefore calls for a dressing up of the speech act in linguistic features which will allay or mitigate the threat to face. There are a number of specific strategies which are available for doing negative politeness. Here are some of the most common: 1
158
Use indirectness: As we have seen from ‘Could you close the window?’, this strategy is built on a mismatch between grammatical form and discourse function. None the less, the function of the utterance will still be conventionally understood, despite this obliqueness. Indeed, only the most pedantic will ‘misunderstand’ a routinely polite indirection like ‘Can you pass the salt?’
DIALOGUE AND DRAMA
2
Use hedges: Hedges are the small particles of language which soften the impact of an FTA. They include ‘sort of, ‘er…’, ‘by any chance’ and ‘as it were’. Modal auxiliaries, such as ‘could’, ‘would’ and ‘might’, are also useful softeners. Hesitation and mumbling (often considered not to be real language at all) can also be used to hedge an FTA.
3
Be pessimistic: This is a curious though pervasive tactic of negative politeness. It basically involves expressing doubt about the chances of your FTA succeeding. It is encoded in expressions like: ‘I don’t suppose you could give us a lift’; ‘There’s no chance you could babysit next week.’
4
Minimise the imposition: This is suggesting that the FTA is not intrinsically serious and that you’re only committing a minor or negligible infringement to someone’s face. For example, an FTA like ‘Could I borrow a tiny little bit of paper’ downplays all aspects of the potential infringement: the amount of paper you request is minimal and you only want to borrow it, not take it for ever. If your interlocutor interpreted this literally, you would be less than satisfied with the outcome.
5
Indicate deference: This strategy, which is most obvious in overtly polite terms of address such as ‘sir’ or ‘madam’, accords social status to your interlocutor. However, deference can also be communicated if you humble yourself in some way. For example, you can downgrade your own abilities and possessions in remarks like ‘I’m not too good at this sort of thing…can you help?’; ‘Look, I’m really ashamed to have to ask you this—I feel such a pillock.’
6
Apologise: This is perhaps the most obvious candidate for inclusion which, in its most explicit form, simply solicits forgiveness from your interlocutor (‘Please forgive me’; ‘sorry’; ‘excuse me’). However, you can also apologise by referring explicitly to the impingement you’re making (‘I’d like to ask you a big, big favour’) or you can indicate your reluctance to have to bother someone (‘I really don’t want to intrude but…’; ‘I hate to have to ask you this…’). Alternatively, you can cite the overwhelming reasons that forced you to commit the FTA, a tactic normally interpreted as an ‘excuse’ (‘There just wasn’t enough time to do this’; ‘I’m sorry, but the bus was late…’). 159
DIALOGUE AND DRAMA
7
8
9
Impersonalise: This means removing references to yourself from the FTA. It is often a feature of official correspondence where an FTA is being committed (‘It would be desirable if…’). Plural forms can also be used for this function as they remove any individual responsibility for the FTA (‘We regret to inform you’). State the FTA as a general rule: This allows you to get yourself ‘off the hook’ by asserting that the conditions that led you to do the FTA are general ones and therefore not directly attributable to you. It is characterised by remarks like ‘Look, you’re not supposed to smoke in this compartment’ or ‘Textbooks must not be taken into the examination.’ Acknowledge the debt: This is simply a straightforward admission in a FTA that you are indebted to someone (‘I’ll never be able to repay you’; ‘I’ll be eternally grateful’).
It is worth reviewing briefly the travel agent exchange (example 5, above) in the light of this classification of negative politeness. Attention was given earlier to how the manageress exploits a cluster of linguistic strategies in a single discourse move. These strategies can now be accounted for systematically within the present model. The manageress employs a number of key negative politeness markers in the execution of her request to her interlocutor. She hedges the FTA by hesitating and using the particle ‘er’. She uses indirectness by employing an oblique grammatical form (a declarative) that is tangential to the discourse function. She impersonates the FTA by removing any reference to her own responsibility for producing it. She states the FTA as a general rule suggesting that the conditions incumbent on her addressee are external ones and not attributable to her. All this is supplemented with the word ‘please’, which is not only an obvious marker of politeness but also acts as a ‘force disambiguator’. It is often tagged on to an utterance to make sure that its request function will be understood. Up to now, we have considered only threats to negative face and the redressive strategies that are available to speakers when doing these FTAs. There remains the other side of our persona, positive face, which reflects our basic desire to be approved of by others. As with negative face, there are many linguistic acts which intrinsically damage positive face. Criticisms, reprimands, complaints and accusations are all quite obviously threats to a person’s desire to be liked and appreciated. Similarly, the mention of irreverent or taboo 160
DIALOGUE AND DRAMA
topics that are inappropriate in the context indicates that the speaker has little regard for the hearer’s wishes and feelings. On the other hand, there is a great deal of linguistic behaviour which is designed to attend to positive face, including compliments, expressions of solidarity and general displays of approval of each other’s personality. Linguistic behaviour which attends specifically to the hearer’s positive face is called positive politeness. Positive politeness is not avoidance-based like negative politeness, but extends more generally to unsolicited ‘polite’ behaviour. It is often used to claim common ground between speakers and can take the form of spontaneous compliments and displays of approval. Positive politeness is basically designed to ‘oil the wheels’ of interaction. Here are some of the more common strategies: 1
Compliment the hearer: This involves saying ‘nice things’ about your hearer. These compliments can be focused on their possessions (‘What a lovely vase’), their appearance (‘Nice tie!’) or their capacities (‘I don’t know how you manage to keep your kitchen so clean’). Compliments are often ostentatiously exaggerated (‘What an absolutely amazing meal that was!’).
2
Use in-group markers: This tactic is designed to close down the social distance between speakers and to proclaim common identity. Popular in-group markers include: ‘mate’, ‘pal’, ‘luv’, ‘guys’ and ‘dear’.
3
Claim common opinions and attitudes: This involves seeking agreement with your interlocutor and trying to avoid disagreement. This may be achieved by restricting conversational material to uncontroversial or safe topics. It can also be achieved by focusing on the particular aspects of a topic which are least likely to give offence. For example, if a friend wearing a garish, poorly designed jacket asks you if you like their new suit, you can always reply ‘Gosh, it fits ever so well.’
4
Hedge your opinions: This means not appearing too dogmatic in your views. The same hedging particles that have a negative politeness function can double up here as positive politeness markers (‘Well, it’s sort of a bit of a botched job, really’).
Positive politeness is often used as a supplement to negative politeness. Basically, it provides an excellent opportunity for ‘buttering somebody 161
DIALOGUE AND DRAMA
up’ before you try to get them to do something. In the first of the following two examples from Brown and Levinson (1987:103), the utterance is headed by a sequence of positive politeness employing the strategy ‘compliment the hearer’. However, the sting in the tail comes soon after in the form of a negative politeness FTA: 43 Goodness, you cut your hair. How lovely!…By the way, I came to borrow some flour. In the next example, an FTA is supplemented with an in-group marker (‘mate’) which Brown and Levinson (1987:108) argue is designed to close down the social distance between the interactants: 44 Here mate, I was keeping that seat for a friend of mine. I would like now to close this subsection by doing a little ‘fine tuning’ to the politeness model outlined above. There are a number of important principles governing the way speakers select strategies which need to be explained. One of these concerns the amount of mitigation which speakers decide to use in a FTA. Generally speaking, the number of mitigating elements is directly proportional to the amount of intended imposition to face. In other words, the greater the threat to face, the greater the number of politeness strategies. Here is a rather odd example where the number of negative politeness strategies used dramatically exceeds that warranted in the context: 45 I’m sorry to trouble you—I know it’s an awful imposition— but could I possibly impose on you and ask you if you could tell me what time it is? The oddity of (45) derives from the mismatch between the inordinate amount of politeness markers exhibited by the FTA and the actual imposition that it poses to the interlocutor.5 The relative social power of interlocutors is a strong determinant of choice of politeness strategy. In asymmetrical encounters, the speaker who holds high relative power need fear little threat to their own face from their interlocutor. Consequently, they can afford to use a less polite strategy such as the bald, non-redressive tactic. By contrast, the less powerful interactant will need to make use of the more polite strategies, such as negative politeness and off-record. The degree of 162
DIALOGUE AND DRAMA
social distance that holds between speakers is also crucial to choice of strategy. People who don’t know each other very well behave differently from people who do. Intimates in conversation can relax to some extent because the risks of damage to face are minimal. Nonintimates, where social distance is much greater, can take no such chances and so are under greater pressure to select politeness strategies appropriate to the context. There is a further consequence of this in that social roles are open to constant renegotiation. Politeness strategies can be used in a constitutive manner to close down or to expand the social distance between speakers. We perhaps all have acquaintances who are that little bit too ‘pally’ or who feign impoliteness through banter to claim a social proximity that isn’t really justified. We also probably all know people whose linguistic routines resist intimacy or whose politeness strategies are such that they sustain social distance. Politeness strategies do not just reflect existing social roles; they often go a very great way towards constructing new ones. Finally, there is an aspect of interaction which offers an excellent opportunity to study the confluence of the three Ss: setting, structure and strategy. This is the type of linguistic behaviour known as phatic communion (Laver 1975; 1981). In a structural sense, phatic communion occurs at the margins of interaction: it constitutes the beginnings and endings of conversations. It normally takes the form of the routine formulae of greeting and parting (‘Hi!’; ‘Good morning’; ‘Goodbye’), although significantly, it also consists of stereotypical remarks on topics like the weather. Perhaps nowhere is face more at risk than at this sensitive time, as speakers attempt to manage the transition from silence into conversation. Phatic communion therefore demands placatory language, not only to defuse the potential discomfort of silence but to ensure that only safe, non-threatening conversational topics are broached. This explains why, in phatic utterances, speakers so often search out emotionally uncontroversial themes—even if this reduces them to making supremely obvious or inanely trivial remarks about aspects of their immediate non-linguistic setting. Consider for instance typical phatic remarks like ‘Great view’, ‘About time these trains were cleaned’ or ‘Rotten weather again’. To put it more systematically, phatic communion draws heavily on the strategies of positive politeness, especially the substrategy which seeks to claim common opinions and attitudes with an interlocutor. It involves seeking agreement and solidarity with the listener, while at the same time trying to avoid face-damaging acts like disagreement, irreverence 163
DIALOGUE AND DRAMA
or taboo language. And the need for this interactive harmony is heightened in unequal power encounters or, as we have just seen, between socially distant interlocutors. Soon, we shall be seeing what happens when this and other conversational routines are not observed in social interaction.
5.4 Discourse analysis and drama dialogue The profitability of using drama texts for the study of conversation has been stressed earlier in this chapter. However, this is not to say that drama and naturally occurring conversation are identical types of communication. The most obvious difference between the two is that characters in plays are simply not real people in the way that interlocutors in conversation are. Another difference is that whereas naturally occurring conversation is straightforwardly ‘face to face’, in drama dialogue the channel of communication is more complex. This is because there are two communicative layers at work in drama discourse (Short 1989b: 149; Elam 1980:135). On the one hand, there is interaction within a play: this is the character-to-character dialogue which is displayed on stage or in the text. On the other, there is communication between the dramatist and audience or reader. At this higher level, the playwright becomes a kind of puppeteer who oversees and controls the displayed interaction. In a sense, the messages passed between characters within the play become, at the higher communicative level, messages about the play itself. Nevertheless, this is not to say that the character-to-character level of interaction cannot be subjected to the same analytic procedures as naturally occurring conversation. On the contrary, in the process of textual construction dramatists (like poets and novelists) have at their disposal all the codes and resources of language. Their creativity can therefore be fully appreciated only when charted against this substratum of naturally occurring language. Translated specifically to drama dialogue, this means that the analytic tools needed for observing what dramatists do must be able to handle the full panoply of verbal interaction. Aware, then, that it comprises this dual communicative structure, we will now move on to examine a short piece of drama dialogue. The extract in the ruled lines is the opening of Edward Albee’s onescene play The Zoo Story (I960).6 The play is based on a short encounter between two middle-aged men in New York’s Central Park. The men, Peter and Jerry, are strangers. Peter, who is described in the stage directions as being ‘neither handsome nor homely’, wears 164
DIALOGUE AND DRAMA
tweeds and smokes a pipe. Jerry is carelessly dressed and his appearance communicates ‘a great weariness’. Peter is seated on a bench reading a book as Gerry enters. Before reading the passage, it would be a good idea if readers were to review their responses to the two exercises built around examples (6) and (7) in section 5.2. JERRY: I’ve been to the zoo. [PETER doesn’t notice.] I said, I’ve been to the zoo. MISTER, I’VE BEEN TO THE ZOO! PETER: Hm?… What?… I’m sorry, were you talking to me? JERRY: I went to the zoo, and then I walked until I came here. Have I been walking north? PETER: [puzzled]: North? Why… I… I think so. Let me see. JERRY: [pointing past the audience]: Is that fifth Avenue? PETER: Why, yes; yes, it is. JERRY: And what is that cross street there; that one, to the right? PETER: That? Oh that’s Seventy-fourth Street. JERRY: And the zoo is around Sixty-fifth Street; so I’ve been walking north. PETER: [anxious to get back to his reading]: Yes, it would seem so. JERRY: Good old north. PETER: [lightly, by reflex]: Ha, ha. JERRY: [after a slight pause]: But not due north. PETER: I…well no, not due north; but, we…call it north. It’s northerly. JERRY: [watches as PETER, anxious to dismiss him, prepares his pipe]: Well, boy; you’re not going to get lung cancer, are you? PETER: [looks up, a little annoyed, then smiles]: No, sir. Not from this. JERRY: No, sir. What you’ll probably get is cancer of the mouth and then you’ll have to wear one of those things Freud wore after they took one whole side of his jaw away. What do they call those things? PETER: [uncomfortable]: A prosthesis? JERRY: The very thing! A prosthesis. You’re an educated man, aren’t you? Are you a doctor?
5
10
15
20
25
30 165
DIALOGUE AND DRAMA
PETER: Oh, no; no. I read about it somewhere: Time magazine, I think. [He turns to his book.] JERRY: Well, Time magazine isn’t for blockheads. PETER: No, I suppose not.
¶
166
I hope readers will forgive me for having rather mischievously duped them into participating in a stylistic experiment. What is more, it will probably be clear by now that the setting exercise and lexical set protocols developed around examples (6) and (7) were designed in anticipation of this precise piece of dialogue. The point of these protocols was to elicit intuitive predictions about appropriacy in discourse. These predictions can now be set against the stretch of dialogue above and any disparities between them explored. Utterance (6) turns out in fact to be Jerry’s opening conversational move. However, what is significant here is that in every elicitation I have carried out in workshops and seminars, this utterance is only ever aligned with setting 2, that of young child talking to an adult. Even though one of the remaining settings details specifically an encounter between two strangers at a park bench, this is never entertained as a viable setting for utterance (6). A similar point can be made about the appearance in the text of the lexical set containing items like ‘cancer’, ‘doctor’ and ‘prosthesis’. This set is only ever aligned with setting 3, the doctor-patient context, and is ruled out as positively ludicrous if applied to the other two scenarios. Moreover, the rather obvious nature of both parts of the exercise (commented upon by many participants!) makes all the more striking the discoursal twists that occur in the Albee passage. What you expect in discourse is a far cry indeed from what Albee provides, and the lack of fit between your predictions and the text itself means that there is a rich stylistic space which can be explored. Basically, and this is a point which will be resumed later, you have made a reading of the text before you have read the text. Any of the analytic models assembled in the previous two sections can be brought to bear on the character-to-character communication exhibited in the extract and each can be directed towards different aspects of the text’s organisation. Here are some suggestions as to how each model may be applied.
DIALOGUE AND DRAMA
In terms of its discourse structure, the dialogue is organised into two blocks of exchanges. These blocks are collections of exchanges which are broadly related by a common topic or theme. Such topically related blocks are often referred to as discourse transactions by discourse analysts. The first topic comprises Jerry’s attempt to discover the direction of the zoo and it extends from the opening line to line 19. The second, which occupies the remainder of the passage, is broadly concerned with exploring the clinical consequences of smoking. A key structural feature of the passage is that Jerry is responsible for initiating every single exchange: it is he who makes all the opening moves and controls the topic of discourse. Peter, by contrast, initiates nothing and is instead charged with the responsibility for supplying follow-ups to Jerry’s moves. The bulk of the exchanges are straightforwardly two-part structures. Some are of the ‘questioning’ type (see subsection 5.3.1) where the act question elicits the act response: JERRY: [pointing past the audience]: Is that Fifth Avenue? PETER: Why, yes; yes, it is.
7
Others are ‘stating’ exchanges where a statement anticipates the act acknowledgement: JERRY: Well, Time magazine isn’t for blockheads. PETER: No, I suppose not.
33
There are two marked exceptions to this pattern of discourse structure. The first is provided by the faltering opening to the conversation: JERRY: I’ve been to the zoo. [PETER doesn’t notice.] I said, I’ve been to the zoo. MISTER, I’VE BEEN TO THE ZOO! PETER: Hm?…What?…I’m sorry, were you talking to me? Here, Jerry’s statement does not receive the acknowledgement which it predicts. Instead, Peter follows it with a question which halts momentarily the progression of the dialogue. As it constitutes a ‘breach’ in the discourse framework, Peter’s question is therefore a challenging move. The second exception to the dominant pattern of discourse structure takes the form of a three-part exchange towards the end of the passage: 167
DIALOGUE AND DRAMA
JERRY: …you’ll have to wear one of those things Freud 25 wore after they took one whole side of his jaw away. What do they call those things? PETER: [uncomfortable]: A prosthesis? JERRY: The very thing! A prosthesis. Jerry’s second turn here functions as an evaluation of Peter’s response. This sort of three-part exchange structure was discussed earlier in the context of classroom interaction (see subsection 5.3.1). Outside the classroom context, however, its use tends to be restricted to situations such as quizzes (when speakers know the answer to the questions they are asking) or to situations where a speaker who has forgotten some fact asks their interlocutor about it. Whatever its precise nature here, the sense of cross-examination which it engenders clearly does little to put Peter at his ease. The onesidedness of the dialogue is further reinforced by the length of speakers’ conversational turns. While many of Peter’s responses are minimal, Jerry tends to hold the floor for much longer than his interlocutor. Moreover, Jerry repeats his material and, in his first utterance, even resorts to shouting it at his addressee. This is curious because repetition, when accompanied by an increase in loudness, is often used as an attention-getting device by children. This point is borne out by the consistency of predictions in the setting exercise, where responses consistently assign Jerry’s opening sequence (utterance 6) to child-adult interaction.7 Although it is common for children to use the strategy to get attention from adults, its use is plainly socially proscribed in spoken interaction between adults. Many structural features of the passage are clearly shaped by the discourse strategies used, and to this extent, a stylistic analysis would not be complete that did not pay attention to pragmatic features of the text. However, a pragmatic analysis does not mean trying to recover the psychological intentions of what are, after all, invented fictional characters. Rather, it involves trying to position yourself within the displayed interaction. What, for example, would you think if you were a witness to this interaction? What would your response be if someone spoke to you, say, in the manner of Jerry? Or again, what would your reaction be if a complete stranger you’d just met in a park kept trying to ascertain the direction in which they had been walking? In terms of Grice’s theory of maxims, Jerry’s over-persistence on this 168
DIALOGUE AND DRAMA
‘direction’ topic constitutes a flout of the maxim of quantity. He says more than is required for the current purposes of the exchange. His obsessive pursuit of this information is lexicalised through abundant repetition of the word ‘north’ as well as references to street names such as ‘Fifth Avenue’ and ‘Sixty-fifth Street’. However, as was established in subsection 5.3.2, the entire rationale behind flouting maxims is to generate implicatures. The question which is raised here is what sort of implicature could Peter possibly be expected to infer from all this? Again, what would we make of someone so obsessively concerned with whether they had been walking due north or just northerly? Jerry’s communicative goals are far from obvious, and from Peter’s perspective, the unease which this induces is clearly signposted in the stage directions. And no sooner does it seem that Jerry’s obsessive questioning is fizzling out, when he launches abruptly into the second block of exchanges without any clear resolution of the purpose of the first. Relevance theory offers a framework for assessing discourse strategies which is compatible in many respects with Grice’s theory of maxims and implicature. A guiding principle for exploring relevance in this text would be to ask: why does Jerry say the things he does? Take his opening sequence, for example. What would one make of a complete stranger announcing out of the blue that he had just been to the zoo? In relevance theoretic terms, Jerry provides an ostensive stimulus which signals his intention to communicate. The addressee’s task is to achieve a contextual effect by establishing the relevance of Jerry’s act of ostension. Now, the setting in which an interaction takes place would normally offer clues about which existing assumptions you could bring to bear in your search for the relevance of an utterance. This can be confirmed by looking back at the protocols for this utterance. There are strong pre-existing assumptions about the type of discourse that occurs in an encounter between two strangers. In fact, when asked to write exchanges for this setting, participants’ responses routinely converge on formulaic phrases like ‘Nice day’ or ‘Is this seat taken?’ However, none of these assumptions about context help establish the relevance of Jerry’s utterance. And as the relevance of the utterance is weak, so the effort needed to process it becomes all the greater. The same principle extends to Jerry’s questions about the direction in which he has been walking. To achieve a satisfactory contextual effect, an addressee must establish the relevance of what a speaker says, and in 169
DIALOGUE AND DRAMA
the case of Jerry’s utterances, this is far from easy. So while Jerry’s intention to communicate is clear, and the literal meaning of everything he says is clear, the intrigue lies in trying to establish the relevance of what he says. Moreover, in view of the context, there is something deeply disquieting about not knowing why an interlocutor singlemindedly pursues a topic whose relevance is so difficult to establish. It is interesting that McCarthy (1987:45) touches upon this aspect of the dialogue in his critique of the play, commenting on how Peter finds himself in the presence of an ‘unpredictable and consequently threatening’ interlocutor. The framework of politeness phenomena, which enables us to explore the intersection of setting, structure and strategy, forms a valuable extra dimension to the analysis. Jerry and Peter’s conversation is located at the initial phase of interaction—a period, it may be recalled, which carries heightened psychological risk. This is especially so given that the two interlocutors are strangers and that maximal social distance therefore obtains between them. As we observed in subsection 5.3.3, this phase of conversation calls for linguistic routines which gently massage the transition from silence into interaction, signalling that the speakers’ intentions are pacific and placatory. These routines of phatic communion are, in other words, designed specifically to attend to face: they comprise acts of face redress; they make use of safe topics of conversation; they avoid doing intrinsically face damaging acts. Moreover, the onus for initiating phatic communion often falls on the speaker who is moving in physical space into the territory of their interlocutor. Given these expectations about language and context, what occurs in the passage is markedly odd, and the protocols, again, are a good way of teasing out this feature of the interaction. An interactive gulf develops between the two interlocutors right from Jerry’s first move. What follows is an explanation of this conversational schism in terms of politeness phenomena. First of all, many of Jerry’s utterances constitute direct threats to the negative face of his interlocutor. In other words, his linguistic acts do not observe the addressee’s basic desire to be free from imposition. In doing these negative FT As, Jerry makes extensive use of the bald, on-record strategy. This is the strategy that is situated at the direct end of the strategy continuum. It offers no redress to the hearer, and, although clear and unambiguous, it is also blunt and rude. In the passage, the bald on-record strategy is encoded in direct questions containing no hedges or mitigation: 170
DIALOGUE AND DRAMA
Have I been walking north? Is that Fifth Avenue? What do they call those things?
(4–5) (7) (26–7)
The tactic extends to demands for information which is the personal preserve of the interlocutor: You’re an educated man, aren’t you? Are you a doctor? (29–30) Questions like this constitute a substantial intrusion into the psychological domain of an interlocutor. As a discourse strategy, it carries a great deal of risk. Consequently, its use is often confined to asymmetrical encounters where it is the speaker who wields the greater power and who therefore has less fear of retaliatory face threats from the interlocutor. However, symmetrical encounters, such as that between two strangers, exhibit no perceived power differential, so direct, unmitigated attacks on the hearer’s negative face are a very risky strategy indeed. Many of Jerry’s utterances also pose a threat to positive face. From his opening outburst onwards, his linguistic acts offer no redress to his interlocutor’s positive self-image and self esteem. He does not hedge his opinions, nor soften his statements and questions with any of the mitigating particles that are available for doing positive politeness. In view of the setting in which the interaction takes place, the term of address Jerry uses to Peter is markedly over-familiar (‘Well, boy!’). Amongst other things, this suggests that the social distance between the two speakers is small, when, remembering the context, it is manifestly high. Another striking feature of Jerry’s discourse is that he has no qualms whatsoever about using acts which intrinsically damage face. It was emphasised throughout the survey of positive politeness, that the use of taboo language or sensitive topics constitutes an inherently serious threat to a hearer’s positive face. For one thing, it signals that the speaker does not care about the hearer’s face-wants and feels no compunction to create interactive harmony. And there can hardly be a more alarming or emotionally controversial topic of discourse than discussing the likelihood of an interlocutor’s chances of contracting cancer. Even more disturbing is the fact that this topic is broached within the first few minutes of a conversation between two total strangers. It is notable that Albee is at pains to highlight the 171
DIALOGUE AND DRAMA
discomfort which Peter experiences under this onslaught, carefully monitoring his reactions through stage directions like ‘anxious’, ‘uncomfortable’ and ‘annoyed’. Peter’s discourse strategies are situated firmly at the opposite end of the politeness continuum from those employed by Jerry. His utterances bear all the hallmarks of the positive politeness strategies characteristic of phatic communion. He qualifies his opinions and is not dogmatic in his views. This is signalled in the text by extensive use of modality, encoded in hedging verbs like ‘think’, ‘seem’ and ‘suppose’, and is supplemented by the use of hesitation, indicated by three-dot sequences ‘…’. Here is a sample of Peter’s positive politeness strategies: Why…I…I think so. I…well no, not due north; it would seem so. No, I suppose not.
(6) (1 8) (13–14) (34)
Elsewhere, Peter apologises to Jerry with ‘I’m sorry’ (3), an act of negative politeness in which the speaker takes the blame for committing an FTA. He employs the term of address ‘No, sir’ (23) which suggests decorousness and which forms an interesting counterpoint to Jerry’s ‘Well, boy’. However, this strategy is immediately thrown back at him as Jerry’s next utterance (24) begins with an echoic and possibly sarcastic ‘No, sir.’ Peter’s strategies meet further interactive rebuffs when the sensitive topic of lung cancer is introduced (21–2). Albee’s stage directions are interesting here: Peter looks up annoyed, but then smiles. The directions may be trying to communicate that the initial irritation Peter experiences at the mention of a taboo subject subsides in the realisation that his interlocutor is, in one respect, offering reassurance. However, the respite is only momentary, as Jerry immediately manoeuvres the topic to cancer of the mouth—a condition from which his interlocutor will not be immune. Peter’s interactive acquiescence, signalled by his use of positive politeness strategies, is the antithesis of the direct, on-record strategies used by Jerry. This interactive asymmetry is sustained for most of the play, with the reticent Peter becoming ever more exasperated by his interlocutor’s doggedly blunt discourse strategies.8 Literary critics have had a lot to say about how the dialogue of The Zoo Story represents the discourse of the ‘absurd’. In common 172
DIALOGUE AND DRAMA
with much criticism of the absurd, particular emphasis is placed on the way Peter and Jerry’s interaction exhibits a sense of alienation, crisis and bewilderment, and on how it unveils the isolation of the individual in human society. What is remarkable about this criticism, however, is the absence of any coherent explanation of what Albee is actually doing with language in order to create these absurdist themes and motifs. The fact that the play is absurd is tacitly assumed—an assumption which only serves to bypass Albee’s particular skill with dialogue. What little critical attention there is on Albee’s technique tends unfortunately to be rather anodyne. What is more, it is difficult to conceptualise the sort of dialogue that there is in The Zoo Story on the basis of critical commentaries on the play. For instance, Bigsby (1969:24) sees the play in terms of a ‘breakdown of communication’ between the characters. Well, the very first exchange cited in this chapter (example 1) was a genuine communicative breakdown and the type of discourse structure it displays is far removed indeed from that of The Zoo Story. To describe the play as a communicative breakdown then is simply misleading. McCarthy (1987:16–20) talks of the isolation and alienation created through the ‘raw violence’ of Jerry’s language. This suggests that Jerry issues a string of unsophisticated verbal insults, triggering a proverbial ‘slagging match’ between the two characters. Again, this is not what happens: Jerry’s discourse strategies do much more than simply deliver wholesale abuse and invective. Critical comments such as these, which focus on character-to-character interaction, have an important influence on the sorts of interpretations that are made about the higher communicative level in the play. Impressionistic descriptions of dialogue within the text are sometimes aligned directly with the supposed mind set of the writer. McCarthy, for example, talks of Albee’s ‘aristocratic’ use of language which he says reflects ‘the situation of the writer’ (1987:8). Similarly, Evans (1977:199) makes the astonishing claim that Albee’s ability to dissect ‘sterile usages of speech’ can be directly attributed to his ‘apparent preoccupation with an inability to beget children…as an image of sterile futility’. The point is simply this. Behind many critical interpretations of the absurdism of the play is the taken-for-granted assumption that we all know what absurd dialogue looks like. However, the lack of agreement between the critics shows that we don’t. Discourse analysis can at least establish a framework for the systematic analysis of character-to-character interaction. This in turn lays the foundations 173
DIALOGUE AND DRAMA
for a coherent, rather than impressionistic, evaluation of Albee’s dramatic technique. It is clear from the stylistic analysis, for instance, that the dialogue Albee produces is not nonsensical or inane. On the contrary, his technique involves taking some basic, commonplace principles of naturally occurring conversation and knocking them off kilter. As we have seen, the impact created by exploiting these conversational routines in this manner is, well, dramatic. In this regard, the analysis of discourse has hopefully laid the foundations for a more rigorous appreciation of what Albee is doing with language. More importantly, given the current task in hand, it will have helped to highlight the routine and commonplace rituals of spoken interaction through observing what happens when they are distorted or subverted.
5.5 Summary Given its scope, and the extensive array of theories covered, it will be worth drawing together here the material covered over the course of this chapter. The provision of a short inventory of significant stylistic features might also help with practical applications of discourse models. The concept of the discourse ‘shunt’ (section 5.2) offers a useful way in to the analysis of conversational exchanges. It may be recalled that this involves collecting three or four ordinary exchanges, arranging them in sequence and then nudging the second member of each pair along one position. Amongst other things, this foregrounds the concepts of structure and appropriateness in discourse because, through the shunt, ordinary exchanges become extraordinary. In view of the material covered since the shunt was undertaken, readers may like to see if they can explain more systematically the peculiarity of the shunted examples, now that they are in possession of a fuller analytic toolkit. The setting exercise, also introduced in section 5.2, involves aligning a single utterance (and also, if possible, a lexical set) with a set of possible contextual scenarios. This can act as a valuable precursor to the stylistic exploration of drama dialogue. Assuming that the primary focus of study is linguistic and not literary, then it also makes good sense to concentrate on drama texts which are discoursally striking in some way. Texts which critics hail as ‘absurd’ 174
DIALOGUE AND DRAMA
are obvious choices, thereby making the work of writers like Samuel Beckett, Harold Pinter and Eugene Ionesco good candidates for selection. Another of Edward Albee’s plays, The American Dream, also lends itself well to this sort of applied discourse analysis. Here is an inventory of key aspects of discourse organisation. This inventory recasts the material covered over the course of the chapter into a checklist of discourse features which can be applied directly to dramatic dialogue.
¶
Discourse structure 1 Are discourse exchanges structured in a neat and orderly way, with opening moves followed by supporting moves? If not, are there any challenging moves which function to hold up or suspend the discourse framework? If there are challenging moves, do they form a pattern? Are they, for example, associated with a particular character in the text? 2
Which character initiates the conversational exchanges and which character responds? Is there a marked asymmetry in patterns of initiation, such as that identified between Peter and Jerry in The Zoo Story?
3
How long are respective speakers’ turns? Does one character hold the floor for longer than another? Does one character strike you especially as interactively reticent? Again, recall how the interaction between Peter and Jerry is differentiated in this way.
4
Are there any interruptions in the dialogue? Does the discourse of one character tend to cut across the discourse of another in any way?
Discourse strategies (maxims and relevance) 1 Do characters observe faithfully Grice’s four conversational maxims? Or, in the absence of ‘maximal efficiency’, do characters flout maxims? Are there flouts of any one maxim in particular? 2
Are the implicatures generated by maxim flouts accessible? Or, for example, do characters speak ‘in riddles’ all the time?
3
In terms of deriving satisfactory contextual effects, is it easy to establish the relevance of what characters say to each other? If not, does what characters say require substantial processing 175
DIALOGUE AND DRAMA
effort? And if so, what are the interactive consequences of these utterances which carry weak relevance? (Recall, for example, the problems that were involved in assigning relevance to what Jerry says.) Discourse strategies (politeness phenomena) 1 Are characters differentiated through their use of politeness strategies? 2 What are the perceived social relationships between characters? For instance, is a clear power differential or high social distance signalled in the text? How are these social relationships encoded into discourse strategies? Is any use made, for example, of the bald on-record strategy? If so, under what conditions? Is it attributable to a perceived power differential or to a suggested social intimacy between speakers? Are any strategies used which appear to conflict with assumed social roles? Is, say, the bald onrecord strategy used when the social distance between the speakers is too great to warrant it normally? 3 In Face-Threatening Acts, is the number of mitigating features used in the utterance in line with the actual imposition caused to the interlocutor? Or are they mismatched in some way?
¶
This inventory can be modified and reshaped to have more general relevance or can be adjusted to accommodate different types of spoken interaction. It has been the central aim of this chapter to illustrate how the stylistic analysis of drama dialogue can cast light on the organisation of naturally occurring conversation. Significantly, it is the sometimes absurd or surreal patterns exhibited by drama dialogue which offer the opportunity for exploring the commonplace and routine patterns that often pass unnoticed in everyday conversation.
Suggestions for further reading General discourse analysis Standard introductions to discourse analysis include Coulthard (1977); Stubbs (1983); Brown and Yule (1983); Levinson (1983). An 176
DIALOGUE AND DRAMA
important overview of the stylistic potential of discourse analysis is Short (1989b). The same writer also offers a useful later survey (1994). Herman (1996) is a substantial application of a range of techniques in discourse analysis to drama dialogue.
The Birmingham tradition of discourse analysis The main inspiration for much work in the Birmingham tradition is Sinclair and Coulthard (1975). Relevant work from the 1980s includes Burton (1982) and Coulthard and Montgomery (1981). The tradition has continued into the 1990s with the publication of Coulthard (1992). The chapter by Francis and Hunston (1992) in this collection is a ‘state of the art’ version of the structural model outlined in this chapter. In addition to Burton’s seminal study (1980), stylistic analyses which employ the structural discourse model and which are worth reading are Toolan (1987) and (1989).
Pragmatics Useful introductions to pragmatics are Grundy (1995) and Blakemore (1992). The second of these offers an extremely clear introduction to relevance theory (1992:24–38). Pilkington (1990) undertakes a relevance theoretic analysis of metaphor while Toolan (1992) offers a sustained critique of Sperber and Wilson’s theory. Jucker (1988) offers a survey of the connections between relevance theory and Brown and Levinson’s theory of politeness phenomena. Simpson (1989) explores absurdism and politeness in dramatic dialogue.
Additional reading Over the last few years, there has been a steady output of material on discourse and drama in the journal Language and Literature. Moreover, much of this work draws on models of pragmatics and discourse analysis that I have not been able to cover in this chapter. Particularly useful is: Calvo (1992) on Shakespeare’s As You Like It; 177
DIALOGUE AND DRAMA
Bennison (1993) on Tom Stoppard’s Professional Foul; Lowe (1994) on Arthur Miller’s The Crucible. Two other important references, in which developments in critical theory add an extra dimension to the analysis of drama dialogue, are Birch (1991) and Elam (1980:135–297).
178
Emphasising the value of literature in the study of language has been the theoretical cornerstone that supports this book. The focus has been not so much on the importance of language to literary study but rather on the importance of literature in language study. Given this emphasis, I hope that not too many readers have been alarmed that the sanctity of the pristine literary text has been violated, and that the works of canonical writers have been mauled, pummelled, dismantled and reassembled in language exercises and workshops. The primary aim has simply been to explain linguistic structure and function. This is not to suggest, however, that what we have done has no relevance to critical reading. On the contrary, all the chapters of the book have attempted to underscore the value that stylistic analysis has for critical interpretation. As far as a holistic and comprehensive survey of English language goes, this is, of course, nowhere near the full picture. All that a short introductory textbook like this can offer is a point of access into the study of the language—a full-blown account of all its structures and functions is simply not possible. Other key areas of language, into which literature would offer useful access,
After word
Afterword
179
AFTERWORD
are phonetics and phonology. These topics, touched upon in Chapter 2, are concerned with the sounds of language, and stylisticians have demonstrated the value of poetry, in particular, as a way in to this branch of language study. Another complex and substantial area of language organisation is grammar, which would probably merit an entire textbook to itself. Moreover, the coalescence of grammar, phonetics and phonology—along with vocabulary—offers access to the study of accent and dialect. Poetry, prose and drama written in regional or non-standard English not only offer valuable source material for this type of applied language study but also raise attendant issues to do with cultural, social and ethnic identities among speakers of English. Perhaps the greatest challenge in the study of English language is how to understand and explain its most basic structures and its most prosaic, day-to-day functions. It is hard to develop a conscious awareness of what we routinely take for granted about language. At the start of this book, I pointed out that one of the primary goals of stylistics—the linguistic analysis of literary discourse—is to get at the heart of everyday language. I hope that I have demonstrated to some degree, over the book, that the very act of explaining how and why something is linguistically striking goes a long way towards illuminating the ordinary and the commonplace in language.
180
1 Introduction p. 22 The technique of re-registration (introduced by Carter and Nash 1983:129) is the stock device used by ‘Merseybeat’ poet Adrian Henri in his poem ‘Bomb Commercials’. The poem is built on a veritable mosaic of different voices, registers and narrative styles which include ‘officialese’, advertising discourse and conversational narrative. Although not discussed here, this poem lends itself well to the sort of practical stylistic analysis undertaken in the course of the chapter. Dorothy Parker’s short poem, ‘One Perfect Rose’ also works well in this type of application. The poem is written largely in the form of a traditional lyrical ballad, focusing on the traditional romantic red rose, yet with the appearance in its final stages of words like ‘limousine’ it exhibits a marked switch in register.
Appendix
Teachers’ appendix
2 From shapes to words p. 30 I have found that this linguistic ‘conundrum’, which is taken from Smith and Wilson (1979:41), serves a useful 181
TEACHERS’ APPENDIX
teaching device. It works particularly effectively if presented as an isolated sentence on an overhead projector. After students have offered their opinions about its coherence, it can be disambiguated by adding the parallel constructions one at a time. p. 44 The study of the lexicon has probably been plagued more than any other area of linguistics by a proliferation of technical terms. Often linguists have good reason for introducing terms of their own and in many cases the new terms are used in slightly different senses from established terms. None of this helps the reader who approaches the discipline for the first time, however, and the irritation that might be felt when confronting this intimidating array is quite understandable. Nor does it help the teacher who needs to disseminate to pupils and students a workable and straightforward set of clearly defined terms. Wherever it has been feasible in this chapter, I have tried to restrict the number of terms used, and to use them as consistently as possible. For the benefit of those teachers who wish to develop this topic, here is a checklist of some of the over-lapping and alternative terminology that they are likely to encounter in the relevant literature: Term used in this chapter grammatical word content word root morpheme
derivational morpheme inflection(al morpheme)
Other available terms form word full word lexical word stem base operand lexical formative content morpheme derivational formative function morpheme inflectional formative
p. 54 A brief word to teachers who may wish to adapt this workshop plan to their own requirements. I have found over the years that Activity 1 is best undertaken by all the individuals making up a workshop or class. However, restrictions of time normally make it good practice to work in groups for the remaining activities, and to concentrate on a few rather than all of the short texts produced by the 182
TEACHERS’ APPENDIX
group as a whole. Any remaining ‘untransformed’ texts could be written up as exercises by participants after the workshop. Activity 3 can be carried out by placing a sample of the transpositions on OHP transparencies and then asking all of the groups for their comments. A useful subactivity is to swap the transpositions between groups, getting one group to comment on another’s efforts. The students who participated in the present version of the workshop were in the second year of their single honours degree programme in English language and literature at Liverpool University. All had taken, in the previous year, an introductory course in English language which comprised three principal components: basic grammatical description; phonetics and phonology; the history of the English language. The present session was a two-hour seminar in stylistics, the first part of which was a broad survey of the material covered in sections 2.2 and 2.3, the second a set of practical activities using the guidelines set out in section 2.4. Given that the time available was restricted, only one of the eight short texts elicited in this session was transposed. I have in general been able to use this set of activities with groups who have very different language requirements and linguistic abilities. This includes students of EFL. p. 56 One of the benefits of this activity is that it can be built up around a few stylistic devices and that, furthermore, these devices are pre-programmable. This ensures that the transposition exercise does not become a linguistic ‘free for all’ but instead works within a set of tightly defined stylistic boundaries. As the end product reveals, the results of the transposition can be quite striking. Moreover, the transformed text will be even more eye-catching if it is assembled using a typeface such as that provided by a computer screen. Although such facilities were not available during the present session, the final text still looked quite startling even though it was handwritten with a marker on a large whiteboard.
3 Words and meanings p. 75 The semantic entailment scale is one means of capturing diagrammatically the nature of semantic inclusion and compatibility between classes of words. The Venn diagram can also be used for 183
TEACHERS’ APPENDIX
this purpose, and teachers may prefer to adopt this type of schema. Here is part of the entailment scale for dog, presented here in the Venn diagram format:
p. 80 Collocational clashes in the rock and pop arena are certainly easy to find—indeed, the shelves of music shops are bedecked with countless examples designating groups, albums and song titles. However, the nature of the genre is such that they can drop in and out of vogue almost by the week. Here are more names of bands, which are at least relatively topical at the time of going to press. Students and pupils will have no trouble in enriching this list. Psychedelic Furs Spandau Ballet Soup Dragons That Petrol Emotion Wishbone Ash Prefab Sprout The New Fast Automatic Daffodils
Upholstered Eldorados Voice of the Beehive Aztec Camera Inspirai Carpets Velvet Underground Icicle Works
p. 87 Teachers may be interested to know why I chose this particular poem. I have found not only that it lends itself well to cloze test (there are many poems that do) but that it is short enough to be easily managed in the classroom context. Each line may be introduced in turn either by overhead transparency or by 184
TEACHERS’ APPENDIX
transcription on to a whiteboard or blackboard. In this way, the entire text can be built up cumulatively and predictions about the lexical gaps in each line can be sought in turn. The suggestions used in this chapter were collected informally in the course of a seminar with a small group of third-year students in 1993. The five students who comprised the group were all taking an option in stylistics as part of their joint honours degree at Queen’s University, Belfast. Although in their final year, none of the students had a particularly specialised knowledge of English language. Before then, all had taken courses in English literature as well as courses from the other disciplines in their respective joint degrees, such as history, French and politics. During the course of the seminar, suggestions were elicited and noted down. Students sometimes made more than one guess for a particular slot, in which case all responses were recorded. Occasionally, the same suggestion was made by different students, and where this impacts on the exercise, it is highlighted in the main part of the chapter. I have used this text in different pedagogical contexts and found that there is still a broad consensus of agreement about suggested entries to the cloze test. In the course of an introductory lecture on stylistics to year 1 students, I use it to elicit suggestions ‘from the floor’ as it were. Aside from the productive interactive function it has in this setting, it also provides a useful tool for highlighting basic semantic constraints on lexical combination. p. 94 The elicitation of core words (or ‘nuclear vocabulary’, as it is sometimes known) using multiple choice text has proved a popular device in EFL teaching. One reason for this is that core words often form commonly occurring idiomatic or metaphorical constructions in English. (See Stubbs (1986:99–115) and Carter (1987:33–46) for a fuller account.) p. 96 For copyright reasons, it has not been possible to reproduce in multiple choice format any part of Auden’s ‘In Memory of W.B. Yeats’, so unfortunately readers will need to track this poem down for themselves. Listed below are some four-term paradigms which can be applied to five suggested blanked out slots in the opening stanza of the poem. Each set contains the original term along with three other items which are compatible with the grammatical context of the deletion: 185
TEACHERS’ APPENDIX
Paradigmatic choices for stanza I: line: 1 passed away disappeared died expired
2 forests farmyards airports motorways
3 covered disfigured adorned fell upon
4 mercury sun rain chill
5 academics authorities sources instruments
Different sorts of issues can be developed through the exercise. For a start, attention can be focused upon the relationship of words to context. Set 1, with its euphemistic constructions, raises attendant issues concerning appropriateness and register. The relationship between lexis and grammar can also be explored. The line from which the item in list 2 has been deleted, for instance, contains a parallel structure of the order: ‘the xs were frozen, the ys [were] almost deserted’. This exerts a strong influence on the selection of the y element. And of course, the semantic gap between what is anticipated and what actually occurs in a text can throw light on the process of literary composition. Here are suggested multiple choice paradigms for the next three stanzas of the same poem, presented again in the format of multiple choice text. The deleted terms, as well as the selections offered for each entry, are only suggestions and need not be taken as a rigid formula. I have found that these lists work well in seminars, but they can easily be adapted to deal with all sorts of issues in lexical semantics.
Paradigms for Stanza II: line 2: streams wolves images yokels
3: fashionable dilapidated nice interestin g
5 personality identity death work
Paradigms for Stanza III: 2 clerics rumours doctors tranquillisers 186
4 squares depths windmills bits
5 suburbs bedroom hospital poetry
6 lost attacked upset became
TEACHERS’ APPENDIX
Paradigms for Stanza IV: 1 famous scattered forgotten unknown
3 lifestyle craft profession wood
4 honour conduct conscience practice
6 guts souls semantics land
It is worth adding that, prior to the cloze procedure and multiple choice activities, I had used ‘straight’ versions of the Roethke and Auden poems in seminars to little effect. ‘Dolour’ proved especially unsuccessful. It seemed that the full version of the poem closed off many of the potentially illuminating interpretations and readings that accrue from the progressive build-up of the text. Another reason for the relative effectiveness of the cloze test approach may have been that it commits students to making a semantic description of the text before they try to interpret it. In other words, it is their criterion of semantic appropriateness that helps establish the interpretative tramlines upon which a reading of the text can be developed. Moreover, the disparity between collocational predictions and collocational realisations throws into sharper relief the process of textual composition. p. 99 It is also worth detailing an article which explores the concept of oxymoron from a semantic perspective. Shen (1987) uses both componential analysis and the concept of the semantic scale to investigate this figure of speech. He argues that there are two main types of oxymoron: the direct (e.g. ‘the silence sounds’) and the indirect (e.g. ‘the silence whistles’). The difference is that in the first type the two opposites are at equivalent points on their respective semantic scales, whereas in the second type, one member of the pair is at a lower point than the other. The indirect type will therefore be a combination of silence and a hyponym of sound. This representation may help to clarify the principle:
187
TEACHERS’ APPENDIX
Shen argues that the indirect type is more common in poetry than the first and that it is altogether a more sophisticated technique requiring greater attention and comprehension.
4 Exploring narrative style p. 104 The implementation of these instructions in an actual workshop setting requires some modification. The jumbled sentences can be transposed on to A4-sized cards. Given the number of participants normally involved in this type of exercise, it is best that the sentences are enlarged as much as possible, but that sizeable gaps are left between each sentence so that the temptation to interpret contiguous sentences as meaningfully related is avoided. The cards can then be presented to students along with a worksheet containing instructions like those given. 1 On the accompanying card you will find a complete (untitled) short story by Ernest Hemingway. The story comprises just eleven sentences. However, these sentences have been ‘shuffled’ so that now they do not occur in the sequence of the original story. Each sentence has been assigned a letter. 2 Working in groups, try to reconstruct the story. You may cut the card into strips if you prefer, thereby enabling you to build up what you consider a ‘normal’ narrative sequence. This need not necessarily be an attempt to reassemble the original Hemingway text: just ‘re-shuffle’ the sentences so that they produce a relatively satisfactory narrative pattern. 3 Now make a note of the sequence you’ve selected (e.g. a b c d e…etc.) and try to justify your decisions. It may help to isolate any features of language which signal that a particular sentence should precede rather than follow another. Underline any such features. 4 Each group will be asked to state the sequence of their narrative and to account for the decisions they’ve taken. Each group’s selection will be discussed. 5 Yes…the original version will be produced! 188
TEACHERS’ APPENDIX
One of the advantages of cutting up the card is that each group, after having recorded their own reconstruction, is able to assemble the narratives proposed by the other groups. This ensures that all participants, when not presenting their own version, have before them the narratives of each of the other groups. p. 104 The participants in the present workshop were second- and third-year students following Liverpool University’s B.A. programme in English language and literature. These students had opted to take the stylistics course as part of the language ‘track’ of their degree; all were well motivated and their work in the first-year had revealed that all had demonstrable ability for language study. I had initially used the original, unaltered version of the Hemingway story as a sample exercise in comparable tutorials in previous years. The main intention had been to highlight a variety of aspects of narrative organisation, with special emphasis on isolating the cohesive devices exhibited by the text. This, in turn, was intended to lead to a comparative discussion of narrative structure by undertaking a contrastive analysis of this text with some ‘social stories’ told in everyday conversations. Despite the optimism of the tutor about the stylistic potential of the text, these tutorials were, frankly, unsuccessful. While many of the more stock responses regarding the ‘flatness’ of Hemingway’s prose style were produced, there were problems in explaining how this ‘flatness’ was created in the text. Furthermore, whilst many potentially illuminating comments on the ‘disinterested’ and ‘cynical’ nature of the narrative style were made, students encountered difficulties in identifying precisely the linguistic features which triggered these responses. As the unsatisfactoriness of the exercise was clearly not due to any limitations on the part of the participants, then it had to lie squarely in the design of the exercise itself. So finally, in a last ditch attempt to salvage the exercise as a whole, the present workshop activity was developed. It is worth adding that the activity was certainly engaged in with more enthusiasm than any previous tutorials based on this story. Group members co-operated well with each other, electing a spokesperson to summarise and justify the narrative pattern agreed upon by the group as a whole. One particularly noticeable development was the way in which each group defended its position in response to the queries and suggestions of other groups. 189
TEACHERS’ APPENDIX
p. 116 For reference purposes and for other teaching applications, here is a fuller breakdown of the Halliday and Hasan cohesion model, along with illustrations of each category. Cohesion—a general term for the ways in which different meaning relations in a text are combined inside sentences and across sentences. COHESIVE RELATIONS Anaphora: (a) Bill started to limp. It was strange to see him in pain. (b) Wash and core six cooking apples. Put them into a fireproof dish.
Cataphora: (a) He started to limp. It was strange to see Bill in pain. (b) This is how to get the best results: you let the berries dry in the sun, till all the moisture has gone out of them. ] COHESIVE TIES Definite reference: In addition to the definite article the, definite reference may be developed through the use of personal pronouns, deictics and implied relationships:
Three blind mice, three blind mice, See how they run! See how they run! Substitution:
Where a ‘token’ is substituted for a lexical word, My axe is too blunt—I must get a sharper one.
Ellipsis:
A type of ‘substitution by zero’, indicated by the symbol : Everton scored three goals and Liverpool Ø two Ø.
Linkage:
190
Realised by co-ordinating conjunctions and linking adverbials: and (additive); but (adversative); so (causal); then (temporal); if (conditional).
TEACHERS’ APPENDIX
Formal repetition: Repeated use of expression (morpheme, lexical item, proper noun, phrase etc.) which has already occurred in the text. Lexical cohesion/elegant variation: Incorporates use of general nouns, synonyms, superordinates and antonyms. Generally entails use of alternative expression (not a pronoun or substitution) for an item in the text: What shall I do with this crockery?—Leave the stuff there, (general noun) I began the ascent of the North face. The climb is really quite easy, (synonym) That’s a new Ferrari. I love Italian cars, (superordinate) The first effort was good. But the second effort was really bad. (antonym) p. 122 A word of advice is necessary on the ways in which this creative writing protocol can be aligned with the main workshop activity. Obviously, tasks 1 and 2 of the protocol cannot be carried out after the main workshop because foreknowledge of the Hemingway text interferes with the composition of an original story. There are a number of ways around this, one of which is to get two or three participants to peel off from the group as a whole before it embarks on the main activity. It makes sense to chose participants for this ‘splinter group’ who are interested in creative writing and/or who are confident enough to produce a story from the newspaper account. Their story, which should be written when the larger group is doing the central workshop, can then be compared with Hemingway’s version at the end of the session. If the small group’s story is transferred to an OHP transparency, this offers excellent opportunities for a three-way comparison, because all participants will now have access to the Hemingway story, the newspaper account and the ‘creation’ of their colleagues. It often make for a fitting end to a workshop to have students marvel at the differences between what their peers have created and what Hemingway has created using the same source material. Even more remarkable is the high literary 191
TEACHERS’ APPENDIX
rating they often give their own work in the face of competition from a major twentieth-century novelist. p. 123 The advantages of restricting the stories in length in this way are that it saves time, forces participants to focus only on details which they regard as essential and yields versions which are comparable in terms of length and economy to the Hemingway version. Although comparisons with the latter may be less rigorous linguistically and may not focus directly on specific features like cohesion and narrative structure, there is none the less some stylistic benefit. For one thing, it may be more appropriate with non-advanced learners of English, or in a context where language expertise is generally not as developed. On the other hand, it may be suitable in creative writing classes where such activities are common but where an exercise such as that proposed here would have the added advantage of providing an insight into Hemingway’s method. In both contexts, such a comparative analysis would provide a point of entry for linguistic and stylistic analysis.
5 Dialogue and drama p. 148 It should be pointed out that the schema presented in this subsection is a very much simplified version of the Birmingham structural framework. For example, the Birmingham model uses specialist terms like elicitation, directive and informative which correspond respectively to my more informal terms question, request and statement. Strictly speaking, the specialist terms are the ones that I should have used in my structural breakdowns. However, there are enough new technical terms in this chapter without having to assimilate twenty-two categories of act, seven types of move and about half a dozen exchange types! There are also two more units above the rank of exchange which are not represented here for the same reason. A related terminological problem is that the analytic categories are constantly under review and many of the original terms have been jettisoned or replaced by Coulthard’s (1992) collection of essays. Those new to the area might find it useful to see what the practitioners themselves have to say about discourse. Here is a useful statement on the structural ‘rationale’ behind the Birmingham model: 192
TEACHERS’ APPENDIX
The advantages of working within a structural description of discourse are clear. The distinctive feature of a structural description is that the elements in the description and their possible combinations must be rigorously defined. This means that descriptions which are based on the same structural criteria are directly comparable. It is possible to reveal similarities and differences between different discourses and different genres of discourse once these have been subjected to the same structural analysis. (Willis 1992:112) p. 155 Teachers might be interested to see a selection of the key, original passages through which maxim theory and relevance theory were originally propounded. Here, first of all, is Grice outlining the principal tenets of his co-operative principle: Our talk exchanges do not normally consist of a succession of disconnected remarks, and would not be rational if they did. They are characteristically, to some degree at least, cooperative efforts; and each participant recognises in them, to some extent, a common purpose or set of purposes, or at least a mutually accepted direction. This purpose or direction may be fixed from the start (e.g. by an initial proposal of a question for discussion), or it may evolve during the exchange; it may be fairly definite, or it may be so indefinite as to leave very considerable latitude to the participants (as in casual conversation). But at each stage, SOME possible conversational moves would be excluded as conversationally unsuitable. We might then formulate a rough general principle which participants will be expected (ceteris paribus) to observe, namely: Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged. One might label this the COOPERATIVE PRINCIPLE. (1975:45) Here now is an important section from Sperber and Wilson (1986) where they sketch in brief their theory of relevance: an act of ostensive communication automatically communicates a presumption of relevance… The relevance of a stimulus is 193
TEACHERS’ APPENDIX
determined by two factors: the effort needed to process it optimally, and the cognitive effects this optimal processing achieves… It is in the interest of the addressee that the communicator should choose the most relevant stimulus… that is the one that will call for the least processing effort. Here the interests of communicator and addressee coincide. Unless the communicator is merely pretending to communicate, it is in her interest to be understood, and therefore to make it as easy as possible for the addressee to understand her. An addressee who doubts that the communicator has chosen the most relevant stimulus compatible with her communicative and informative intentions—a hearer, say, who believes that he is being addressed with deliberate and unnecessary obscurity—might doubt that genuine communication was intended, and might justifiably refuse to make the processing effort required. All this is mutually manifest; it is therefore mutually manifest that the communicator has chosen the most relevant stimulus capable of fulfilling her intentions. (156–7; original emphasis) p. 166 I have used this passage from The Zoo Story on my third-year stylistics seminars at Queen’s University over the last few years. Any responses to the protocolled activities of section 5.2 were elicited in the course of these seminars. p. 175 Regarding the choice of material for stylistic analysis, the Albee text was used in this chapter because it exemplifies many important aspects of discourse. Although any text will suffice in principle, it often makes good sense to use a text that is considered odd or peculiar in some way. Once chosen, it is worth looking out for a number of linguistic features which not only shed light on the principles of strategy and structure but also reveal interesting devices in the text itself. p. 176 Perhaps the most daunting aspect of discourse stylistics, especially from a pedagogical point of view, is how to impart and develop knowledge about discourse theory. Although the shunting and setting exercises are a good way of ‘sugaring the pill’, there is no doubt that the teaching of linguistic theory is a formidable task. Yet there needs to be at least some theoretical framework if any purchase 194
TEACHERS’ APPENDIX
is to be gained on the basic mechanics of social interaction. This necessity prompted the development of the structure and strategy axes in section 5.2 of the chapter. The axes were designed as an overarching framework for organising and collecting a diverse body of research. Assuming that there will already be some familiarity with the choice and chain axes from Chapter 3, transposition to the concepts of structure and strategy in discourse should be conceptually more straightforward. While the axes allow the full gamut of discourse theories to be brought together, the question remains as to how to implement theory in the context of teaching and learning. I find that some pragmatic theories can be illustrated by role play activities. For example, the politeness continuum represented by Figure 5.5 can be prefaced by asking participants in workshops to perform imaginary tasks in different settings. One such task is to think of two ways of getting someone to open a window: the first in the most direct manner possible and the second by just dropping a hint. The task can be embellished by suggesting different settings for the request. One setting could be an interaction between a sergeant major and a raw recruit and another an encounter between two strangers on a train. Then, following one of the basic principles of stylistic analysis, create a mismatch by switching the two utterances to the wrong settings. Assess the results of the mismatch, and in particular try to explain any oddity that accrues. In this type of activity, not only are constraints on discourse appropriateness thrown into relief but the opposite poles of the politeness continuum are clearly highlighted.
195
The more elaborate definitions in this glossary are given for terms which are used across different chapters of the book but which do not receive extensive treatment in any particular place. Shorter ‘refresher’ definitions are offered for those terms which are dealt with in detail in specific sections of the book. Accent Accents are varieties of language distinguished from each other by pronunciation. There is a tremendous proliferation of accents of English, not only in terms of international variation (i.e. Australian, Canadian, South African accents), but also within national boundaries. British English, for instance, contains numerous easily recognisable and markedly different accents such as Scouse, Glaswegian and Cockney. The most ‘prestigious’ accent of English, known as Received Pronunciation (RP), has its historical origins in the upper-class speech of London and the surrounding area, although it is now no longer the accent of any particular region. In spite of its high social status, it is estimated that under three per cent of the population of the United Kingdom speak RP. See also dialect.
Glossar y
Glossary
197
GLOSSARY
Aspect Aspect is a feature of the verb which specifies whether the verbal action has been completed or is in progress; or whether it has had limited or extended duration. Two important types of aspect in English are: Continuous aspect
formed by using the ‘ing-form’ of the main verb and a form of the verb ‘to be’ Perfective aspect
formed by using the the past participle of the main verb and a form of the verb ‘to have’ At present, he is living in Sheffield, (present continuous) That year, he was living in Sheffield, (past continuous) For generations, Nepal has produced brilliant mountaineers, (present perfect) For generations, Sparta had produced famous warriors, (past perfect) See also tense.
Coherence A term used in the study of discourse which is often contrasted with cohesion. Whereas cohesion is chiefly a semantic phenomenon, coherence is more psychological in emphasis in that it refers to the processing and inferencing strategies which people use to interpret texts. A coherent text is one judged to make sense on a general level, rather than on the basis of its cohesive features. For example, ‘I am sorry I’m late. The traffic was bad’ exhibits coherence without having any cohesion. Alternatively, one would be hard put to retrieve coherence from the following sequence even though it contains obvious cohesive links: New York is a city. Cites are made from buildings. Buildings are often tall. ‘Tall’ doesn’t mean ‘short’. See discourse, cohesion.
Cohesion A term describing the ways in which elements in a text are bound together. Cohesive ties are the semantic links that operate within and across sentences. See also coherence. 198
GLOSSARY
Connective Connectives are words which link grammatical units together. When the units are of equal grammatical status, then the connectives are referred to as co-ordinating conjunctions: He ate his supper and he went to bed. He ate his supper so he went to bed. When the two grammatical units are not of equal status—that is, when only one of them can stand on its own—the connectives used to link them are referred to as subordinating conjunctions:
When he had eaten his supper, he went to bed. If he has eaten his supper, he must have gone to bed. Other subordinating conjunctions are ‘although’, ‘because’ and ‘since’.
Connectivity See connective. Declarative A declarative is an especially common type of grammatical mood in which the subject of a sentence precedes its verb. Declaratives are typically, though not necessarily, used to make statements: Mary shut the door. The cat is in the garden.
Dialect Dialects are varieties of language which are distinguished from one another by differences in grammar and lexis. Unlike a register, a dialect will often reveal important information about the social and regional characteristics of its speakers. The most widely distributed dialect of English, known as Standard English, is the variety which is taught to non-native speakers and is normally used in writing: I haven’t seen any of those kids. (Standard English dialect) I ain’t seen none of them children. (Non-standard English dialect) See grammar, lexis.
Discourse A key term in language study which encompasses a host of different approaches, theories and analytic models. In its broadest sense, discourse refers to the way language is organised above the level of the sentence. The units of analysis in the study of discourse are text and utterance. See also pragmatics. 199
GLOSSARY
Exclamative A type of grammatical mood which is rather imprecisely defined. Exclamatives normally begin with words like ‘what’ and ‘how’, contain some rearrangement of the subject-verb-complement sequence, and in their written forms conclude with an exclamation mark: What a lovely hat! How tall Sinead is! What a great day that was!
Existential sentence Existential sentences simply assert that something exists. They are identified by a combination of the dummy subject ‘there’ (which is simply a formal grammatical requirement in English) and some form of the verb ‘to be’: There was a phone call. There is no need for this. There are many good textbooks around.
Explicative A narrative device used to explain why certain events and actions happened in a story. Explicatives often present information about which the narrator had no knowledge at the time the story happened, but which has come to light since. The device is often marked by ‘that…’ clauses and past perfect (see aspect) verb forms: and when we realised that we were really out of danger then we found out that we had been so tense that our feet had been up against the panel…
Foregrounding Foregrounding is a term developed by the Russian Formalists and Prague school structuralists which refers to any linguistic feature (at any level of language) whose prominence in a text is motivated for literary purposes. Foregrounded items are ‘defamiliarised’ in that they deviate from the established background patterns of language in a text. Formalist Formalism is a critical-interpretative method developed chiefly by Russian theorists and poeticians in the earlier part of the twentieth century. Epitomised by the work of Victor Shklovsky, formalist interpretations are predicated mainly on the formal linguistic features of a literary work and not on its contexts of production and reception. Special attention is given to the manner by which certain elements in a text are 200
GLOSSARY
foregrounded. Attracting criticism from Marxist critics in particular, formalists have been accused, perhaps unfairly, of being overly text-immanent in their approach. See foregrounding, text-immanent. Grammar A key term in language study. The grammar of a language is the set of core rules which combine phrases, clauses and sentences into well-formed structures. In the academic study of English language, grammatical differences between different dialects of English are not considered ‘good’ or ‘bad’, but rather as systematic, rule-governed differences that define those dialects: I’ve my dinner eaten. (Irish-English grammar) I’ve eaten my dinner. (Standard English grammar) I’m just after seeing an accident. (Irish-English grammar) I have just seen an accident. (Standard English grammar) See also dialect.
Graphology Graphology refers to the visual medium of language. It describes the general resources of language’s written system, including punctuation, spelling, typography, alphabet and paragraph structure, but it can also be extended to incorporate any significant pictorial and iconic devices which supplement this medium. Intensifier An intensifier is an adverb which is used to describe a gradable adjective: extremely hot very wide slightly obscure rather poor more cold
Some adjectives like ‘military’ or ‘deliberate’, which in certain grammatical constructions are not gradable, cannot take intensifies: a slightly military junta an extremely deliberate liar
Lexis Simply, words and vocabulary. A study of lexis may involve looking at the structures of words, the meanings of words, the 201
GLOSSARY
history and origins of words or the geographical distribution of words. Modal auxiliary Modal (or secondary) auxiliaries are a closed set of words which precede the main verb in a verb phrase. They are: ‘can’, ‘could’, ‘must’, ‘may’, ‘might’, ‘will’, ‘would’, ‘shall’, ‘should’ and ‘ought to’. Modals have several functions. One is to convey modality; another is to signal future time reference in the absence of a future tense in English. Modality Modality refers to a speaker’s or writer’s attitude towards, or opinion about, the truth of the sentences they utter. It also extends to their opinion about the degree of obligation which attaches to what they say. There are various ways of conveying modality in language: You must/could/might be right, (modal auxiliaries) I believe/think/assume you are right, (modal main verbs) It is certain/probable/doubtful that you are right, (modal adjectives) You are certainly/probably/possibly right, (modal adverbs)
Modals See modal auxiliary. Mood A grammatical term which describes different types of sentence structures. The criterion often used to identify mood is based on the relation between the subject and the first part of the verbal element in a sentence: Martha Is going to town (declarative) Is Martha going to town? (interrogative)
See also grammar. Narrative A language-based definition of narrative is that it is a method of recapitulating past experience by matching a sequence of clauses to the sequence of events which (it is implied) actually occurred. A minimal narrative is a sequence of two clauses which are temporally ordered: that is, a change in their order will affect the interpretation of the time sequence of the original story: John drank two whiskies (clause 1) and he fell into the river (clause 2).
Natural narrative See narrative. 202
GLOSSARY
Past continuous A past continuous verb phrase is formed by placing the verb ‘to be’ in its past tense form in front of the ‘ing form’ of the main verb. The past continuous is often used in storytelling to describe events that were ongoing when another, more central event occurred: I was just sitting at home, when… We were living in Nottingham that year, when…
See also aspect. Phonetics The study of the articulatory and acoustic properties of speech sounds. Phonetic analysis can be subdivided into three principal areas of investigation: the production of speech sounds (articulatory phonetics); the transmission of speech sounds (acoustic phonetics); the perception of speech sounds (auditory phonetics). See also phonology. Phonology The study of the abstract relationship between meaning and sound in language. Phonologists examine intonation patterns in spoken discourse as well as the individual sound segments which make up words. The unit of analysis in phonology is the phoneme. See also phonetics. Pragmatics Pragmatics is that branch of language study which seeks to explain the meaning of utterances in contexts of use. Studies in pragmatics often focus on strategies of social interaction such as indirectness, hints and politeness. See discourse, semantics. Psycholinguistic(s) Psycholinguistics is the scientific study of language and the mind. Psycholinguists examine, among other things, the development of language in young children as well as the general mental processes involved in human communication. Register A variety of language associated with a specific context of use. Sermons, lectures and recipes are all types of register. A register is defined according to the use to which language is being put, and not through the social and regional characteristics of its user. It is not to be confused with dialect, therefore: I have not seen any of those children. (Standard English dialect, formal register) I haven’t seen any of those bloody kids. (Standard English dialect; informal register).
See also dialect. 203
GLOSSARY
Semantics Put most simply, the study of meaning. A more precise, and therefore more complex, definition would be that semantics is concerned with the abstract meaning potential of units of language of varying sizes. The unit of analysis in semantics is the seme. Unlike pragmatics, semantics is more concerned with the core meaning of linguistic units irrespective of their various contexts of use. Speech act Coined by the philosopher J.L.Austin, the term speech act refers to what is done when something is said (e.g. stating, warning, declaring, threatening etc.). An utterance may therefore be regarded as performing a specific function as well as having a semantic value. Austin argues that in performing a speech act, a speaker simultaneously performs the following three kinds of ‘sub-act’: • a locutionary act—the act of saying a meaningful, grammatical sentence in a language • an illocutionary act—the act performed in saying something (a command, request, warning, question and so on) • a perlocutionary act—the act performed by saying something (frightening, convincing, persuading and so on) See also pragmatics.
Tense Tense is a grammatical term which describes the correspondence between the form of a verb and our concept of time. The basic tense distinction in English is between present and past: I walk (simple present: formed by using the base form of the verb) I walked (simple past: formed by using the ‘-ed’ form of the verb)
Unlike French or Latin, there is no specific future tense in English corresponding to the present and past, in so far as there is no grammatical operation which is performed on the main verb to yield such a tense. Consequently, future time is expressed in a variety of ways in English: On Tuesday she leaves for Paris, (simple present) She will leave for Paris on Tuesday, (modal auxiliary and base form of main verb) On Tuesday she is leaving for Paris, (present continuous)
See also aspect. 204
GLOSSARY
Text-immanent A text-immanent interpretation assumes that textual meaning resides entirely within the formal linguistic properties of a text and is not influenced by dynamics of communication, the context of textual production or by pragmatic aspects of communication. See Formalist.
205
1 Introduction 1 2
3
4 5
6
7
Notes
Notes
This excerpt is taken from: Abrams, M.H. (1981) A Glossary of Literary Terms (4th edition) New York: Holt Rinehart & Winston, p. 192. When stylistics does become genuinely ‘scientific’ it is transformed into something else, known as the ‘empirical study of literature’ or ‘literary science’. This is not the method that will be used in the book, although if readers want to acquaint themselves with this type of work, then Steen (1994) is a representative sample of research in the area. The journal Poetics is also heavily orientated towards the empirical study of literature. A useful critical review of the Leavisite movement can be found in Eagleton (1983). I have also written a piece on this tradition of criticism (Simpson 1990), which looks into the linguistic strategies that its practitioners use to get their points across. This excerpt is part of a recipe from: Jaffrey, Madhur (1982) Indian Cookery London: BBC Publications, p. 163. This extract is from pages 11–12 of the Pan Books edition of The Little Drummer Girl published in 1984. (The novel was originally published in 1983 by Hodder & Stoughton Ltd.) Readers seeking a literary-critical introduction to le Carré’s fiction might like to consult Barley (1986), which contains a chapter on The Little Drummer Girl. Simpson (1988) offers a stylistic analysis of the opening of this novel. This passage is taken from: Pevsner, Nikolaus (1953) The Buildings of England: Derbyshire Harmondsworth: Penguin Books Ltd. I am grateful to Professor Walter Nash for bringing this text to my attention. Thanks are due to the students of Liverpool University’s 207
NOTES Department of English Language and Literature for responding to this questionnaire. Their responses were collected over a morning in the Department. The questionnaire was distributed between classes, not during any one class, which is why it was possible to obtain responses from both undergraduates and postgraduates. The exercise as a whole is not intended to be a rigorous, quantitative survey and obviously more systematic experimentation would be required in order to produce a statistically significant set of results. In the context of this chapter, it simply offers some valuable extra perspectives on the particular text under scrutiny.
2 From shapes to words 1
2 3 4 5 6 7
There is, unfortunately, an anomaly in the graphological system in that there is no allograph symbol corresponding to the square brackets, ‘[ ]’, of phonetics. Realisations of graphemes are encoded with ‘’, the same symbol, in other words, as the graphemes themselves. This extract from Samuel Beckett’s Murphy is taken from the Picador edition (1973), p. 10. This extract is from page 1 of the Independent on Sunday, 20 July 1992. This quotation from Chaucer is taken from line 188 of The Miller’s Tale (Bantam Books edition, edited by Kent Hieatt and Constance Hieatt, 1964). This quotation from Alice in Wonderland is taken from the Signet edition (1960), p. 91. This quotation from Ulysses is from the Penguin edition (1980), p. 44. This poem can now be found in: e e cummings (1968), Complete Poems: Volume 2 1936–1962, London: MacGibbon & Kee, p. 727.
3 Words and meanings 1 This quotation is taken from Lewis Carroll’s Alice through the Looking Glass, [1871], Signet edition (1960) p. 186. 2 Carter (1987:58–65) reserves the term fixed expressions for constructions like euphemisms, proverbs and stock phrases. The principle common to all these expressions is that (1) their component words and structures are fixed and (2) the meaning of the whole is not obvious from the meaning of the constituent parts. Here are some examples: The little boy’s room’; ‘A stitch in time saves nine’; ‘When all is said and done’; ‘As sober as a judge’; ‘A watched kettle never boils.’ 3 By ‘traditional dictionaries’, I have in mind, for example, the various editions of the Chambers Dictionary or the Oxford English Dictionary. The accuracy of word-definitions in such dictionaries is often largely dependent on the intuition of the lexicographers who compile the dictionary. However, recent, corpusbased dictionaries have broken with this tradition. The Collins COBUILD dictionary, for example, lists words as they occur in attested (‘real’) constructions which have been transcribed from a naturally occurring corpus of language. 4 Another problem with synonymy is that the privileges of occurrence of each item in a pair are not always identical. For instance, deep and profound share 208
NOTES
5
6 7
8
9 10 11
12
13 14
the same sense, but while someone’s thoughts may be either, a river can only be described with the first term. Similarly, the term film is somewhat broader in scope than movie: we may speak of a camera running out of film, but not out of movie. This example is one of the many amusing newspaper headlines studied in an article by Lindemann (1990). Lindemann challenges the popular view that headlines are telegrammatic and informative, arguing instead that they often ‘disinform’ through riddles, puns and cryptic allusions. Both examples are taken from Alumni summer 1991 issue, p. 21. The idea of lexical specificity and the pragmatic implications of over- and under-specification was originally developed in detail in Cruse (1977). See also Cruse (1986) for a book-length treatment of this and other topics in lexical semantics. Interactive ‘put-downs’ like this are not however the sole function of marked under-specification. Cruse (1977) details a series of environments where they can even function as markers of in-group expertise. At the races, pundits may remark That is a fine animal’ which produces less specificity than the contextually neutral term horse. Similarly, a jeweller may exclaim These are fine stones’ (as opposed to ‘diamonds’ or ‘emeralds’), while a flight attendant may urge passengers to move to the back of the ‘aircraft’ (a superordinate of aeroplane). This axiom comes from one of the founders of modern descriptive linguistics, J.R.Firth. His influential (1957) book contains a chapter on semantic analysis. This quotation from Astrophil and Stella is taken from the Oxford Authors edition, 1989, p. 153. One of the main inspirations for componential analysis is an article by Katz and Fodor (1963). For a fuller critique of the development and practice of componential analysis, see the references to additional reading provided at the end of the chpater. The feminist critic Dale Spender has argued that it is the system of language itself which is sexist, man-made and permanently under the control of the patriarchy. (Spender 1980, passim). A critique of Spender’s position, with specific reference to componential analysis, is provided in Cameron’s survey of feminism and linguistic theory (1985:59–62). The edition from which Roethke’s ‘Dolour’ is taken is: Words for the Wind: The Collected Verse of Theodore Roethke (1961), Bloomington: Indiana University Press; p. 55. W.H.Auden’s ‘In Memory of W.B.Yeats’ is taken from The Norton Anthology of English Literature: Volume 2, fifth edition (1986), London: Norton & Company, pp. 2299–300.
4 Exploring narrative style 1 2 3
This story, first published in 1925 by Charles Scribner’s Sons, is taken from p. 285 of The Essential Hemingway, Jonathan Cape, 1970. Two of the many studies which explore mind-style and point of view in narrative are Fowler (1986:127–46) and Simpson (1993:11–85). Originally printed in the New York Times, 20 December 1922, p. 1, col. 3. The Paris edition of the Tribune is likely to have reproduced an identical story. (See Reynolds 1972:82 for a fuller account of Hemingway’s source material.) 209
NOTES 4
I am grateful to Sonia Zyngier, her colleagues and her students at the Universidade Fédérale in Rio de Janeiro for participating in this creative writing exercise and for producing such an illuminating series of narratives. The narrative cited here is one of many elicited in the course of this workshop.
5 Dialogue and drama 1
2
3
4 5
6
7 8
210
The precise sources for the four exchanges used for the shunt are as follows. The first is from an encounter with a colleague at work. The second, the doctorpatient encounter, was overheard in a hospital ward during visiting hours in March 1994. The third exchange is from a televised interview between David Frost and the leader of the Liberal Democrats, Paddy Ashdown (The Frost Programme; BBC TV; 6 February 1994). The last results from a recent, unsuccessful attempt on my part to inspire my nineteen-month-old son to eat some breakfast cereal. In some non-standard dialect systems the rules governing the formation of imperative sentences can be rather more complex. Irish-English dialects, for example, often permit the insertion of a subject as in ‘Come you here!’ or ‘Don’t youse be goin’. Example (39) was uttered by BBC commentator, Harry Carpenter. The occasion which produced it was boxer Frank Bruno’s entrance into an arena to a thunderous rendition of ‘Land of Hope and Glory’ (16 March 1994). Simpson (1992) is a pragmatic study of the humour that this sort of gaffe produces. I am grateful to Michael Finn, an honours student at Queen’s University, for bringing this example to my attention. It is worth adding that the mismatch in (45) forms a principle which is exploited by dramatists of the absurd. It is one of the techniques used to produce startling and bizarre dialogue. Simpson (1989) explores this absurdist use of politeness phenomena in a play by Eugene Ionesco. This extract from Edward Albee’s The Zoo Story is taken from: Absurd Drama (ed. Martin Esslin) 1976; Harmondsworth: Penguin. Literary-critical introductions to this writer and his work can be found in Bigsby (1984) and McCarthy (1987). On the use of repetition as a feature of child language, notice how, in the fourth exchange of the shunt in section 5.2, the youngster repeats ‘No want it!’ The eventual outcome of the play is that Jerry manages to impale himself on a knife he has just thrown to the luckless Peter. Jerry’s dying confession is interesting because for a few brief moments there is a total change in his attitude towards Peter. This change in attitude is signalled by a concomitant shift hi discourse strategies, and, while I have not the space to deal with them here, they would none the less form the basis of a useful comparative exercise.
Adams, V. (1973) An Introduction to Modern English Word-Formation Harlow: Longman. Baker, C. (1969) Ernest Hemingway: A Life Story New York: Scribner’s. Bakhtin, M.M. (1986) Speech Genres and other Late Essays Austin: University of Texas Press. Barley, A. (1986) Taking Sides: The Fiction of John le Carré Milton Keynes: Open University Press. Baum, S.V. (1972) ‘E. E.Cummings: The Technique of Immediacy’. In Friedman (ed.), pp. 104–20. Bennison, N. (1993) ‘Discourse Analysis, Pragmatics and the Dramatic “Character”: Tom Stoppard’s Professional Foul’. Language and Literature, 2, 2, 79–100. Bigsby, C.W.E. (1969) Albee Edinburgh: Oliver & Boyd. ——(1984) A Critical Introduction to Twentieth Century American Drama vol. 2, Cambridge: Cambridge University Press. Birch, D. (1989) Language, Literature and Critical Practice London: Routledge. ——(1991) The Language of Drama: Critical Theory and Practice Houndmills: Macmillan. Blakemore, D. (1992) Understanding Utterances Oxford: Blackwell. Brown, G. and Yule, G. (1983) Discourse Analysis Cambridge: Cambridge University Press. Brown, P. and Levinson, S. (1978) ‘Universals in Language Usage: Politeness Phenomena’. In Goody, E.N. (ed.) Questions and Politeness Cambridge: Cambridge University Press, pp. 56–289. ——(1987) Politeness: Some Universals in Language Usage Cambridge: Cambridge University Press.
References
References
211
REFERENCES Brumfit, C. and Carter, R.A. (eds) (1986) Literature and Language Teaching Oxford: Oxford University Press. Burton, D. (1980) Dialogue and Discourse London: Routledge. ——(1982) ‘Conversation Pieces’. In Carter and Burton (eds), pp. 86–115. Calvo, C. (1992) ‘Pronouns of Address and Social Negotiation in As You Like If. Language and Literature, 1, 1, 5–27. Cameron, D. (1985) Feminism and Linguistic Theory London: Macmillan. Carter, R.A. (1981) ‘Linguistics and Literary Irony’. UEA Papers in Linguistics, 14– 15, 30–51. ——(ed.) (1982a) Language and Literature: An Introductory Reader in Stylistics London: Unwin Hyman/Routledge. ——(1982b) ‘Responses to Language in Poetry’. In Carter and Burton (eds), pp. 28–55. ——(1984) ‘Why have I Started to Talk to You like This?: Narrative Voices, Discourse Pragmatics and Textual Openings’. In Van Peer, W. and Renkema, J. (eds) Pragmatics and Stylistics Louvain: Acco, pp. 95–132. ——(1987) Vocabulary: Applied Linguistic Perspectives London: Unwin Hyman/ Routledge. ——(1989) ‘What is Stylistics and Why can we Teach it in Different Ways?’ In Short (ed.) (1989a), pp.161–75. ——and Burton, D. (eds) (1982) Literary Text and Language Study London: Arnold. ——and Nash, W. (1983) ‘Language and Literariness’. Prose Studies, 6, 2, 123–41. ——(1990) Seeing Through Language Oxford: Blackwell. ——and Simpson, P. (eds) (1989) Language, Discourse and Literature London: Unwin Hyman/Routledge. Chiaro, D. (1992) The Language of Jokes London: Routledge. Coulmas, F. (1989) Writing Systems of the World Oxford: Blackwell. Coulthard, M. (ed.) (1977) An Introduction to Discourse Analysis Harlow: Longman. ——(ed.) (1992) Advances in Spoken Discourse Analysis London: Routledge. ——and Montgomery, M. (eds) (1981) Studies in Discourse Analysis London: Routledge. Cruse, D.A. (1977) The Pragmatics of Lexical Specificity’. Journal of Linguistics, 13, 153–64. ——(1986) Lexical Semantics Cambridge: Cambridge University Press. Crystal, D. (1987) The Cambridge Encyclopedia of English Language Cambridge: Cambridge University Press. ——and Davy, D. (1969) Investigating English Style Harlow: Longman, de Beaugrande, R. (1987). ‘Schemas for Literary Communication’. In Halasz, L. (ed.) Literary Discourse Berlin: de Gruyter, pp. 49–99. Downes, W. (1994) ‘Register in Literature’. In Asher, R.E. and Simpson, J. M.Y. (eds) The Encyclopedia of Language and Linguistics vol. 2, Oxford: Pergamon Press, pp. 3509–11. Dumas, B. (1974) E.E.Cummings: A Remembrance of Miracles Plymouth: Vision Press. Eagleton, T. (1983) Literary Theory: An Introduction Oxford: Blackwell. Elam, K. (1980) The Semiotics of Theatre and Drama London: Methuen. Evans, G. (1977) The Language of Modern Drama London: Dent. Firth, J.R. (1957) Papers in Linguistics 1934–1951 Oxford: Oxford University Press. Fowler, R. (1986) Linguistic Criticism Oxford: Oxford University Press. Francis, G. and Hunston, S. (1992) ‘Analysing Everyday Conversation’. In Coulthard (ed.), pp. 123–61. Freeborn, D. (1986) Varieties of English (2nd edition) Houndmills: Macmillan. 212
REFERENCES Friedman, N. (ed.) (1972) E.E.Cummings Englewood Cliffs, New Jersey: Prentice Hall. Fromkin, V. and Rodman, R. (1988) An Introduction to Language (4th edition) London: Holt, Rinehart & Winston. Gregory, M. and Carroll, S. (1978) Language and Situation London: Routledge. Grice, H.P. (1975) ‘Logic and Conversation’. In Cole, J. and Morgan, J. (eds) Syntax and Semantics III New York: Academic Press, pp. 41–58. Grundy, P. (1995) Doing Pragmatics London: Edward Arnold. Halliday, M.A. K. (1966) ‘Descriptive Linguistics in Literary Studies’. In Mclntosh, A. and Halliday, M.A.K. Patterns of Language: Papers in General, Descriptive and Applied Linguistics Harlow: Longman, pp. 56–69. ——(1978) Language as Social Semiotic London: Edward Arnold. ——(1994) An Introduction to Functional Grammar (2nd edition) London: Edward Arnold. ——and Hasan, R. (1976) Cohesion in English Harlow: Longman. Haynes, J. (1989) Introducing Stylistics London: Unwin Hyman/Routledge. Herman, V. (1996) Dramatic Discourse London: Routledge. Hoey, M. (1989) ‘Discourse-Centred Stylistics: A Way Forward?’ In Carter and Simpson (eds), pp. 123–38. Hudson, R. (1980) Sociolinguistics Cambridge: Cambridge University Press. Hurford, J.R. and Heasley, B. (1983) Semantics: A Coursebook Cambridge: Cambridge University Press. Jackson, H. (1988) Words and their Meanings Harlow: Longman. Jucker, A.H. (1988) The Relevance of Politeness’. Multilingua, 7, 4, 375–84. Katz, J.J. and Fodor, J.A. (1963) ‘The Structure of a Semantic Theory’. Language, 39, 170–210. Kempson, R. (1977) Semantic Theory Cambridge: Cambridge University Press. Kidder, R.M. (1979) E.E.Cummings: An Introduction to the Poetry New York: Columbia University Press. Labov, W. (1972) Language in the Inner City Philadelphia: University of Pennsylvania Press. Laver, J. (1975) ‘Communicative Functions of Phatic Communion’. In Kendon, A., Harris, R. and Key, M. (eds) The Organisation of Behaviour in Face to Face Interaction The Hague: Mouton, pp. 215–40. ——(1981) ‘Linguistic Routines and Politeness in Greeting and Parting’. In Coulmas, F. (ed.) Conversational Routine The Hague: Mouton, pp. 289–304. Leech, G.N. (1969) A Linguistic Guide to English Poetry Harlow: Longman. ——(1981) Semantics (2nd edition) Harmondsworth: Penguin. ——and Short, M. (1981) Style in Fiction Harlow: Longman. Levinson, S. (1983) Pragmatics Cambridge: Cambridge University Press. Lindemann, B. (1990) ‘Cheap Thrills We Live By: Some Notes on the Poetics of Tabloid Headlinese’. Journal of Literary Semantics, 14, 1, 46–59. Lowe, V. (1994) ‘“Unsafe Convictions”: “Unhappy” Confessions in The Crucible’. Language and Literature, 3, 3, 175–95. Lyons, J. (1977) Semantics Cambridge: Cambridge University Press. Matthews, P.H. (1974) Morphology: An Introduction to the Theory of Word-Structure Cambridge: Cambridge University Press. McCarthy, G. (1987) Edward Albee Houndmills: Macmillan. Montgomery, M., Durant, A., Fabb, N., Furniss,T. and Mills, S. (1992) Ways of Reading London: Routledge. Nash, W. (1985) The Language of Humour Harlow: Longman. Palmer, F. (1982) Semantics (2nd edition) Cambridge: Cambridge University Press. 213
REFERENCES Pilkington, A. (1990) ‘A Relevance Theoretic View of Metaphor’. Parlance, 2, 2, 102–17. Reynolds, M. (1972) Two Hemingway Sources for In Our Time’. Studies in Short Fiction, 9, 81–6. Sampson, G. (1985) Writing Systems: A Linguistic Introduction London: Hutchinson. Shen, Y. (1987) ‘On the Structure and Understanding of Poetic Oxymoron’. Poetics Today 8, 1, 105–22. Short, M. (ed.) (1989a) Reading, Analysing and Teaching Literature Harlow: Longman. ——(1989b) ‘Discourse Analysis and the Analysis of Drama’. In Carter and Simpson (eds), pp. 139–68. ——(1994) ‘Discourse Analysis and Drama’. In Asher, R.E. and Simpson, J.M.Y. (eds) The Encyclopedia of Language and Linguistics vol. 2, Oxford: Pergamon Press, pp. 949–52. ——(forthcoming) Studying the Language of Plays, Poems and Novels Harlow: Longman. ——and Candlin, C. (1986) ‘Teaching Study Skills for English Literature’. In Brumfit and Carter (eds), pp. 89–109. Simpson, P. (1987) ‘The Narrative Structure of Hemingway’s The Old Man and the Sea’. Belfast Working Papers in Language and Linguistics, 9, pp. 167–227. ——(1988) ‘Access through Application’. Parlance, 1, 2, 5–28. ——(1989) ‘Politeness Phenomena in lonesco’s The Lesson’. In Carter and Simpson (eds), pp. 171–93. ——(1990) ‘Modality in Literary-Critical Discourse’. In Nash, W. (ed.) The Writing Scholar Newbury Park: Sage Publications, pp. 63–94. ——(1992) ‘The Pragmatics of Nonsense: Towards a Stylistics of Private Eye’s “Colemanballs”’. In Toolan, M. (ed.) Language, Text and Context London: Routledge, pp. 281–305. ——(1993) Language, Ideology and Point of View London: Routledge. ——and Montgomery, M. (1995) ‘Language, Literature and Film: The Stylistics of Bernard MacLaverty’s Cal’. In Weber, J.J. and Verdonk, P. (eds) Twentiethcentury Prose Fiction: From Text to Context London: Routledge, pp. 138–64. Sinclair, J. McH. and Coulthard, M. (1975) Towards an Analysis of Discourse Oxford: Oxford University Press. Smith, F. (1973) ‘Twelve Easy Ways to Make Learning to Read Difficult’. In Smith, F. (ed.) Psycholinguistics and Reading London: Holt, Rinehart & Winston, pp. 183–96. Smith, N.V. and Wilson, D. (1979) Modem Linguistics: The Results of Chomsky’s Revolution Harmondsworth: Penguin. Spender, D. (1980) Man Made Language London: Routledge. Sperber, D. and Wilson, D. (1986) Relevance: Communication and Cognition Oxford: Blackwell. Steen, G. (1994) Understanding Metaphor Harlow: Longman. Stubbs, M. (1982) ‘Stir Until the Plot Thickens’. In Carter and Burton (eds), pp. 57– 85. ——(1983) Discourse Analysis Oxford: Blackwell. ——(1986) Educational Linguistics Oxford: Blackwell. Thorne, J.P. (1981) ‘Generative Grammar and Stylistic Analysis’. In Freeman, D.C. (ed.) Essays in Modern Stylistics New York: Methuen, pp. 42–52. Toolan, M. (1987) ‘Analysing Conversation in Fiction: The Christmas Dinner Scene in Joyce’s Portrait of the Artist as a Young Man’. Poetics Today, 8, 2, 393–416. ——(1988) Narrative: A Critical Linguistic Introduction London: Routledge. 214
REFERENCES ——(1989) ‘Analysing Conversation in Fiction: An Example from Joyce’s Portrait’. In Carter and Simpson (eds), pp. 195–211. ——1992) ‘On Relevance Theory’. In Wolf, G. (ed.) New Departures in Linguistics New York: Garland Publications, pp. 146–62. Traugott, E.C. and Pratt, M.L. (1980) Linguistics for Students of Literature New York: Harcourt Brace Jovanovich. Trengove, G. (1986) ‘What is Robert Graves Playing at?’ In Brumfit and Carter (eds), pp. 60–9. Van Peer, W. (1986) Stylistics and Psychology: Investigations of Foregrounding London: Croom Helm. ——(1987) ‘Top-Down and Bottom-Up: Interpretative Strategies In Reading e e cummings’. New Literary History, 18, 3, 597–609. ——(1989) ‘How to Do Things with Texts: Towards a Pragmatic Foundation for the Teaching of Texts’. In Short (ed., 1989a), pp. 267–97. ——(1993) Typographic Foregrounding’. Language and Literature, 2, 1, 49–61. Verdonk, P. (1993) Twentieth Century Poetry: From Text to Context London: Routledge. Wales, K. (1989) A Dictionary of Stylistics Harlow: Longman. ——(1992) ‘Zodiac Mindwarp Meets the Horseflies’. English Today, 29, 50–1. Watson, B. (1972) ‘The Dangers of Security: E.E. Cummings’ Revolt against the Future’. In Friedman (ed.), pp. 31–45. Wegner, R.E. (1965) The Poetry and Prose of E.E.Cummings New York: Harcourt Brace. Werth, P. (1994) ‘Extended Metaphor—A Text-world Account’. Language and Literature, 3, 2, 79–103. Widdowson, H.G. (1975) Stylistics and the Teaching of Literature Harlow: Longman. Willis, D. (1992) ‘Caught in the Act: Using the Rank Scale to Address Problems of Delicacy’. In Coulthard (ed.), pp. 111–22.
215
Abstract 116, 118 absurdist plays 21, 130, 131, 134, 176; see also Albee, E. accent 180, 197 accusations 160 acronyms 44 actional frames 111, 113, 118 acts 145–7; bound 146; content-bearing 146; subordinate 146; see also speech Adams, V. 59 adjectives 14, 16, 28, 34, 42, 43, 58 adverbs 34, 110, 116 advertising language 31–2, 35, 70–1, 181 affixes 36–7, 39 agreement 163 Albee, E.: absurdist play 131, 164–14 passim, 175, 194 alliteration 16 allographs 26 allophones 26 analogy 48, 50 analysis extension 115–20; natural narrative 116–20 anaphora 107, 108, 109, 115, 190 antonymy 72–4, 83, 92, 108, 115;
complementary opposition 72; gradable 72–3; relational opposition 73; semantic derogation 74 apology 159 approval, displays of 161 articles 34, 107 asides 124 aspect 198 asymmetry 142, 171, 172 attitude 76 Auden, W.H. 95–6, 98, 185, 187 authoritarianism 6
Index
Index
back-formation 58 backtracking 47 Baker, C. 121 Bakhtin, M. 18–19 Barley, A. 17 Baum, S.V. 58 Beckett, S. 27–8, 175 Bennison, N. 177 Bigsby, C.W. E. 173 Birch, D. 178 Birmingham model 143–4, 146, 147, 192–3 Blakemore, D. 177
217
INDEX blend 43 bound morphemes 36–8, 39, 50, 57, 145, 146; see also affixes; infixes; prefixes; suffixes breakage 50 Brown, G. 176 Brown, P. 155, 156–7, 158, 162, 177 Brumfit, C. 18 Burton, D. 144–5, 177 Calvo, C. 177 Cameron, D. 74 Candlin, C. 7, 22 Carroll, L. 41 Carroll, S. 21 Carter, R.A. 18, 22, 59, 96, 98–9, 127, 181, 185 cataphora 107, 115, 190 cause-and-effect relationships 108, 111, 114, 125 Chaucer, G. 7–8 Chiaro, D. 98 choice and chain axes 77, 78, 195 classroom discourse 143–4 clipping 43 cloze procedure 84, 85–92, 94, 95, 96, 184–5, 187 co-operative principle 148, 150, 193 Coda 117, 119, 125 cognitive processing 29 Cohen, L. 22 coherence 126, 127, 133, 182, 198 cohesion 21, 102–27, 188–92, 198, 207–8; additive, adversative and causal conjuctions 115; chain 109; creative writing 120–5; devices 189; model 190; relations 190; ties 190–1; see also analysis extension; short story reconstruction collocation 78–9, 132, 134, 144, 145, 184; cloze procedure and stylistic analysis 86–7; multiple choice text and stylistic analysis 93; patterns 84, 85, 96; predictions 187; range 94, 95 collocational clashes 79–83, 88, 89, 90–1, 97
218
columnar format 27–8 combinations see under words commands 140 common opinions and attitudes 161, 163 communicative breakdown 131–2, 146–7, 173 comparator 51 complaints 160 complementary opposition 72 Complicating action 116, 118–19 compliments 161, 162 componential analysis 81–4, 88, 187 compounds 42–3 conceptual clusters 43 conjunctions 115 connectivity 115, 199 connotation 66–7, 69, 93 content words 33, 34, 35, 36, 39, 43, 50, 182 context 63, 68–9, 73, 75 contextual effects 152–3, 154, 157, 169 core narrative 94, 116, 118, 185 Coulmas, F. 59 Coulthard, M. 176, 177, 192 creative writing 120–5 criticisms 160 cross-cultural differences 63 Cruse, D.A. 98 Crystal, D. 21, 98 cummings, e e: poem 44, 57, 58, 59, 90; analysis extension 53–6; linguistic models 45–53 passim; morphology 37, 41 Davy, D. 21 de Beaugrande, R. 127 debt acknowledgement 160 declarative 51, 139–42, 160, 199 deference 159 definite reference 190–1 deformed text 56 denotation 66–7, 69, 73 derogation 74 descriptive frames 111, 113, 118, 183
INDEX deviation 54 dialect 10, 180, 199 dialogue 131–42; communicative breakdown 131–2; lexical set 138–9; mood 140–2; paradigmatic axis 134–5; request-reaction 135–7; setting 135–9; shunting 132–5, 138; strategy 134–8, 142; structure 121, 134, 137; syntagmatic axis 134–5; topic 133–4 discourse analysis 130–78, 192–5, 199, 208; maxims and relevance 148–55; scale 145; strategies 149, 156–7, 163, 170, 195(see also maxims; politeness phenomena; relevance theory); structure 143–8; theory 194; see also dialogue; drama dialogue; politeness phenomena disruptive patterns 54 Downes, W. 22 drama dialogue 20, 21, 164–74; classroom interaction 168; communicative breakdown 173; exchanges 167; maxims theory 168–9; politeness phenomena 170–2; question/response 167; relevance theory 169; statement/ acknowledgement 167; structure 166–7; three-part exchange 167–8; transactions 167 Dumas, B. 52, 58–9 dysfluencies 131 effect-and-cause relationships 114 Elam, K. 164, 178 ellipsis 108–9, 190 euphemisms 93, 95 Evaluation 117, 120, 124–5 Evans, G. 173 exchanges 147 exclamative 51, 124, 200 existential sentences 112, 200 explicatives 117, 125, 200 face see politeness false starts 131
first-person narrative 123–4 foregrounding 33, 45, 200 form-function asymmetry 142 formal repetition 191 formalist reading 120, 200–1 Fowler, R. 31, 127 fragmentation 48, 51, 52 Francis, G. 177 Freeborn, D. 22 Freudian slips 71 Friedman, N. 44 Fromkin, V. 59, 98 geographical origins 26 Germanic derivations 67–8 grammar 96, 180, 186, 201 grammatical words 33, 34, 35, 43, 182 graphemes 25–6, 27, 34, 45, 47, 64 graphetic substance 25 graphology 25–33, 103, 181–3, 201, 206; advertising language 31–2; breakage 90; perceptual strategies 29–30; psycholinguistic influence 28–9; see also graphemes greeting 163 Gregory, M. 21 Grice, H.P. 148–9, 151, 155, 157, 158, 168–9, 175, 193 Grundy, P. 177 Halliday, M.A.K. 21, 107, 108, 115, 126, 127, 190 Hasan, R. 107, 115, 126, 127, 190 Haynes, J. 127 Heasley, B. 98 hedging 159, 160, 161, 172 Hemingway, E.: short story 102, 104–5, 127, 188–9, 191–2; creative writing 120–5; general patterns of response 106–10; idealised narrative 110–12; natural narrative 117–20; original version 112–14; reconstruction 103 Henri, A. 181 Herman, V. 177 hesitations 131, 159, 160, 172
219
INDEX Hoey, M. 127 homonymy 69–72 horizontal processing 52 Hudson, R. 21 Hunston, S. 31, 127 Hurford, J.R. 98 hyponymy 74–6, 187 hypothesis formation 29 ideograms see logograms idioms 64, 93 imperatives 11, 139–42 impersonalisation 160 implicatures 149, 151, 154, 156, 158, 169, 175 imposition, minimising 159 in-group markers 161 indicatives 140, 141 indirectness 142, 150, 158, 160 inferencing 151–3, 154 infixes 36–7, 39 inflections 39 intensifies 28, 58, 117, 201 intention to communicate 153, 154 interpolations 124 interpretations 29 interrogative 51, 139–42 intimacy 163 Ionesco, E. 175 Jackson, H. 98 Joyce, J. 7–8, 42–3 Jucker, A.H. 177 Kempson, R. 98 Kidder, R.M. 58 Labov, W. 116–17, 124, 126, 127 Latinate derivations 67–8 Laver, J. 163 le Carré, J. 13–16, 18 Leavis, F.R. 6 Leech, G.N. 6, 59, 98, 114 Levinson, S. 155, 156–7, 158, 162, 176, 177 lexemes 64, 78, 80, 83, 87, 88, 89, 94; incompatible 90; juxtaposition 79
220
lexical cohesion/elegant variation 191 lexical semantics 28, 62–99, 132, 183–8, 204, 206–7; advertising language 70–1; analysis 127; antonymy 72–4; boundaries 63; choice 68; compatibility 81; connotation 66–7, 69; context-free theory 63; denotation 66–7, 69, 73; homonymy 69–72; hyponymy 74–6; idioms 65–6; incompatibility 81; link 157; paradigm 95; polysemy 69–72; puns 70–2; register 67–9; scale 74–5, 83, 187; seemingly identical 63; shifting 63; space 96; specificity 75–6, 93; strategies 95–6; superordinates 74–6; synonymy 66–9, 74 lexical set 11, 85, 89, 91, 92, 94, 111 lexicon 33, 34, 35, 40, 41, 42, 182 lexis 186, 201–2 line boundaries 31 line endings 47 line-by-line reading strategy 52 linguistic models for stylistic analysis 44–53; graphemes 45, 47; morphology 47–52; psycholinguistic perceptual strategies 45; word-formation 45 linguistic reduction 34–5 linkage 190 literary communication 13 literary language 7–19; advertising discourse 9–10, 14; rhetorical devices 8–9, 14 logograms 27, 45, 47, 55, 58 Lowe, V. 177 lower case 45, 54, 55, 58 Lyons, J. 98 McCarthy, G. 170, 173 marker 146, 161 Matthews, P.H. 59 maximal efficiency 157, 158 maxims 148–55, 175–6; flouting 169, 175; manner 148, 150–1; quality 148, 149, 151; quantity 148, 149–50, 169; relation 148,
INDEX 150, 151, 155; theory 168–9, 193 meanings see lexical semantics metaphors 14, 79, 149 Miller, A. 177 mnemonics 80 modal auxiliaries 50, 159, 202 modal verbs 34 modality 124, 125, 172, 202 modals 117 modifiers 14; multiple 15–16 Montgomery, M. 127, 177 mood 51, 140–2, 157, 202 morphemes 64, 144; derivational 37, 38, 39, 40, 57, 182; free 48, 49, 50; inflectional 38, 40, 182; plural 38; reversative function 41; see also bound; root morphology 35–57, 181–3, 206; asymmetry 40–1; back-formation 41–2; breakage 49, 90; compounds 42–3; single 50; see also morphemes moves 144–7; challenging 145, 147, 167, 175; opening 145, 175; supporting 145, 146, 147, 175 multiple choice text 84, 93–6, 185, 186, 187 mumbling 159 narrative 21, 102, 202; categories 116–17; first-person 123–4; minimal 116; style see cohesion; third-person 124; see also core; natural Nash, W. 22, 98, 181 natural narrative 21, 102, 124, 125, 126 newspaper headlines 34–5 non-sequiturs 131, 150 nonce-words 42, 57 nouns 34, 43, 58, 78, 109; antecedent 38; phrases 14, 15–16, 106; see also verb-noun paradigms obligation 137 Orientation 116, 118–19, 125
ostension 151–2, 154, 169, 193 over-cohesion 114, 115 over-specification 76 oxymoron 79, 187–8 Palmer, F. 98 paradigm/paradigmatic axis 77, 78, 79, 85, 88, 89, 93, 134–5 parallelism 9, 10; rhetorical 16; syntactic 14 parentheses 47, 55, 58 Parker, D. 181 particles 35, 36, 37, 48, 56, 58 parting 163 past continuous 116, 119, 203 perceptual strategies 29–30, 33, 45, 48, 56 pessimism 159 phatic communion 170, 172 phonemes 25, 26, 64 phonetics 25, 179–80, 183, 203 phonology 25, 179–80, 183, 203 Pilkington, A. 177 Pinter, H. 175 placatory language 163 poetry 20, 21, 79; concrete 27; see also Auden, W.H.; cummings, e e politeness phenomena 137, 150, 155–64, 172, 176, 195; bald, non-redressive strategy 157; bald, on-record strategy 170, 176; drama dialogue 170–2; face 155, 156, 157, 163, 170; Face-Threatening Acts 156–60, 162, 170, 172, 176; in-group marker 162; negative face 155, 156, 158–60, 162, 170, 172; off-record 155, 157, 158, 162; on-record 157, 158, 172; phatic communion 163–4; positive face 155, 156, 160–2, 163–4, 171, 172; pragmatic concepts 156, 157, 158 polyphony 18 polysemy 69–72 power and status 137 pragmatics 76, 77, 130–1, 156, 157, 158, 168, 195, 203
221
INDEX Pratt, M.L. 37, 41, 59 prefixes 36, 39, 40, 41, 50, 57 prepositions 34 prompt 146 pronouns 15, 34, 58, 107, 115; anaphoric 108 prose 20, 21 psycholinguistics 28, 29, 203 punning 9, 10, 51, 70–2 question/response 140, 145, 146, 167, 192 re-formation 29 re-registration 181 referent 65 register 10–19, 67–9, 88, 93, 94, 186, 203; context 12–13; field 11–12; mode 11–12; re-registration 17–19; recipe 10–11, 12; sermon 12; technical 92; tenor 11–12; tourist-travelogue 15–17 relevance theory 148–55, 157, 158, 175–6, 193 repetition 109, 191 reprimands 160 request-reaction exchange 140, 146, 192 Resolution 117 resonances 65, 67 Reynolds, M. 121 rhyming 9 Rodman, R. 59, 98 Roethke, T. 87, 91, 187 root morphemes 35, 39, 40, 42, 47, 58, 182; blending 43; discourse structure analysis 145, 146 Sampson, G. 59 semantics see lexical sense 29, 65, 69 setting 163, 170, 174–5, 194 Shakespeare, W. 7, 177 shapes to words see graphology; morphology Shen, Y. 187–8 Short, M. 6, 7, 22, 114, 164, 177 222
short story see Hemingway shunting 80–1, 132–5, 174, 194; dialogue 132–5, 138 Sidney, P. 79 silence 187 Simpson, P. 127, 177 Sinclair, J. McH. 177 Smith, N.V. 181 social distance 163 solidarity 161, 163 speech act 158, 204 speech genres, primary and secondary 18–19 Sperber, D. 151–3, 157, 158, 177, 193–4 stanzaic organisation 10 statement/acknowledgement 140, 167, 192 Sterne, L. 27 Stoppard, T. 177 structure 144, 147–8, 163, 170, 175, 195 Stubbs, M. 127, 176, 185 stylistic analysis techniques 84–96; cloze procedure 85–92; multiple choice text 93–6 stylistics 2–7; generic application 6, 15; heuristic value 4–5; intersubjective role 5; linguistic function 5; literary interpetative guise 5 substitution 190 suffixes 36, 37, 38, 39, 50, 57–8, 64; grading 57 summons 146 superordinates 74–6, 83 surface features 126 symmetrical encounters 171 synonymy 66–9, 74, 85, 87, 90–4; euphemistic 95 syntagm/syntagmatic axis 134–5; cloze procedure and stylistic analysis 85, 88, 89, 91; multiple choice text and stylistic analysis 93, 94, 95; words and combinations 77–8, 79, 80, 84 taboo topics 161, 163–4, 171, 172 tense 125, 204 text-immanent reading 120, 205
INDEX third-person narrative 124 Thorne, J.P. 97–8 Toolan, M. 126, 177 topic 133; see also taboo Transformational-Generative model 96–7 Traugott, E.C. 37, 41, 59 Trengove, G. 22 typographs see logograms under-specification 76 upper case 46, 51, 54, 55 Van Peer, W. 59, 99, 127 Venn diagram 183–4 verbs 58, 109, 124, 140; actional 119; hedging 172; morphemes and words 34, 38, 41–2, 43, 50; past continuous 116, 119; verb-noun paradigms 42 Verdonk, P. 98 visual ambiguity 31 visual information 29, 31, 32, 33 visual system 52
vocabulary 88 Vonnegut, K. Jnr 127 Wales, K. 98 Watson, B. 58 Wegner, R.E. 44, 58 wellformedness 133 Werth, P. 127 Williams, W.C. 31 Willis, D. 193 Wilson, D. 151–3, 157, 158, 177, 181, 193–4 word-play 9, 10 words and combinations 77–84; antonymy 83; collocation 78–9, 84; collocational clash 79–83; componential analysis 81–4; paradigmatic axis 77, 78, 79; semantic scales 83; shunting 80–1; superordinates 83; syntagmatic axis/ combinations 77, 78, 79, 80; see also lexical semantics Yule, G. 176
223
