Chapter 9Common LanguageCorpus, creativity and cognition*Ronald Carter
Nottingham University

Abstract
This chapter takes further debates concerning the nature of literary language and the presence of literariness in a range of discourses by exploring the extent to which everyday conversational discourse displays literary properties. The author argues that studies of literary discourse, and of the continuities between literary and non-literary discourse, have tended to focus on written language or on representations of spoken discourse in fictional or dramatic dialogues. This emphasis has made for questionable connections between literature, literacy and the written language because it assumes that spoken language is no more than a less patterned version of written language. Using the Cambridge and Nottingham Corpus of Discourse in English (CANCODE), the author shows how verbal inventiveness is pervasive in ordinary talk. The chapter concludes that common, everyday language is far from being either everyday or common – on the contrary, it is pervasively “poetic”.
Keywords
  • creativity;
  • linguistic and literary theory;
  • literariness;
  • literary language;
  • pleasure;
  • spoken discourse;
  • verbal play

1.Introduction: Writing and literary language?

This chapter explores the extent to which examples of everyday spoken discourse display literary properties. Immediately several questions are begged and below I shall examine some of the histories of definition attendant on words like “everyday” and “literary”. Studies of literary discourse and of the continuities between literary and non-literary discourse have for the most part focused on written language or on representations of spoken discourse in fictional or dramatic dialogues or on a range of individual sentence-level speech acts, reinforcing questionable connections between literature, literacy and the written language by assuming that spoken language is no more than a less patterned version of written language.
Interest in spoken discourse has been re-awoken by the large collections of naturally occurring spoken data which have become increasingly accessible, such as the BNC (British National Corpus) and the spoken sections of the COBUILD corpus at the University of Birmingham. The data on which this chapter is based come from the CANCODE corpus assembled in the Department of English Studies at the University of Nottingham.

2.CANCODE data

2.1Punning and playing

CANCODE stands for “Cambridge and Nottingham Corpus of Discourse in English” and is a multi-million-word corpus of spoken English. The working corpus from which examples can be drawn totals over five million words, although the main aim of the project was originally to construct a qualitative corpus and not simply a large quantitative corpus. In other words, data was carefully collected with reference to a range of different speech genres, with an emphasis on “common” informal discourse collected mainly in non-institutional settings. (For further discussion of the corpus design, see Carter and McCarthy 1995a; McCarthy and Carter 1995; McCarthy 1998.) Although the emphasis in work on the corpus is on lexico-grammatical description, with the aim of producing a new pedagogical grammar of English, what cannot be ignored (indeed, it leaps out at researchers from almost every transcript of the data) is the inherent creativity of significant proportions of common language use.
Here are some representative samples from the corpus.
Example 1.
Members of a family in Cardiff (1993) are preparing food for a party:
A:
Now I think you’d better start the rice
B:
Yeah … what you got there [4 secs]
A:
Will it all fit in the one
B:
No you’ll have to do two separate ones
C:
Right … what next [17 secs]
C:
Foreign body in there
B:
It’s the raisins
C:
Oh is it oh it’s rice with raisins is it
B:
no no no it’s not supposed to be [laughs] erm
C:
There must be raisin for it being in there?
Example 2.
Members of a family are telling ghost stories over dinner in Canterbury (1993). This story refers back to an earlier newspaper story involving a man having a premonition of death on the ship Hood. Speaker A begins by describing how the members of the crew were all lined up prior to being selected for the Hood and for an anticipated escape:
A:
Oh yes, I mean they were all eager to get on it they were really looking forward to being the chosen ones [B: Mm] and he was one of the ones who was called up [B: yeah] and he was getting ready to go and the chief petty officer came back and said, oh no it’s a mistake
C:
We’ve got one extra
A:
Dymock, Dymock, er you’re not needed [B: Mm] and er he was a bit disappointed and he went back, carried on with what he was doing and the boat sailed out and was torpedoed and
C:
by a German ship [B: Oh yes]
D:
Everyone, everyone died
C:
Anyway, all hands lost but legs saved [All: laugh]
B:
Well, sailors were always getting legless, weren’t they, anyway [All: laugh]
A:
Finding their sea legs
B:
Yeah
The creativity of speakers and listeners in Examples 1 and 2 here produces the most basic and widespread forms of verbal play. The comic identification in Example 1 of the word “foreign body” with a raisin and puns on the word, “raisin/reason” and in Example 2 “hand” and again in 2 puns on the word “leg” in the fixed expression “find your sea legs” and the word “legless” (meaning unable to stand up or “find your legs” as a result of being drunk) are the hinges of punning and ambiguity on which the discourses turn (Nash 1985: 137–47). We should also notice in 2 a creative play with the metonymy (“hands”) in hyponymic pattern with “legs”. Similar creative play can be illustrated by Examples 35 which follow. In these contexts the verbal play is with larger units or “chunks” of language:
Example 3.
Two friends in London (1996) are discussing a third friend’s stormy marriage and the fact that, as a result of continuing infidelity, relations between the couple are “frozen” and they are barely talking:
A.
… he’s at it again but he really wants you know just to sit down
B:
like they just talk about how they both feel
A:
out of the frying pan into the deep freeze this time
Example 4.
Two colleagues in Leicester (1995), who are social workers, are discussing a third colleague who has a tendency to become too involved in individual cases:
A:
I don’t know but she seems to have picked up all kinds of lame ducks and traumas along the way.
B:
That that’s her vocation.
A:
Perhaps it is. She should have been a counsellor.
B:
Yeah but the trouble with her is she puts all her socialist carts before the horses.
Example 5.
Two students in Bristol (1995) are talking about the landlord of a mutual friend:
A:
Yes, he must have a bob or two
B:
Whatever he does he makes money out of it just like that
A:
Bob’s your uncle
B:
He’s quite a lot of money erm tied up in property and things. He’s got a finger in all kinds of pies and houses and stuff. A couple in Bristol, one in Cleveland I think.
These examples show speakers playfully and creatively extending metaphors of “hot” and “cold” (core figurations in the expression and evaluation of human feeling) alongside a creative extension of a fixed saying (“out of the frying pan into the fire”, Example 3). Examples 4 and 5 show idioms being displaced into new patterns. For example:
“She puts all her socialist carts before the horses” (Don’t put the cart before the horse) “He’s got a finger in all kinds of pies and houses and stuff” (He’s got a finger in every pie)
And Example 5 also contains the intriguing possibility of subliminal phonological / phonaesthetic echoing (Nash 1986: 128–51) across speaking turns of the word “bob” in the fixed expression “bob or two” and the idiom “Bob’s your uncle”.

2.2Morphological inventiveness: New words for old

Examples 35 illustrate how speakers can create new meanings by means of reformulation of what have been conventionally described in linguistics as immutably fixed chunks of language. Also prevalent in CANCODE data are instances where speakers invent new words from existing words, “morphological creativity” in Carter and McCarthy (1995b).
Here are two examples in the ‘invented’ words heart drawers and crawly:
Example 6.
A group of students in Nottingham (1996) are discussing items for a jumble sale:
A:
I mean stuff like this is what we need because it’s quite quite an easy thing to design and it that looks really pretty. So what’s that?
B:
Yeah. It’s got a lovely heart as well.
C:
Heart drawer
A:
Heart drawers [all laugh]
C:
Heart drawers
A:
Somebody’s knickers. And how much was that? Seventy five?
C:
Seventy five pence …
Example 7.
A:
I’ve done the letter
B:
Right
C:
But it’s a bit bloody abrupt. It’s quite sort of
B:
Ah
D:
You can sort of
B:
No that’s fine
A:
That’s fine?
C:
That’s fine
B:
Are you sure?
C:
Yeah. You don’t want to feel crawly do you?
A:
No it’s not crawly is it. It’s all right if I want this information isn’t it?
C:
Yeah

2.3Echoing and converging

In parallel with the morphological creativity in Example 7, there is also another phenomenon at work. Speakers regularly pattern each other’s words, producing parallel structures in the form of lexical and syntactic echoes. For example, the repetition of:
“That’s fine”; “sort of”; “crawly”; “yeah”.
In the following extract from CANCODE data, the situation and speakers are the same as in Example 1 (members of a family preparing food in a kitchen in Cardiff). In this sample there are no overt markers of verbal play or morphological or other kinds of “inventiveness”. But there are patterns similar to those noted in Example 7:
Example 8.
A:
How long does it take?
B:
Erm
C:
Oh that’ll make a noise
B:
Takes about thirty five minutes yeah that’ll that that’ll destroy your tape
A:
Thirty five minutes
B:
Yeah
C:
Yeah
A:
I thought the microwave did everything in about two minutes
C:
You may as well turn it off now then
A:
Yeah you can do it on the cooker for thir, in thirty five minutes
B:
Then if you have to watch it … you just ignore it
A:
Mm
D:
You don’t have to wash the saucepan either do you?
B:
You don’t have to wash the saucepan after [laughs] [C:mm] you don’t have to erm drain the water off either
A:
I didn’t know that microwaves ran that long
B:
Yeah you don’t have to erm drain the water off either cos er
A:
I’ll switch it off when you turn that on
This multi-party sequence of exchanges contains several examples of the seemingly random topic switching and overlapping, interruptions, unanswered questions, hesitations and false starts which characterize informal conversations. On the surface there is much divergence, disconnection and incoherence. Beneath the surface there is, however, much convergence and coherence marked in a distinctive range of pattern-reinforcing linguistic features, especially repetition. For example, the speakers use each other’s words, employ parallel syntactic forms and generally pattern question and answer replies in such a way as to confirm that there are high degrees of affective connections and convergence. For example, the re-cycling and echoing of the following words and structures: “take”; “thirty five minutes”; “do/did”; “yeah”; “drain the water off”; “you don’t have to wash/drain”; “off/on”.

2.4Pattern-reforming and pattern-reinforcing

Patterns are always potentially present in language and language users always have options whether or not and what kind of patterns to create. CANCODE research suggests two main motivating choices: pattern-reforming choices and pattern-reinforcing choices. In the case of reforming choices speakers play more directly and overtly with language. They draw attention to the patterns by re- forming and re-shaping them. In extreme versions of reforming there is a more radical position created by the ‘reform’. Co-conversationalists are prompted to pleasure and laughter, to more evaluative and affective viewpoints and to a more innovative re-shaping of our ways of seeing. Speakers who reform patterns can sometimes do so by radically displacing or deviating from expected linguistic patterns.
In the case of pattern-reinforcing choices, the patterns which are developed are less overt and work primarily to create an affective convergence and a commonality of viewpoint. The patterns may not draw attention to themselves in the same way as pattern-re-forming choices; and it is also more likely that rules for existing linguistic structures will be conformed to rather than departed from. It is not, of course, suggested that these patterns operate discretely, for both pattern-reforming and pattern-reinforcing can be and often are combined.

2.5Summary

Nottingham CANCODE research is revealing a number of characteristic features of such spoken discourse: first, that common, everyday language is far from being either everyday or common (on the contrary, it is pervasively “poetic”); second, that verbal play with language is often undertaken for humorous purposes, serving in part to bring people closer together (these kinds of data can be easily found in multi-million-word corpora by running concordances for the word <laughs>; third, that this kind of linguistic creativity and inventiveness is almost always contextually embedded insofar as it depends to a measurable degree on the social relations which obtain between participants (relations between people in the above data are informal and participants are on a mainly equal social and psychological footing); fourth, that creativity involves not only more overt attention-drawing pattern-reforming forms but also more covert pattern-reinforcing forms, the latter especially across speaking turns; fifth, that it is a frequent, not exceptional feature of everyday language use and that it is also a common practice to share pleasure and convergence in and through language.
Since such language use is discoursally frequent, some of the definitions of literary language (for example, that it is “deviant” discourse) may need to be revised. It is also interesting to ask whether such definitions of “literary language” can be appropriately extended to the kinds of exchanges exemplified in CANCODE data.

3.Literary language: A brief history of definitions

Two main models for definition can be discerned and these can be grouped, rather loosely, into inherence and socio-cultural models, although the division is by no means a clear-cut one. We shall begin with the more formalist inherency definitions because they are both historically antecedent and more overtly observable.

3.1Inherency models; deviation theory and self-referentiality

Inherency definitions are predicated on a division between poetic and practical language. According to deviation theory, literariness or poeticality inheres in the degrees to which language use departs or deviates from expected configurations and normal patterns of language and thus defamiliarizes the reader. Literary language use is therefore different because it makes strange, disturbs, upsets our routinized “normal” view of things and thus generates new or renewed perceptions. In a much quoted poetic example, Dylan Thomas’s “a grief ago” would be poetic by virtue of its departure from semantic selection restrictions which state that only temporal nouns such as week or month can occur in such a sequence. As a result, however, grief comes to be perceived as a temporal process.
Another influential “inherency” definition is particularly associated with Roman Jakobson. In a famous paper, Jakobson (1960) articulated a theory of poetic language which stressed the self-referentiality of poetic language. Thus, in the examples:
I hate horrible Harry
or
I like Ike
the verbs “hate” and “like” are selected rather than, for example, “loathe” or “support” because they establish a reinforcing phonaesthetic patterning. The examples cited (the latter is Jakobson’s own and is a specific slogan in favour of the former American president Dwight Eisenhower, whose nickname was Ike) demonstrate that poeticality can inhere in such everyday language as political advertising slogans. Jakobson’s definition is, like definitions of deviation theory, founded in an assumed distinction between “poetic” and “pragmatic” language. According to Jakobson, in non-literary discourse the signifier is a mere vehicle for the signified. In literary discourse it is brought into a much more active and reinforcing relationship serving, as it were, to symbolize or represent the signified as well as to refer to it.
This emphasis on patternings and parallelisms and on the self-referential and representational nature of literary discourse is valuable; but it should be pointed out that (1) Jakobson’s criteria work rather better in respect of poetry than of prose and; (2) he supplied no clear criteria for determining the degrees of poeticality or “literariness” in his examples, and did not make clear what exactly makes some messages more unequivocally aesthetic examples than others; and (3) like deviation theorists, Jakobson stresses too much the production of effects, neglecting in the process the recognition and reception of such effects. The reader or receiver (or listener) of the message and his or her socio-cultural position tend to get left out of account.

3.2Socio-cultural models

Initial accounts of literary language which attempt more boldly to underscore the role of the reader or receiver interacting in a socio-linguistic context with the sender of a verbal message have been generally termed speech act theories of literary discourse. One of their main proponents was Richard Ohmann.
Ohmann’s basic proposition is that the kinds of conditions which normally attach to speech acts such as insulting, questioning and promising do not obtain in literary contexts. Instead we have quasi- or mimetic speech acts. As Ohmann puts it:
A literary work is a discourse whose sentences lack the illocutionary forces that would normally attach to them … specifically, a literary work purportedly imitates (or reports) a series of speech acts, which in fact have no other existence … Since the quasi-speech acts of literature are not carrying on the world’s business – describing, urging, contracting, etc., the reader may well attend to them in a non-pragmatic way and thus allow them to realize their emotive potential.(Ohmann 1971: 2)
Thus, the literary speech act is typically a different kind of speech act – one which involves (on the part of the reader) a suspension of the normal pragmatic functions words may have in order for the reader to regard them as in some way representing or displaying the actions they would normally perform.
Ohmann’s theory, like all inherency models, again suffers from an essentialist opposition between literary and non-literary which careful (historical) consideration does not really bear out. Pratt (1977), for example, has convincingly demonstrated that non-fictional, non-pragmatic, mimetic, disinterested, playful speech acts routinely occur outside what is called literature. Hypothesizing, telling white lies, pretending, playing devil’s advocate, imagining, fantasizing, relating jokes or anecdotes, even using illustrations to underscore a point in scholarly argument, are then, by Ohmann’s definition, literary. (See further McCarthy and Carter 1994: ch. 4.)
In a more extreme version of these arguments, Eagleton (1983: 10) argues that “anything can be literature” and that it is all a matter of how we choose (or are chosen) to read a text. Because texts are extremely varied and the social and cultural positions from which readers read texts even more varied, a definition of literature can only be relative to specific contexts. In certain institutionalized contexts such as a department of literature in a school or university, definitions of literature will be made by the selection of texts for study which will in turn do no more than reflect the “interests”, predispositions and theories of those teachers, publishing houses or examination boards which make the prescription.
Do such perspectives help us in our explorations of CANCODE data? Are the inherent features only noticed by certain “institutionally trained” speakers? (Apparently not, since a number of features are universally commented on or, more commonly, stimulate laughter.) Do features such as punning “stand out” in some way discoursally as deviant or are they accepted as normal? And what about the kinds of features indicated in Example 8 which go unremarked upon but which on one level may be very deliberately patterned but are not manifestly either deviant or self-referential? Are pattern-reforming and pattern-reinforcing in speech purposeless non-pragmatic embellishments or do they fulfil more basic purposes?

3.2.1Presentationality

Petterson’s work (1990) develops another socio-cultural definition of literature by locating linguistic analysis within more established domains of literary and cultural theory. Petterson demonstrates that what constitutes literary language and literariness is historically and culturally variable and that many of the issues to do with the deviancy and self-reflexivity of literary language belong more comfortably within the 20th century and within modernist accounts of literary formations. Petterson’s main conclusion is that literature (in its most central contemporary sense) in most historical and most western cultural settings is a verbal composition which is especially marked by its presentationality. For Petterson literature differs from other categories he explores such as functionally informative and directive discourse because “to verbally understand serious literature is … to understand that it is presentational and to apprehend it in the intended manner is to seek to obtain emotional, cognitive or formal aesthetic satisfaction from it” (Petterson 1990: 256).
Petterson’s position finds general support in studies devoted to more universalist and cognitive preoccupations with aesthetic functions. Bauman (1977, 1986), for example, stresses the “designed for performance” aspect of literary presentation, drawing many examples from more oral-based cultures and arguing that “aesthetic” satisfaction may simply result from a display and appropriate evaluation of verbal skill, the evaluation being based on a judgement of the extent to which clear public rules for the performance have been adhered to. (See also established work on “verbal duelling”/“flyting”). Bever (1986) points out that aesthetic satisfaction is produced not simply by straightforward arousal but rather by the pleasure which ensues when difficulty is followed by resolution, a good example being when a complex narrative is resolved by an ingenious plot outcome or in which a specific moral problem is addressed and solutions proposed. Parallelism (a feature not dissimilar from the pattern-reinforcing noted in CANCODE data) is another example in which the first part is resolved by the reinstatement of a related pattern in the second part, a process which serves as a kind of psychological closure or completion.
In earlier studies Berlyne (1971) had gone even further, and with an apparent recognition of the demands of many modernist texts, had argued that a presentation of aesthetic arousal can be autonomous and that, indeed, complexity need not be resolved and may even be all the more pleasurable for remaining so. Cognitive anthropologists in this tradition are, however, less concerned with specific textual functions than with the universal, non-culture-specific character of verbal art.

3.2.2Preliminary conclusions

The definitions above are largely based on a recognition of the social and cultural nature of literary language as discourse and rightly situate speakers and listeners (as well as writers and readers) as essential constituents of the definition. The definitions assume the inherently distinctive nature of literary language and establish that such distinctiveness is relative to social and cultural contexts and to the assignment of such distinctiveness by human subjects. This assignment of distinctiveness assumes, however, that “literary” language can only be measured against the norms of “ordinary” discourse and therefore assumes by omission that ordinary discourse cannot be literary. The explanatory models discussed so far do not, however, account adequately for the kinds of cognitive-affective or psycholinguistic responses to literary language suggested by them.
In relation to Examples 18 given earlier can we account for the pattern-reforming and pattern-reinforcing tendencies of the (largely) socially symmetrical spoken CANCODE discourse in terms of providing pleasure? Can we say that the (aesthetic) pleasure results mainly from the registering of new morphologies, extended idiomatic constructions and convergent parallelisms, or are there more dynamic “problem-solving” processes at work? Do listeners co-construct a discourse so that both speakers and listeners achieve greater mutuality? Is the interaction more or less successful if precise interpretation is not achieved? Answers to such questions demand greater attention to psychological and psycholinguistic factors.
For a long time the field of poetics and stylistics has relied for its inspiration on inherency and socio-cultural explanatory models. Their adequacy is, however, not accepted by a growing and ever more influential body of work in the field of psycholinguistics and of literary study within the explanatory models and paradigms of cognitive science. Such models have started from studies of the endemically figurative character of language.

3.3Cognitive models

A continuing inspiration for much ongoing study of figurative language is the seminal work by Lakoff and Johnson (1980), Lakoff and Turner (1989) and Turner (1991). The starting point and continuing emphasis of this research are that human language and the human mind are not inherently literal. In writings by cognitive linguists figurative language is not so much seen as deviant or ornamental but rather as ubiquitous in everyday language, especially spoken language. Discussions of figurative language proceed on the assumption that the fundamental roots of language are figurative.
Such an assumption and the research paradigms which follow from it present a radically different set of beliefs from the beliefs about human thought and language which have traditionally dominated the disciplines of the humanities and social sciences in the western intellectual tradition. Gibbs (1994) has argued for a cognitive wager which contrasts with the more standard generative wager. The generative wager hypothesizes that explanations of language and of language universals in particular are structure-dependent and that linguistic constructs are autonomous of general conceptual knowledge. The cognitive wager of Gibbs and other cognitive linguists aims to show that there is no autonomous language faculty and to illustrate that language is not independent of the mind. Gibbs argues that figurative schemes of thought structure many fundamental aspects of our ordinary, conceptual understanding of experience.
An example of this position is provided by Gibbs with reference to the polysemous word “stand” which has a range of everyday meanings: for example, “He could not stand the pressure”; “The law still stands’”; “‘The barometer stands at 29.56”; “The house stands in a field”. Gibbs points out that the basic meaning of “stand” is one of a physical movement or physical act. Other meanings of “stand” extend this basic sense, often metaphorically, to convey meanings of verticality, resistance to attacks (as a result of “standing firm”, remaining vertical in the face of attempts to unbalance or knock you down) and endurance (to remain upright); for example: “He stands over six feet tall”; “He stood up to all the attacks against his theory”; “The law still stands”. One interesting conclusion from these examples is that there is a link by metaphoric extension between physical action and mental representation. The figurative often has an origin in physical, bodily experience and the figurative framework of everyday thought motivates a surprising number of meanings in this and other examples, so that phrases such as “to take a stand on something”, “to uphold” (principles/the law), to remain an “upright” person derive from the same underlying, conceptually coherent domain. Traditional studies in lexical semantics attempt to uncover the componential set of features underlying each separate word “stand” and begin from an assumption of literalness. Cognitive linguists put “the body back into the mind”, arguing that
metaphor, and to a lesser extent metonymy, is the main mechanism through which we comprehend abstract concepts and perform abstract reasoning … [and that] metaphorical understanding is grounded in nonmetaphorical preconceptual structures that arise from everyday bodily experience.(Gibbs 1994: 11).
While traditional lexical semantic studies search for literal meaning on the grounds that literal meaning best reflects the truth values of an objectively determined external world, cognitive linguists such as Gibbs recognize that so-called literal language is itself constituted by fundamental processes of figuration.
Studies by Gibbs and others (Johnson 1987; Lakoff and Turner 1989; Sweetser 1990) raise fundamental questions for our understanding of the nature of literary language. Metaphor has always been seen as a fundamentally literary property as a result of the apparent propensity of its users to create new insights into human experience and values; and metaphorization has been conventionally regarded as a liberating process in which divergent and deautomatizing ways of thinking are made possible. Gibbs offers an alternative mapping of creative metaphoric processes by illustrating the extent to which poetry can depend on basic underlying metaphors which structure our most fundamental, common everyday experiences.
In the case of love poetry, for example, one of the prototypical ways in which ordinary language use construes love is in terms of a basic metaphor of love as a nutrient. That is, love is conventionally described in terms of the energy and sustenance it provides. For example: “They were kept going by their love for each other”; “She’s been starved of affection for too long”; “hungry for love”; “I was given new strength by her love”. Other everyday metaphorical construals are of love as a force which can cause a loss of control: for example, “She’s crazy about him”; “They are quite besotted with each other”; “He was burning with passion”; “addicted to love”; “sexual obsession”; “she pursued him relentlessly”. Sometimes these two basic metaphoric frameworks converge so that the nutrient itself can be intoxicating. Gibbs (1994) cites a poem “I taste a liquor never brewed” by Emily Dickinson in which he illustrates how the poet’s description of the experience of love is founded upon such familiar, everyday metaphoric concepts.
Gibbs’ work on metaphor implicitly accepts the notion of a cline between creative and everyday metaphors in ways parallel to that outlined earlier. But he has less to say on processes of the contextual interpretation of metaphor and on the contextual embeddedness and on what the communicative ‘risks’ are when metaphors are creatively extended, especially in literary contexts. Gibbs does, however, provide evidence to support the speculation in Toolan (1996) and Carter (1997: ch. 7) that there is a delimitable set of core, productive and culturally salient vocabulary items that predominate in both conventional and creative metaphors.

4.All language is literary language

The opposition of literary to non-literary language is an unhelpful one and the notion of literary language as a yes/no category should be replaced by one which sees literary language as a continuum, a cline of literariness in language use with some uses of language being marked as more literary than others, broadly following the main outlines of the argument advanced in Carter and Nash (1990: ch. 2). Although the most immediate focus is on text-intrinsic linguistic features, it should not be forgotten that whether the reader (or listener) chooses to “read” or respond to a text (spoken or written) in a literary way, as a poetic text as it were, is one crucial determinant of its literariness.
Such a position is echoed in the writings of many 20th century literary theorists and philosophers, beginning with pre-modernist thinkers such as Nietzsche. Tambling (1988) points out that for Nietzsche all language is literary language because even supposedly referential language has no original reference point and because even a referential statement is a rhetorical device, one which is designed to persuade the listener or reader to act or to think in a particular way. The position is close to that of Derrida’s discussions of textuality in which for him, as Tambling argues:
“Literary language” is a pleonasm … all language is literary, because it is all mere writing (the earlier meaning of “literary”), and it can all be read for the guileful, ambiguous and indeterminate uses of language that literature employs.
[…] “what is literary language?” is not a question to be asked merely by those who study “literature”: it affects those who write history, philosophy, political science or science itself.(Tambling 1988: 74)
Lecercle (1990) adopts a not dissimilar line, arguing that many models of language of the twentieth century have been impoverished by a failure properly to examine the more creative “remainder” of language. The “remainder”, as he terms it, is necessarily ignored by more formalist inspired models which idealize language systems and which, following the lead of Saussure in particular, focus on langue to the exclusion of parole. The “remainder” only comes to our notice when we examine real uses of conversational language or investigate it diachronically or look at particular forms such as poetry or explore everyday metaphors, puns, riddles and verbal games. Lecercle questions whether such uses of language can ever be fully formalized but he argues that linguistic systems would be all the richer for recognizing their existence, pointing out in the process that the preoccupation of much modern linguistics with invented data, sentence-level grammar and a narrowly truth-condition-determined semantics does not allow any direct engagement with such data and their associated issues.

4.1Pleasure and verbal play: Risks and rewards

Several of the above-mentioned commentators on “literary” creative and non- literal discourses have pointed to the pleasure conferred by and derived from verbal play. The notion that ordinary language users regularly and typically communicate in ways from which pleasure may be derived has gained considerable ground in recent years and correspondingly the assumption that pleasure is the sole preserve of the highest forms of artistic encounter may need to be re-inspected.
Cook (1995, 1996) argues that “verbal play” is characteristic of discourse in which language is used to “fill up the spaces between necessary activity” (Cook 1996: 198) underlining the essentially non-utilitarian nature of literary discourse. Like casual conversation, in which the motivation to talk is not necessarily determined by any specific purpose, literature is what Cook defines as a “space-filling discourse”. Cook also argues that literary linguistic analysis should focus more systematically than hitherto on comedy, popular song, graffiti, newspaper headlines and advertisements.
In Cook (1995) there is, in a way which relates interestingly to work described above by Gibbs, an attempt to explain the pleasure derived from patterning in poetry in terms of mental mechanisms which fulfil basic needs of the human species as a whole; and among the more basic of mental mechanisms are those which are renewed and refreshed as a result of the destabilizing effects of literary patterning. One suggestion is that playing with words may be genetically determined and that recasting or deviating from established patterns is a natural and normal ‘biological’ reflex of the human mind which cognitive linguists are in any case increasingly inclined to regard as figuratively predisposed. Again a main conclusion is that literariness in language is normal.
Cook (1994) employs key concepts which are alive in a number of approaches discussed in this article: he stresses the interactive nature of reading, highlighting features of textual design which activate involvement and interpersonality; and, in particular, he underlines the discoursal deviance which can act as a specific trigger to what he terms “schema refreshment”. The notion of schema refreshment is based on work in artificial intelligence (AI) and is what might be termed a textual-cognitive criterion, reflecting the subtitle to Cook’s study “The Interplay of Form and Mind”. Cook analyses a wide range of discourse types for their schema-disrupting properties, arguing that what is generally understood to be aesthetically valued literature, at least within certain cultural (usually western) conceptions of the term, will be the texts which introduce schematic or cognitive refreshment. That is, an advertisement or a joke may disrupt at a textual or linguistic level of deviance but often does not offer any real challenge to cognitive renewal in terms of how the world is perceived.
Bhaya et al. (1988), Toolan (1996) and Carter (1997) explore the extent to which the existence of creative and inventive language involves risks for the creator. The risks mainly involve failures of uptake, the embarrassment of unsuccessful performances and lapsed ‘presentationality’. Puns and extended or newly minted metaphors, in particular, involve acts of language use which are relatively self-conscious and there are risks involved in the undertaking. On the other hand, the risks can bring communicative rewards. Pleasure can be conferred, convergence created and the presenter can in certain contexts achieve an enhanced regard or what Boxer and Cortes-Conde (1997) term ‘identity display’.

4.2More CANCODE data

Example 9.
Two friends – primary school teachers in Nottingham (1996) – are discussing arrangements for changing classes:
A:
So if Monday was clear which we need to check with …
B:
I teach Lenton on Monday
A:
Right
B:
If they’re not going on a trip or anything like that
A:
Well they won’t go on a trip for sure
B:
So they should be here
A:
So it could be Monday if the test’s on Tuesday. Right.
B:
And that wouldn’t upset any apple carts would it? No?
Example 10.
Three Art College students who share a house in Carmarthen, Wales, are having tea and chatting informally among themselves (1994):
A:
I like Sunday nights for some reason, I don’t know why
B:
[laughs] cos you come home
A:
I come home
B:
You come home to us
C:
And pig out
B:
Yeah yeah ….
B:
It’s an earring, it’s an earring
A:
Oh lovely, oh lovely
B:
It’s fallen apart a bit but
A:
It’s quite a nice one actually, I like that, I bet, is that supposed to be straight?
B:
Yeah
C:
Oh I think it looks better like that
B:
And there was another bit as well, another dangly bit
A:
What, attached to
B:
The top bit
A:
That one
B:
Yeah … so it was even
A:
Mobile earrings
C:
Well, that looks better like that, it looks like that
These data illustrate the greater risks which attach to pattern-reforming choices as in Example 9 Speaker B introduces an idiom (“upset apple carts”) which risks being seen as (in)appropriately inventive, affective and lightening amid the information transfer; and in Example 10 the imagistic creativity produced by comparing earrings to a mobile may also not receive the desired uptake, though the informality and social symmetry of the context reduce attendant risks. And alongside such pattern-reforming runs an extensive set of pattern-reinforcing echoes and repetitions (especially in Example 10) which on another level create a mutuality and commonality conducive to the verbal play.
Descriptions of CANCODE data involve us in clines, in definitions of “inherent” language features and in accounts of different social and cultural contexts of use. They also involve us in cognitive assessments and psychological judgements. In such cases pleasure can be communicated but there is a cline or clines which runs from overt to covert, from intended to unintended and, most crucially, from recognized to unrecognized, the last polarity raising long-debated complexities of intentionality and its related measurement.
What is clear is that a poetics of everyday spoken discourse will be inadequate if it only relies on a single descriptive or explanatory framework. Such a poetics cannot therefore be an inherency nor a socio-cultural nor a cognitive poetics but will be likely to be all the more powerful for combining components of all three.

5.Developing a socio-psychological aesthetics

It is clear that a poetics which attempted to capture and define some of the characteristically “literary” features of the CANCODE data will need to recognize the inherently non-literal nature of language, the potential extent of its “presentationality”, the human processes of purposeful engagement with and interpretation of such language and the contextual factors determining its production and reception: in other words, a linguistic poetics which captures both psychological reflexes and social contexts. At the same time, although five million words of naturally occurring data represent substantial evidence, the provisional and limited nature of the data has to be recognized and the following attempt to provide defining categories accepted as a starting point only.
On the basis of the examples from CANCODE and related data cited in Examples 110 earlier the following observations can be advanced.
  1. Non-literal language is more common than is conventionally assumed within most contemporary paradigms for the discipline of linguistics. In the data examined metaphors and linguistic patternings of various kinds are pervasive.
  2. Non-literal language is purposeful and not merely ornamental. One main function is to play with such language in order to confer pleasure. (Michael Toolan [personal communication] suggests, for example, that one of the main motivations for playful, creative language in conversation is that we be thought of as someone who is ‘fun to be with’ and that word play facilitates a display of such an identity.)
  3. In the CANCODE corpus the most frequent examples of verbal play in the use of such language include:
    1. speaker displacement of fixedness, particularly of idioms and formulaic phrases;
    2. metaphor extension;
    3. morphological inventiveness;
    4. verbal play and punning with overlapping forms and meanings;
    5. ‘echoing’ by repetition, including phonological echoes;
    6. strategically affective use of standard idioms.
  4. Verbal play is, however, not decontextualized. It is not simply a biological reflex. In CANCODE data verbal play coincides with social contexts which are informal and where relations between participants are broadly symmetrical. Thus features listed in the previous paragraph do not commonly occur in the data in relatively more formal settings such as interviews, university tutorials or conversations between strangers.
  5. Informal settings are, however, the contexts of use which are most frequently encountered by most speakers. In such settings the main purposes of verbal play can be interpreted to be involving, affective and ‘convergence-creating’. That is, speakers and listeners jointly co-construct playful discourse with the aim of aligning and harmonizing and sharing ways of seeing, ‘re-forming’ and ‘reinforcing’ the informality of the relationship.
  6. There are clearly two levels of assessment of ‘creative’ interactions: first, more overt presentational uses of language, open displays of metaphoric invention, punning, uses of idioms and departures from expected idiomatic formulations (pattern-reforming); second, less overt, may be even subconscious and subliminal parallelisms, echoes and related matchings which regularly result in expressions of affective convergence, in implicit signals of intimacy and in consciously unrecognizable symmetries of feeling (pattern-reinforcing).
  7. There is always an element of risk-taking in such verbal procedures. The punning may not work and attempts at morphological inventiveness may fail but there are corresponding rewards in the establishment of greater intersubjective accord, intimacy and involvement and in the establishment of discourse with a clear affective overlay. The effects can be “schema-refreshing”; for example, in the case of jokes and anecdotes and other similarly constructed instances. The main purpose is, however, one of producing not primarily new ways of seeing but rather democracies of affective interaction (see Carter and McCarthy 1997a, 1997b).
  8. The frequency of such linguistic creativity, in CANCODE data at least, means that to some extent non-literal language exchanges are unmarked while more literal exchanges are marked. Such a perspective reverses the usual assumption that non-literal uses of language are deviations from a standard, “literal”, referential norm.

6.Questions for research: Linguistic and literary theory

The following topics and issues are suggested as areas for further research, including pedagogic research. The areas continue lines of enquiry discussed so far but they do not by any means constitute an exhaustive list.
  1. One of the most basic points of a pedagogical postscript to this article is to reinforce that all learners of a language and all students of literature already possess considerable powers of response to, and creativity with, the literariness potential, the ‘remainder’ which exists in all languages. Methodologies and practices should therefore continue to devote classroom space to exploration of the continuities between literary and non-literary language. Literature teaching should proceed from a more complete openness to assumptions concerning what students already can do with language. The greater the language awareness which results from discussion of ordinary language use, especially informal conversational use, the more such awareness can lead to greater explicitness and fuller awareness of processes of interpretation. Being more conscious of processes of interpretations, particularly in relation to “everyday” interpretive work, is likely to lead to greater confidence and competence in interpretations of texts which are more “semantically dense”.
  2. The question of degrees of interpersonality and “inter-subjective accord’” can be related to the need for fuller definition of the speech genres in which they are typically enacted. For example, is verbal play more likely to co-occur with informal contexts of interaction? In narrative genres rather than genres of argumentation? In casual conversation, with its varying degrees of intimacy, rather than in, say, information reports? Databases of conversational English need to be mined to discover the degrees of interpersonal embedding which occur relative to individual speech genres. In this exploration a specification of clines of casualness in conversation will be a necessary prerequisite.
  3. It is clear that the whole question of pleasure and the kinds of pleasure communicated by and in creative, literary interaction is under-researched. The topics and questions to be investigated might include: the extent to which the most prominent triggers of pleasure can be identified and the extent to which they are linguistically marked. Is pleasure more likely to be produced by the resolution of complex, problematically dense forms or by the resistance of a text to resolution of cognitive or formal complexity? How culture- or social group-specific is such “aesthetic” pleasure (Herrnstein-Smith 1988)? How far is the pleasure conveyed by literary language connected with the kinds of language which challenge conventional meanings, breaking up distinctions that have been culturally formed?
  4. Research reported in this article underlines the pervasively ‘literary’ character of ordinary language. If so much ordinary discourse involves ordinary language users in creative processes, then what is it about human involvement with language which might explain such a predisposition? An answer to this question in part demands that conventional approaches to meaning which centre on truth and reference and literalness are suspended in favour of a view of language in which its essential character is inherently figurative and in which users regularly derive an essential pleasure from their engagement and involvement with such language in use. Traditional approaches to semantics which seek referential, context-free, truth-conditional determination of meaning need to be complemented by an approach which does not see poetic language as a distortion or violation, nor as a mere embellishment of objective content, but which recognizes non-literalness as a basic linguistic and communicative characteristic of human languages.
More fully to account for such properties of language requires that existing analytical preoccupations are at least made more inclusive and above all made to engage with authentic texts, especially spoken texts. To do so requires consideration of the affective, interpersonal and bodily characteristics of language. A stronger version of such a position entails a reversal of the primary concerns that dominated 20th century linguistics, making for a 21st century linguistics which makes the non-literal, the affective and the creative the normal, everyday, common starting-point for investigation.

Acknowledgements

I am grateful to my close colleague Michael McCarthy, co-director with me of the CANCODE project, for helpful comments on an earlier draft of this chapter and for allowing me to use CANCODE data, some of which he collected, for exemplification. I am also grateful to Michael Toolan, and Katie Wales for some perceptive points regarding the argument developed in the chapter.

Note

*This chapter is a revised version of an article which appeared in Language and Literature (1999) Vol 8(3): 195–216. The editor is grateful to SAGE publications Copyright Publications (London, Thousand Oaks, CA and New Delhi) for their kind permission to reprint this material.

References