i) What is stylistics?
Modern stylistics has existed in one form or another since Charles Bally coined the term stylistique in 1909, but its roots arguably lie ‘in the elocutio of Aristotelian rhetorical studies’. In spite of its longevity, however, stylistics as a discipline continues to resist definition, owing perhaps to its restless absorption of the newest linguistic models and theories, and its chameleonic adaptation to the prevailing linguistic Zeitgeist. A corollary of this particular characteristic is that stylistics has collected many different names along the way: literary stylistics, literary linguistics, linguistic stylistics, linguistic criticism, the new stylistics, practical stylistics, and so forth. At its most basic level, stylistics is the study of literary style, and its proponents base their analyses of literary texts in predominantly, although not necessarily exclusively, linguistic analysis. Paul Simpson, a practitioner in this field, writes that ‘what…sets stylistics apart from other types of critical practice is its emphasis, first and foremost, on the language of the text…what captures the essence of the stylistic method is the primacy which it assigns to language’.
In recent years two disciplines closely related to stylistics have grown rapidly: first that of cognitive linguistics, ‘where language, thought, and conceptualization are seen to be embodied’, a field of inquiry with which I am not concerned in this thesis, and second, that of critical linguistics. The scope of critical linguistics extends far beyond texts perceived as literary, and its motivating principle is ‘to explore the value systems and sets of beliefs which reside in texts; to explore, in other words, ideology in language’. It is perhaps this attention to ‘the social function of linguistic structures in literature’ which has prompted David Robey to write that ‘stylistics has…begun to converge with aspects of Marxist criticism’, a statement that is more properly true of critical linguistics. However, the two disciplines of stylistics and critical linguistics do often overlap – in the study of transitivity, for example, to which I shall return in chapter three.
To fully grasp the aims and objectives of stylistics, it is useful to have some idea of what was perceived as deficient in the pre-existing critical status quo. The twentieth century witnessed a great deal of critical activity and the emergence of numerous critical approaches ranging from Russian Formalism to deconstruction and semiotics to psychoanalytic theory. And yet for all this activity, when critic Ian Watt was asked in 1960 to write a paper on the style of Henry James, he described himself as ‘virtually helpless…as far as any fully developed and acceptable technique of explicating prose is concerned’. The ‘acceptable technique’ was still conspicuous by its absence in 1964 when Richard Ohmann wrote ‘the most serviceable studies of style continue to proceed from the critic’s naked intuition, fortified against the winds of ignorance only by literary sophistication and the tattered garments of traditional grammar’. And more recently still, Michael Toolan has written that
reading and writing about complex texts are skills, and…literary linguistics can be an invaluable crutch or catalyst…. There remain many graduate teachers of English…who lack the procedures – and procedural confidence – even to get started on their own communicable assessment of a Stevens poem or a Heaney sonnet.
There are now some excellent stylistics textbooks available, such as those by H. G. Widdowson, Rob Pope, and Paul Simpson, not to mention Geoffrey Leech and Mick Short’s seminal Style in Fiction, designed to help both teachers and students of literature negotiate the arguably uncharted terrain of literary style. Katie Wales, the author of A Dictionary of Stylistics, notes that stylistics ‘also helps students to be more independent in their judgments, by forming their own interpretations of literary texts based on close readings and to be more confident in articulating them’. The confident articulation of literary judgments has been facilitated by the precise and rigorous vocabulary of linguistic analysis that has been imported into stylistic studies. Keith Green writes that:
stylistics arose partly because of the need in literary criticism to work with a set of agreed-upon and defined terms for the analysis and description of a particular kind of language, the language of literature. Such a language…would be built upon modern linguistic analysis.
Relatively speaking, linguistic terms have well-established and inflexible meanings; there is a large core terminology and set of concepts that most linguistic models draw on, but it is also the case that alternative theories and/or models tend to be accompanied by new technical terms. Nevertheless, it is still true to say that there is more agreement over the precise meaning of linguistic terms than there is over some of the terms routinely used in literary criticism – ‘style’ itself being a case in point, a term much-used and perhaps little understood. And it is important that people who have essentially the same goal in view should understand one another. Vague and flabby terms simply add to the general confusion.
Hence the discipline of stylistics has grown considerably in response to the various demands of students, teachers, and critics: a demand in the first instance for a workable technique of literary analysis applicable in particular to prose texts. Stylistics furnishes the student of literature with a starting-point. A preliminary linguistic analysis based, for example, on the checklist provided by Leech and Short in chapter three of their Style in Fiction provides a way of collecting data to be analysed.
My final point in this brief survey is that in grounding the analysis in the text itself, stylistics avoids the pitfall of bending the text to fit the theory. In a spectacular swing away from the text-centred theories that dominated the first half of the twentieth century, critical approaches to literature from roughly the 1960s onwards grew increasingly alienated from the text, prompting Wolfgang Iser to write in 1974:
all too often literary critics tend to produce their theories on the basis of an esthetics that is predominantly abstract, derived from and conditioned by philosophy rather than by literature – with the regrettable result that they reduce texts to the proportions of their theories, instead of adapting their theories to fit in with the texts.
By way of contrast, Keith Green notes that ‘stylistics in its anglicized form has tended to eschew the philosophical complexities and self-reflexive obsessions of literary theory.’ For some, stylistics answers the need to return to the words on the page, with the advantage of having discarded the inhibiting notion of ‘literariness’ championed by the Russian Formalists and having also shed the somewhat whimsical concept of verbal iconicity advocated by the Anglo-American New Critics.
There are, it would seem, many benefits to be reaped from a close working partnership between the two separate disciplines of linguistics and literary criticism. Critics gain a methodology and a vocabulary with which to formulate and support their literary hypotheses, and linguists acquire a fertile testing ground – literary texts – on which to try out their latest theories: ‘linguistic models offer a “way in” to a text, while the text itself allows for a challenging application for those models’. But in spite of the mutual advantages to be gained from such a partnership, there exists a long and well-documented history of antagonism between linguists and literary critics, with, it appears, much blinkered obstinacy and wilful misunderstanding on both sides of the fence.
Literary hackles were early raised by the strident and aggressive war-cries of linguists who made far-reaching claims for their subject, and linguists were in their turn infuriated by the critics’ ignorance of basic linguistic concepts, their lack of familiarity with the range of published material, and their assumption that linguists were simply mentally ill-equipped to deal with literary texts. It is the fate of the stylistician to have been caught in the No Man’s Land between the two camps, linguistic and critical. A more detailed understanding of the peculiar position occupied by the stylistician can be partially achieved by a brief consideration of some of the objections raised against stylistics and a look at how – or if – these objections have been countered.
One criticism of stylistics frequently voiced is that there is little, if any, difference between a textual analysis purporting to be linguistic and the literary critic’s activity of close reading. For the linguists at least this is not considered a problem. In 1966 Fowler wrote that ‘modern descriptive linguistics is a natural companion to modern criticism because both are text-centred: both involve analysis, close reading, and both set a premium on accuracy and usefulness of description’. But some critics apparently do not want such a companion and feel affronted by those linguists who presume to encroach upon their territory – hankering after ‘literary forbidden fruit’, as Helen Vendler puts it. In ‘The Limitations of Stylistics’, Peter Barry comments upon Fowler’s essay ‘Language and the Reader: Shakespeare’s Sonnet 73’ as follows: ‘nothing in this interpretation of the poem’s imagery would be beyond the scope of conventional close-reading – indeed, it seems precisely the kind of observation which close-reading fosters’. Barry also observes that presenting an insight ‘in linguistic dress…does not convert it into something intrinsically linguistic’. The frustration felt by the literary critics would be entirely understandable were it not for the fact that there is an important difference between a stylistic analysis and a close reading, as explained by Short:
there is some considerable overlap between stylistic analysis and the more detailed forms of practical criticism. The difference is, in part, one of degree rather than kind. Practical critics use evidence from the text, and therefore sometimes the language of the text, to support what they say. But the evidence tends to be much more selective than that which a stylistician would want to bring to bear. In that sense, stylistics is the logical extension of practical criticism. In order to avoid as much as possible the dangers of partiality, stylisticians…try to make their descriptions and analyses as detailed, as systematic and as thorough as possible.
This summary of the stylistician’s analytical approach may be somewhat idealistic, but Short’s words do at least highlight the distance between a comprehensive stylistic analysis and close reading – the difference is in the levels of selectivity and partiality. But whether or not a stylistician can really lay claim to a complete absence of selectivity in his analyses brings me to the second criticism often levelled at stylistics as a discipline – that its claim to objectivity is spurious.
In Exploring the Language of Poems, Plays and Prose, Short writes that ‘an enduring problem in literary criticism is that critics sometimes come to partial conclusions by concentrating on some aspects of a text to the exclusion of others’. But, critics argue, the same is true of stylisticians. If stylistics is to lay claim to a scientific objectivity, then each and every stylistic analysis of a text should be utterly exhaustive – otherwise the stylistician, like the critic, lays himself open to the charge of focusing only on those areas of the text which actively support his original hypothesis. But can there be such a thing as a truly exhaustive description of a literary text? Such a description, even of a short poem, would be of considerable length; exhaustive descriptions have been attempted in the past, only to be vilified for their tedium and their ultimate failure to shed any new light on the text in question. Any reasonably exhaustive analysis of a poem is bound to throw up a great deal of information which is of little or no interest to a literary critic – and an exhaustive analysis of a novel would take years, possibly decades, to complete, and is therefore quite impractical. So one must concede that the stylistician is indeed selective when it comes to choosing areas of a text on which to focus. Opponents of stylistics have not failed to point this out and stylisticians themselves have not attempted to deny this: ‘any attempt to write a complete grammar of the poem would be gratuitous and would obscure the object of the analysis, for “one doesn’t want to know every syntactic fact about a poem; one wants to know the significant ones”.’
But in fact stylisticians do not make their claim to objectivity on the basis of whether or not their analyses are exhaustive. The stylistician’s objectivity is allegedly centred in his methods and the uses to which the data gathered is put. He revolves around linguistic data and literary hypothesis in a circular movement first described by linguist-critic Leo Spitzer, and summarised here by Katie Wales:
[Spitzer’s] concept of the philological circle…is characteristic of the stylistician’s procedure still today: constantly and delicately moving between hypothesis, linguistic analysis of data, and critical explanation and aesthetic response, with a revised hypothesis if necessary.
The deficiency of practical criticism as it stands is that the movement from intuition to text is not balanced by any movement back towards the original hypothesis. The critic’s first intuitive interpretation is allowed to stand unchallenged. It may even be the case that words and phrases that are unhelpful, or that actively contradict the critic’s hypothesis, are simply ignored. Spitzer’s method of constantly oscillating between text and interpretation should, in theory, remove all, or at least part, of the danger of selecting only those details which usefully support an initial arbitrary reading.
Other stylisticians have voiced their opinions in the objectivity debate. Short argues that if a stylistician formulates an interpretation of a literary text before conducting a linguistic analysis, he is not being any less objective:
objectivity comes not from the order in which you do things, but in being systematic and careful in your argument, not overlooking unhelpful facts, and, in more general terms, constructing a clear and detailed relationship between your interpretative hypothesis on the one hand and your analysis on the other.
As with the previous examples, it is the close relationship between linguistic data and literary interpretation that is highlighted. Simpson takes a different tack in playing down any original claim to objectivity that the stylistician may or may not have made. He describes what he believes to be a common misconception:
the analyst stands by disinterestedly while the linguistic machine squeezes out of a text whatever meanings have been put there by the writer. Yet few stylisticians claim such objectivity. They prefer to recognise instead that all interpretations are in some sense context-bound and are contingent on the position of the analyst relative to the text.
In recognising that interpretation and criticism do not take place in isolation – that these activities are communal rather than individual – Simpson brings stylistics into the realm of literary pragmatics, a field of inquiry that ‘is concerned with the user’s role in the societal production and consumption of texts’.
To recap the arguments so far, then. The literary critics accuse the stylisticians of doing nothing more than producing close readings of literary texts and passing them off as linguistic analyses. The stylisticians reply that there is a certain amount of truth in this, but that their readings differ from those of the critics because a stylistic reading is supposedly more comprehensive, more meta-theoretical and more self-aware. But while stylisticians gather more extensive linguistic data, there is always an element of selectivity involved. So is stylistics therefore just a more technical version of practical criticism? Apparently not, because the stylistician’s claim to objectivity derives from his working methods and meta-methods.
However – and here we come to the most serious charge levelled at stylistics and those who practise it – critics have claimed, with some justification, that these working methods are very far from being infallible; it has been suggested that the link between the linguistic data gathered and the interpretation finally offered is frequently tenuous, with the result that in the end a stylistic interpretation of a given text is equally as arbitrary in nature as the exegesis supplied by a literary critic relying solely on his critical perception and intuition. Stanley Fish is one such critic who has put forward this argument. In response to Ohmann’s work on transformational grammar, Fish has written that ‘a stylistician will interpose a formidable apparatus between his descriptive and interpretive acts, thus obscuring the absence of any connection between them’. John Russell earlier voiced a similar concern: in what is a measured and tactful response to Ohmann’s article ‘Literature as Sentences’, Russell comments on the effort involved in sifting through the linguistic apparatus necessary to Ohmann’s way of proceeding. While conceding that Ohmann’s work seems ‘just what is needed in one way for alerting critics brought up under traditional grammar and rhetoric’, Russell notes that he is ‘led to wonder, since meaning is the final goal of the analyst, whether the yield is worth the effort if so much that is intuitional must be worked over first’, and he concludes that ‘a traditional rhetorical analysis…would take one further and faster toward the same goal’. The charge, then, is that stylisticians inevitably produce the same or similar interpretations as critics intuitively produce, but a stylistician needs more time to arrive at the same destination because he takes a lengthy detour down twisting linguistic byways. In addition, this detour is an unnecessary one, because the end result is the same: a critical reading based on intuition instead of empirical facts in spite of the stylistician’s claims to the contrary.
How is one to answer this charge? There is certainly a grain of truth in the critics’ claims. It is easy enough to find examples of linguistic readings of literary texts that either describe the grammatical structure of the text without contributing anything to a discussion of its meaning, or there are those readings whose comments on the meaning of the text are only loosely linked to the accompanying linguistic analysis; I have already cited Barry’s criticisms of Fowler’s work as an example of the latter. However, it is equally possible to find instances of linguistic – or stylistic – readings that do make sensitive and perceptive comments on literary texts, and whose comments are grounded in linguistic analysis: for example, Fowler’s discussion concerning point of view in Hemingway’s ‘The Killers’, and Leech and Short’s astute and discerning textual interpretations in chapter three of Style in Fiction. I would argue that one also has to ask what was expected at the outset – is stylistics supposed to provide the reader with a foolproof method of determining the ‘meaning’ of a literary work? The idea of the stylistician and the literary critic reaching the same destination contains an implicit assumption that there is a single ‘meaning’ to be found, which is unlikely to be the case. George Dillon writes with refreshing frankness on this subject:
what sort of knowledge, then, can linguistic analysis provide the student of literature…? It cannot be a discovery procedure for finding interpretations, nor can it be a proof or validation of an interpretation… One engages in formal analysis to specify and articulate one’s own response and perhaps to share it with others. It seems as though we are providing the grounds of our response, and from grounds to causes is but a short step, but in fact we may only be working out the consequences and ramifications of our response, according to rather loose and flexible canons governing how responses can be grounded in texts and with ample scope left for the ingenuity of the critic in formulating the poem’s ‘fashion of speaking’.
With these words, Dillon sits contentedly in a comfortable middle-ground. Linguistic analysis does not claim to provide or validate the ‘answers’, nor does it dictate the method by which these answers should be derived. The claims of those who advocate a stylistic approach are arguably far more modest: that linguistic analysis provides both a way of engaging with the text and the means of articulating and supporting a critical response, a response which does not pretend to be definitive, but which is firmly based in the language of the text. It might be objected at this point that we are back to practical criticism again, but the advanced form of close reading offered by stylistic study is not the only tool available in the stylistician’s kit.
The collected jumble of means and methods which make up the stylistician’s tool-kit is another easy target for the invective of suspicious literary critics. The eclectic nature of this tool-kit reflects the eclectic nature of stylistics itself. In 1972 Fowler wrote that stylistics ‘is a very diffuse and diverse set of interdisciplinary endeavours…no single descriptive practice is being recommended.’ This diffuseness of methods and approaches may be better understood when one considers the variety of activities in which stylisticians have traditionally been employed: pedagogy, pragmatic analysis, critical discourse, technical studies of poetic meter, attempts to classify and describe authorial style, forensic linguistics, and so on. But there is no denying that stylisticians make use of many theories, linguistic or otherwise; in 1964 Ohmann commented that ‘the attempt to isolate the cues one attends to in identifying styles and in writing stylistic parody has sprawled out into an almost embarrassing profusion of critical methods’, and he proceeds to list twelve different methods of which he personally is aware.
The problem, as perceived by the detractors of stylistics, is precisely this practice of picking and choosing whichever method of analysis happens to be the most convenient or the most fashionable. Such a facility lends stylistics a disreputable air and denies it status as a methodology into the bargain. Stylisticians have also been vilified for advocating methods of analysis which have produced non-predictive rules: rules which produce the desired results for one text but which cannot be applied with equal success to another. What is in fact revealed here is the stylistician’s commitment to theory-building. When an analytical method yields unsatisfactory results, the inadequacies of the theory are highlighted, thereby providing useful data for the ever-flexible stylistician. Inflexible approaches to the study of language seem doomed to failure, so perhaps the stylistician is wise to keep his tool-kit so well-stocked. He also draws on other related fields such as psycholinguistic theory, pragmatics and semantics, to name but a few. And of course, stylisticians do not discount the writings of critics – quite the opposite. Short has argued that these writings are themselves part of the stylistician’s tool-kit:
stylistic analysis is just as interested…in established interpretations as in new ones. This is because we are also profoundly interested in the rules and procedures which we, as readers, intuitively know and apply in order to understand what we read. Thus, stylisticians try to discover not just what a text means, but also how it comes to mean what it does.
The question of a reader’s intuition is an interesting one and it is a question to which I shall return in a later chapter. For the present, I wish to conclude this section of the introduction by addressing an issue central to stylistics, that of the supposed existence of a ‘literary’ language, a language that is different from ‘ordinary’ language.
The question of whether or not there is a special kind of literary language is one that cannot be ignored, and it is undoubtedly one that has important ramifications for this thesis. Simpson points out that it is an axiom of modern stylistics that there is no distinction to be drawn between language in everyday use and the language one finds in works valued as literature. If a category of language that is specifically literary does not exist, there is no reason why linguistic theories should not be applied to literary texts, and it is certainly true that many linguists have not ventured into reading literary texts except as language data. Halliday writes that ‘what the linguist does when faced with a literary text is the same as what he does when faced with any text that he is going to describe’. The literary language debate also has significance for the second question posed in this thesis, whether form is inseparable from content. Arguably, one of the properties of literary language is its self-referential nature: the language calls attention to itself, and has no practical function beyond this. If there is no such thing as literary language, then what is said is perhaps more important than how it is said, which leads to the conclusion that any utterance contains a paraphrasable content which is independent of form. But there are several arguments to be weighed in the balance, and I shall begin with that of linguistic register.
The ability to recognise and reproduce different registers forms part of a language user’s ‘communicative competence’. Registers are distinctive varieties of language used in different situation types. Halliday noticed that a single word or phrase is often enough to cue recognition of a register: for example, the word ‘begat’ is recognisable as biblical language, while ‘indemnity’ is a word one associates with insurance documents. It can be seen, then, that there is an integral link between register and situation type – it is unlikely that one would use the phrase ‘dearly beloved’ outside of a church service, for example – but it takes only a moment’s reflection to realise that this cannot apply to literary texts. Fowler notes that ‘the literary text’s language is not embedded in a real context of situation…but it creates its own situation, topic, and world for the reader to enter.’ The fictional situation is, after all, fictional, and the author’s use of register is mimetic of language use in everyday reality. These arguments reflect the ‘sovereign’ nature of the text, to which I shall shortly return.
It is usually not the case that texts exhibit only one register. Such texts do exist, however, and these texts are described as ‘monosemic’, ‘over-registrated’, or ‘hegemonic’; they are works in which ‘there are no resonances or networks of multiple and expanding meaning’, for instance, legal documents and instruction manuals. In general, most texts are plural, or polysemic, in that they contain a number of registers. Mikhail Bakhtin’s term for this phenomenon is ‘heteroglossia’, and literary texts in particular make use of many different registers. The following example is taken from Stevie Smith’s Novel on Yellow Paper:
Should I Marry a Foreigner?…You do not say, dear, if he is a man of colour. Even if it is only a faint tea rose – don’t. I know what it will mean to you to GIVE HIM UP but funny things happen with colour, it often slips over, and sometimes darkens from year to year and it is so difficult to match up. White always looks well at weddings and will wash and wear and if you like to write to me again, enclosing a stamped addressed envelope, I will give you the name of a special soap I always use it myself and do not stretch or wring but hang to dry in a cool oven. My best wishes for your happiness, dear, I think it was very sweet of you to write.
The registers included here are those of a written reply from an agony aunt in a magazine (‘You do not say, dear’, ‘write to me again, enclosing a stamped addressed envelope’), an advertisement for soap powder (‘will wash and wear’), and care instructions for a garment which one might find on the label (‘do not stretch or wring but hang to dry in a cool oven’). However, the fact that a text is polysemic cannot be taken as an indication that its language is therefore literary; as previously stated, many texts not considered literary are heteroglossic nevertheless.
One could ask whether literary language should itself be considered a register: however, there are problems with this suggestion. Taken out of context, literary language can appear decidedly un-literary. Terry Eagleton uses two examples to demonstrate this point. He writes that
if you…murmur ‘Thou still unravished bride of quietness,’ then I am instantly aware that I am in the presence of the literary. I know this because the texture, rhythm and resonance of your words are in excess of their abstractable meaning…. Your language draws attention to itself, flaunts its material being.
By way of contrast, Eagleton quotes the following from Knut Hamsun’s novel Hunger: ‘This is awfully squiggly handwriting!’, and points out that
the context tells me that it is literary; but the language itself has no inherent properties or qualities which might distinguish it from other kinds of discourse, and someone might well say this in a pub without being admired for their literary dexterity.
It might be relatively easy to locate lines of poetry in a literary register, but as Eagleton demonstrates, this is more difficult with prose works, especially those that are more transparent than opaque. An additional difficulty is that the existence of a literary register would be intrinsically linked to a definition of literature itself, and this definition has proved notoriously elusive. The existence of a canon of works considered to be literature is not sufficient as a criterion because the canon reveals far more about the society which compiled it than it does about those works which constitute it. Besides, as Ben Burton and Ronald Carter point out, ‘canons are not immutable…tastes change and evaluations shift as part of a process of canon formation’.
Just as anything can be literature, Ronald Carter and Walter Nash have recognised that anything can be literary. In an article entitled ‘Language and Literariness’, Carter and Nash note that literary works make use of registers that are considered non-literary, but by implication this means that any word can be literary if it is employed in a literary text. Carter and Nash label this process as ‘reregistration’: a word or phrase connected with one register is transferred to a literary work, and once this happens the original register can no longer be seen to apply because the integral link between register and situation type has been broken. As Fowler puts it, ‘the external register is therefore in the new text for some purpose other than its original function. Inevitably it becomes “estranged”.’ Bakhtin analyses this shift in terms of the ‘utterance’. He divides speech genres into primary (simple) and secondary (complex), novels being an instance of the latter. Bakhtin writes that ‘primary genres are altered and assume a special character when they enter into complex ones. They lose their immediate relation to actual reality and to the real utterances of others.’ In the Stevie Smith passage quoted above, the mixture of registers from an advice column, a soap advertisement and a set of care instructions combine to give a humorous picture of a husband who is an accessory that must be made to ‘match up’; he is likened to a garment that will stay as good as new year after year, but only with careful use (and only if he was the ‘right’ colour to begin with!).
In addition to their concept of reregistration, Carter and Nash have also coined the term ‘sovereignty’. They note that the literary text differs from other texts because it is ‘not “used” for any practical purpose; it teaches its own use; it is sovereign in its own domain of language’. Sovereignty ‘denotes the self-supporting capacity of the text, its power to generate and develop a pattern of meaning, without reference to externals and without requiring of its readers any prior knowledge other than the common stock of experience’. In addition, the reader is not required to do anything in a functional sense, as he would be if he were reading a recipe, for example. Carter and Nash also put forward ‘displaced interaction’ as a possible marker of a literary text: displaced interaction refers to the interaction of author and reader (Halliday’s tenor, or interpersonal function), and it relates to the distance between addresser and addressee that the text is required to bridge.
It would be useful at this point to summarise and recap the arguments offered so far. There is considerable doubt over whether such a thing as a language use that is specifically literary actually exists; the reasons for doubting its existence include the lack of a definition of literature itself, and the fact that there is nothing inherently literary in much of the language which appears in a literary context. Literary language has doubtful status as a register because the lack of formal criteria make it difficult to identify, and in addition to this, literary texts make frequent use of non-literary registers, thus potentially redefining every language use as literary. This borrowing from other registers occurs in many types of language use, thus polysemy alone cannot be a possible marker of a literary text. Other possible markers include sovereignty, self-referentiality and displaced interaction: these are among the criteria listed by Burton and Carter, who argue for a functional view of literary language:
literary language is not special or different, in that any formal feature termed ‘literary’ can be found in other discourses.
Yet, literary language is different from other language uses in that it functions differently. Some of the differences can be demarcated with reference to such criteria as medium dependence, reregistration, semantic density produced by the interaction of linguistic levels, displaced interaction, polysemy, and discourse patterning. What is prototypically literary will be a text that meets most of the above criteria.
This ‘prototypical’ approach suits those who argue for a cline of literariness rather than an inflexible, problematic ruling about what is and is not literary language. As far as this thesis is concerned, I feel that this approach serves to endorse the use of linguistic descriptions of language in literary texts, if the criteria listed by Burton and Carter are to be adequately explored and described. The argument that literary language functions differently is a consideration that will have to be incorporated into the form/content debate, to which I shall turn in the next section of this introduction.
ii) Form = Content?
In embarking on this thesis I decided to set myself the project of answering to my own satisfaction the familiar question, whether form and content in literature are the same thing: is it possible to paraphrase a literary text? In carrying out my research into stylistic criticism and in preparing my own critical studies of style I have found this question a useful focus. I intend to argue that for given definitions of form and content, the two are indeed inseparable.
The belief that form and content are inseparable is a tenet of monism, whose advocates include Jan Mukařovský of the Prague School of poetics, New Critics such as John Crowe Ransom and Cleanth Brooks, and more recently, the novelist and critic David Lodge. To the monist way of thinking, a change in the wording of an utterance provokes an alteration in the meaning, which is to say that it is impossible to paraphrase a literary text. Brooks published his essay ‘The Heresy of Paraphrase’ at a time when critics were engaged in adjusting their focus away from authorial intention and from literature as socio-historical documentation, and were coming to rest their gaze instead on the language of the text, favouring close and attentive reading above historical and biographical research. Much water has flowed under the bridge since Brooks wrote The Well-Wrought Urn, and he was of course referring specifically to poetry instead of prose, but there is still much that is appealing in the idea that form and content are one and the same thing, particularly when one is confronted with a prose text that does not render itself easily to paraphrase. The monist position can be an extreme one, however; form and content are seen not only as inseparable but as indistinguishable. The excesses of monism can be checked by a degree of rationality. For example, the New Critics’ somewhat immoderate concept of iconicity is one such excess, and Barry refers to this concept as the ‘enactment fallacy’:
to the habit of exclusive concentration on the poem’s verbal envelope the New Critics…added the doctrine of the organic fusion of form and content, making it obligatory to see formal details as intimately connected with content, since they had to enact meaning if they were not to be puritanically condemned as merely decorative.
The notion of the verbal icon fell from favour once it was acknowledged that writers are in fact largely constrained by the language in which they write, and it is simply not possible for formal details to be always enacting content. There is obviously a difference, however, between the idea that form enacts content, and the supposition that form and content are inseparable: the two positions are not mutually exclusive, and it is possible for the former position to be untenable without greatly disturbing the equilibrium of the latter. The monist position can also be criticised for its emphasis on the concept of ‘literariness’, as originally devised by the Russian Formalists; if this concept can be placed on a rational footing by using the criteria outlined in the concluding paragraphs of the previous section of this introduction instead of relying on the subjective instincts of the critic, then once again the monist position is rendered more attractive.
In writing about the impossibility of paraphrasing a literary text Brooks was referring to poetry rather than prose, and the monist way of thinking is certainly more in accordance with poetic texts. One might ask why, considering that I intend to defend the view that form and content are inseparable, I have chosen to explore prose texts instead of poetry; my task would certainly have been easier had I chosen to write about the latter. Indeed, much of the critical discussion relating to form and content refers largely to poetry – and yet I have chosen prose. My reasons for doing so are quite simple. I do not see why the arguments traditionally put forward for suggesting that form and content are indivisible in poetry should not equally apply to prose, whether that prose be transparent or opaque. I dislike the implication that poetry is intrinsically more ‘literary’: the verbal patterning that exists between the various linguistic levels builds up wider resonances and deeper levels of meaning in both verse and prose. I believe that prose rhythms can be representational, although I accept that this is more generally a feature of poetic texts. In addition, I consider it misleading to suppose that language draws attention to itself only in poetic texts: consider, for example, the intricate prose style of Henry James in The Ambassadors, and how delicately the lines are poised. Indeed, in chapter one, I demonstrate that a passage from Mervyn Peake’s Gormenghast  can easily be reconstructed as a poem, owing to the unusual density of parallel constructions in a handful of lines. But my argument is not that some prose should be considered akin to poetry because it has poetic features: on the contrary, even the most sparse prose is the product of authorial decision and every authorial choice has to be weighed in the balance.
To return to the main argument of this section, it is important to clarify that which I wish it to be understood by my use of the terms ‘form’ and ‘content’, because these definitions are central to my argument. For my purposes, ‘form’ refers to the words on the page, how these words are arranged into sentences and paragraphs, the punctuation of a text, its layout, any additional considerations of graphology or orthography – in short, everything concerning the physical appearance of the text, and everything to which the reader has immediate access. More importantly, I also understand form to refer to the whole text, which as previously noted, is the only valid unit of study when dealing with literary texts. If ‘content’ is inseparable from form, then meaning resides in areas such as lexical choice at word level, and syntactic arrangement and punctuation at sentence level. On a textual level, meaning can be found in the arrangement and structuring of material, similar to the Russian notion of syuzhet, or the story as shaped and edited by the storyteller. To the definition of content must also be added the dimension of context, or the way in which literary language functions: the novel is a different kind of language event from an instruction manual. However, I do not wish to claim outright that form is inseparable from content only in literary texts; nor do I wish to deny that it is possible to paraphrase some texts (the instruction manual again). I think it is safe to conclude that the further one travels along the cline of ‘literariness’ towards ‘literature’, the more it is true that form and content are the same thing; but this premise also holds true for non-literary texts that exhibit signs of literariness: some advertisements are very carefully constructed, and many jokes would not be funny – that is, would cease to be jokes – if they were stripped of their layers of verbal patterning. So, to recap, form = the whole text, and content = the whole text + context (literary language).
I would now like to take a closer look at the linguistic attempts to isolate the paraphrasable element of an utterance. These attempts were grounded in the theories of deep/surface structure and transformational grammar, and their success has been negligible. As far as my argument is concerned, the problem with these concepts is that they deal with utterances on a sentential level only; no account is taken of the ways in which sentences are punctuated, or how they work together within the text as a whole, with the inevitable result that many levels of meaning are ignored.
The deep structure of a sentence comprises two important components: proposition and modality. The proposition is the bare bones, the relationship between a noun and a predicate expressed in the simplest terms. Fowler describes the predicate as the ‘semantic nucleus’ of a proposition; it is often realised as a verb or an adjective in the surface structure. The three basic types of predicate are action, state or change of state. If the predicate represents an event, or a state of being, the nouns of the proposition represent the corresponding participants or objects. These participants/objects are assigned different roles in the deep structure, namely, agent, object, experiencers, patients, beneficiaries, instrument and location. The modality component relates to point of view in fiction, and it expresses the writer’s attitude towards both his subject, and his reader.
The deep structure undergoes a transformation via various rules of realization, and the result is the surface structure, ‘an indirect expression of underlying semantic organization’. So an utterance which is ambiguous has more than one possible deep structure expressed by one surface structure; a paraphrase, on the other hand, is one of several alternative surface structures relating to a single deep structure. At first glance, it would seem that it is, after all, possible to change the form and retain the meaning, or, as Fowler puts it:
sentences which are superficially dissimilar but ‘mean the same’ (are synonymous) are said to have the same deep structure. We can see that there is thus no one-to-one relationship between meaning and form; meaning is constant while form or surface structure diverges.
However, there are at least three problems with this argument: neither the role of the reader nor the part punctuation has to play are properly accounted for, and the notion of meaning, or content, as it is understood here, is inadequate for the purposes of literary language.
The symbols on the page are the only tangible contact the reader has with the semantic meaning buried in the deep structure:
our only access to the underlying meaning of texts is via the orders, forms and choices of words which we encounter on the surface, that is to say, we experience meaning only in the form given by the realization rules, the transformations, which the text employs. Meaning always comes to us processed by the form in which it is expressed.
If the reader can only access the deep structure (content) via the surface structure (form), then surely, for the reader, the form is the same as the meaning? Having claimed that meaning and form do not enjoy a one-to-one relationship, Fowler states that ‘a writer may transform his deep structures into surface structures which radically modify our apprehension of the propositional meaning of the text’. It would seem, then, that the way in which the writer finally expresses the deep structure on paper can completely alter the way in which the fictional world is perceived by the reader:
different surface structures make a radical difference to the impression the text makes on the reader: to his sense of the author’s tone, of the rhythm of the text, of its affiliations with other texts; above all, to the reader’s impression of the place of a text and of its author among the thought-patterns of a culture.
Surely then, the only possible conclusion can be that meaning resides in the surface structure, and that form is therefore inseparable from content. This argument can be further supported by reference to the role of punctuation. Punctuation carries meaning: one only has to remove the punctuation of a text to see how vital it is to writing. And yet we often manage without it – text messages, for example, rarely carry punctuation, and it is a sad fact that many native English speakers live in woeful ignorance of the use and function of the apostrophe, and yet people still manage to communicate successfully. In literary texts, however, the absence or intentional misuse of punctuation is usually significant of something connected to the wider themes and purposes of the text; for example, in Daniel Keyes’ short story ‘Flowers for Algernon’, Charlie Gordon’s atrocious punctuation and spelling signify, within the novel discourse, his low intelligence:
I had a test today. I think I faled it. and I think that may be now they wont use me. What happind is a nice young man was in the room and he had some white cards with ink spilled all over them. He sed Charlie what do you see on this card. I was very skared even tho I had my rabits foot in my pockit because when I was a kid I always faled tests in school and I spillled ink to.
When Charlie’s IQ is scientifically augmented, his punctuation and spelling are faultless. In addition to the argument above, it should also be remembered that punctuation and textual layout are vitally important to the fictional representation of speech and thought: this is an area to which I shall return in chapter two.
The third objection to the deep/surface structure theory is that it employs a definition of meaning, or content, that I consider to be severely impoverished; for example, a definition of meaning as Fowler intends the term to be used is ‘cognitive or propositional meaning residing in deep structure’. For the critic, this is simply too bald a definition to be of any use: ideas of propositional meaning, descriptive synonymy and truth conditions may satisfy the linguist, but the critic has need of a more detailed approach. And it seems that so much linguistic information has come to be included in the deep structure of an utterance, that the resulting structure hardly differs from its companion on the surface. In his article ‘Style and the Concept of Deep Structure’, Fowler denies the existence of optional transformations and argues that if transformations are to be meaning-preserving, everything must be accounted for in deep structure, including syntactic arrangements and lexical choices. Fowler preserves the deep/surface structure distinction, however, by maintaining that transformations do not contribute to meaning if they leave ‘the paraphrasable content of a text untouched’. But it is difficult to understand exactly what comprises this ‘paraphrasable content’ if deep structure is to include both the words that will appear in the final surface structure and their syntactic arrangement.
In the same article, Fowler quotes extensively from the work of Ohmann, a staunch defender of transformational grammar. Fowler summarises Ohmann’s conclusion as follows: ‘the style of an author or a text can be expressed by a statement of the characteristic transformations which are employed, or of the characteristic combinations of different types of transformation’. For Ohmann, the ‘notion of style calls for different ways of expressing the same content’. To summarise this argument as briefly as possible: in any human activity – and Ohmann uses the analogy of a game of tennis – there are fixed and variable components, that is, rules that must be obeyed, and the freedom of the individual to act within the boundaries of these rules. So, a tennis player must abide by the rules of the game, but he can choose the shots he wishes to play. Although the analogy does not correspond without difficulty to the activity of writing prose – it is possible, for example, to break grammatical rules for aesthetic effect without rendering a text incomprehensible to a reader – it holds well enough to support Ohmann’s argument that
the idea of style implies that words on a page might have been different, or differently arranged, without a corresponding difference in substance. Another writer would have said it another way. For the idea of style to apply, in short, writing must involve choices of verbal formulation.
Given that the reader’s only access to the underlying meaning is via the surface structure, and given that there are in all probability several choices of expression available to the author, any final authorial decision must therefore also articulate all possible variants that were eventually rejected. This conclusion lends support to Ohmann’s proposal that in order to arrive at a description of an author’s style which is not simply impressionistic and can be supported by valid linguistic data, it is necessary for the critic to explore those sentences which the author did not write, as a means to formulating a hypothesis, linked to the thematic and aesthetic qualities of the text, which can explain why the author chose to write as he did.
The next problem to be tackled is how to formulate these alternative constructions without straying from the original. Ohmann postulates that transformational rules can allow the critic or reader to do precisely this:
a transformation works changes on structure, but normally leaves part of the structure unchanged. And in any case, the new structure bears a precisely specifiable relationship to the old one, a relationship, incidentally, that speakers of the language will intuitively feel.
Ohmann notes that ‘some transformations import new content, others eliminate features of content, and no transformation leaves content absolutely unaltered’. Although this statement appears to undermine his line of argument, Ohmann is careful to qualify what he means by content: ‘the most useful sense of “content” – cognitive content – may be such that transformations do generally leave it unaltered’.
Ohmann, like Fowler, uses only a limited definition of content, and it is arguably a definition that bears little relevance to literary texts; and although Ohmann confidently asserts that ‘another writer would have said it another way’, we are still unsure what it actually is. It is also interesting to note that Leech and Short attempt an investigation into possible stylistic alternatives in chapter four of Style in Fiction, in which they provide a detailed analysis of a line from one of Katherine Mansfield’s short stories (without the aid of transformational grammar!). Their conclusion is that the line written by Mansfield is superior to its alternative renderings on semantic, syntactic and phonological grounds – and they demonstrate effectively that in fact it is not possible to rewrite Mansfield’s sentence. Ohmann’s own attempts at critical analysis based on the methods dictated by transformational grammar have not met with universal acceptance. In his article ‘Literature as Sentences’, Ohmann uses transformational grammar to analyse sentences from James Joyce’s Araby and Conrad’s The Secret Sharer. This article was met with various rebuttals, including a paper entitled ‘Linguistic Structure and Literary Meaning’ by David Hirsch. Hirsch argues that the theory of deep structure is inadequately equipped to deal with literary language, and he demonstrates clearly the deficiencies of Ohmann’s artificially generated sentences which supposedly form the core of Joyce’s chosen sentence. Hirsch concludes as follows:
the fact is that no other arrangement of Joyce’s words (or substitutes for them) could possibly communicate the same content…. This is not to say that form and content can never be separated. It seems likely that in our everyday utterances we communicate meanings in one form that could as easily have been communicated in another. But the language of poetry is different. It communicates cognitive and emotive meanings in a special way.
We are back, then, to the idea that literary language functions differently, and that as far as literary language is concerned, form and content are indivisible. Instead of attempting to formulate dubious paraphrases in an attempt to isolate and describe an author’s style, it surely makes more sense to investigate the text in terms of medium dependence, reregistration, semantic density, displaced interaction, polysemy, and discourse patterning – the suggested criteria for identifying literary language. To the equation form = content, I would argue that we can now add another component: form = content = style.
iii) Texts and Theories
The remainder of this thesis is divided into four chapters on the subjects of foregrounding, speech and thought representation, point of view, and the role of the reader. I have referred to four narrative fictions: Mervyn Peake’s Gormenghast, Emily Brontë’s Wuthering Heights, Terry Pratchett’s Going Postal and Charlotte Perkins Gilman’s The Yellow Wall-Paper. The scope of the thesis moves gradually outward from the text to the narrator to the reader. Chapters two and three follow on naturally from one another in that the fictional representation of speech and thought is a subsection of fictional point of view. In chapter four, linguistic analysis alone is shown to be incapable of accounting for reader response.
Linguistic description can help the critic to explain clearly and objectively how textual elements are foregrounded: the problem, naturally, is how to identify these elements in a manner that is both consistent and ratifiable. Gormenghast is useful for two reasons: first, it is arguable that the entire text is foregrounded, and second, Peake’s novel is opaque rather than transparent in that it solidly resists paraphrase. In making the attempt myself, some meanings had to be discarded and some senses lost. Wuthering Heights is also considered in chapter one to provide the contrast of a relatively transparent text. Linguistic theories have already proved useful as far as the fictional representation of speech and thought is concerned. Brian McHale argues that Free Indirect Discourse, like foregrounding, is one of the possible markers of a literary text. The effects of speech representation are subtle and these effects are lost in paraphrase. Pratchett moralises freely through his use of Free Indirect Discourse, and Going Postal is both interesting and useful because its major protagonist is a seasoned criminal, yet he presents the reader with the most moral of arguments. One expects morality from an upright character such as Sam Vimes, but not from Moist von Lipwig. Paul Simpson’s transitivity and modality systems provide easy and useful access to the study of point of view with the advantage that these systems force the student of literature to draw a conclusion based on textual evidence alone. Gilman’s narrator is unreliable and the reader is eventually alienated from her. I return to Wuthering Heights in the final chapter when I discuss literary allusion as part of my investigation into how a reader makes sense of a text.
CONTINUED IN CHAPTER ONE: FOREGROUNDING
 K. Green, ‘Literary Theory and Stylistics’, Encyclopedia of Language & Linguistics (2006), p. 261. Green notes that ‘elocutio dealt with the appropriateness of the expression and the relevance of its stylistic choices.’
 P. Simpson, Language, Ideology and Point of View (1993), p. 3.
 K. Green, ‘Literary Theory and Stylistics’, Encyclopedia of Language & Linguistics (2006), p. 265.
 P. Simpson, Language, Ideology and Point of View (1993), p. 5.
 D. Robey, ‘Modern Linguistics and the Language of Literature’, Modern Literary Theory: A Comparative Introduction (1986), p. 70.
 D. Robey, ‘Modern Linguistics and the Language of Literature’, Modern Literary Theory: A Comparative Introduction (1986), p. 71.
 I. Watt, ‘The First Paragraph of “The Ambassadors”: An Explication’, Essays in Criticism (1960), p. 253.
 R. Ohmann, ‘Generative Grammars and the Concept of Literary Style’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor), p. 262. Ohmann wrote in an earlier paper that the study of style had necessarily been inclined ‘to rely only on those impressionistic, metaphorical judgments which have too often substituted for analysis: dignified, grand, plain, decorative, placid, exuberant, restrained, hard, and the whole tired assortment of epithets which name without explaining.’ R. Ohmann, ‘Prolegomena to the Analysis of Prose Style’, Style in Prose Fiction (1958), H. C. Martin (editor), pp. 1-24.
 M. Toolan, The Stylistics of Fiction: A Literary-Linguistic Approach (1990), pp. 42-43.
 H. G. Widdowson, Stylistics and the Teaching of Literature (1975), R. Pope, Textual Intervention: Critical and Creative Strategies for Literary Studies (1995), and P. Simpson, Stylistics: A Resource Book for Students (2004). Stylistics textbooks have therefore been available since the mid-1970s at least, but Toolan wrote the words quoted above in 1990 (see footnote 9). Toolan’s words are really about the fact that stylistics had not at that time been widely adopted in English departments.
 K. Wales, ‘Stylistics’, Encyclopedia of Language & Linguistics (2006), p. 214.
 K. Green, ‘Literary Theory and Stylistics’, Encyclopedia of Language & Linguistics (2006), p. 261.
 I am indebted to Jim Miller for this information.
 Ohmann has defined style as ‘a characteristic use of language’ (‘Generative Grammars and the Concept of Literary Style’ (1964), p. 262), but critical essays on ‘style’ seem to be wide-ranging in their discussion, taking in everything from the social conditions surrounding the production of the work to the minutiae of the author’s psychology. In his essay ‘On the Style of Vanity Fair’ (reprinted in Style in Prose Fiction (1958), Harold C. Martin (editor), pp. 87-113), G. Armour Craig focuses his discussion on the distance Thackeray establishes between the novel’s narratorial voice and its characters; the reader is left to infer that which the narrator pretends not to know and does not tell. To be fair, Craig’s essay does not pretend to be about Thackeray’s style; the title makes reference to the style of this particular novel. Even so, what is under discussion here is surely a matter of technique – Craig’s essay would be helpful in a discussion centred on narrative point of view, but it tells us very little about Thackeray’s ‘characteristic use of language’. An essay by Albrecht Strauss reprinted in the same volume (‘On Smollett’s Language: A Paragraph in Ferdinand Count Fathom’, pp. 25-54) appears more promising. While Strauss’ essay is arguably nearer the mark, one is still left with a list of techniques – use of formulaic method, stock phrases and eighteenth-century periphrasis, recurrence of animal imagery, and so on – all of which would enable a reader to produce a Smollettian parody, but, as Strauss himself notes, does not ‘account for the robustness and verve of style which most readers of Smollett will rightly consider to be peculiarly his’ (pp. 49-50). Strauss suggests that Smollett’s voice is so difficult to isolate because ‘the persona of the detached narrator’ is largely absent. Strauss also links his observations to genre: he notes that in Smollett, what fails in pathos succeeds brilliantly in farcical comedy. Between them, Craig and Strauss discuss literary techniques, literary genre and narratorial point of view in their essays on ‘style’, but it could be argued that neither critic manages to identify the ‘characteristic use of language’ of the authors in question.
 G. Leech and M. Short, Style in Fiction (1981), chapter three, pp. 75-82.
 W. Iser, The Implied Reader (1974), introduction, p. xi. To quote from personal experience, I recently consulted a collection of critical essays in the hope that someone would be able to shed some light on the text of Franz Kafka’s The Trial (Twentieth Century Interpretations of ‘The Trial’: A Collection of Critical Essays (1976), J. Rolleston, editor). Instead I found that each critic interpreted the story according to his or her favoured theory. It did not escape my attention how easily Kafka’s text could be bent around any particular reading applied to it: a circumstance which may reveal how slippery the text is, and how difficult it is to pin any one interpretation to Joseph K’s ordeal. The most interesting reading by far was a semiotic investigation of Kafka’s text (T. M. Kavanagh, ‘Kafka’s “The Trial”: The Semiotics of the Absurd’, ibid., pp. 86-93), which pinpoints K’s inability to decipher the codes and signs.
 K. Green, ‘Literary Theory and Stylistics’, Encyclopedia of Language & Linguistics (2006), p. 262.
 P. Simpson, Language, Ideology and Point of View (1993), p. 4.
 This antagonism is incisively documented in chapter one of M. Toolan’s The Stylistics of Fiction: A Literary-Linguistic Approach (1990), pp. 1-27.
 R. Fowler quotes both Harold Whitehall and Roman Jakobson in his essay ‘The New Stylistics’, Style and Structure in Literature (1975), R. Fowler (editor) pp. 1-18. Whitehall wrote ‘as no science can go beyond mathematics, no criticism can go beyond its linguistics’, and Jakobson is quoted as follows: ‘poetics deals with problems of verbal structure…. Since linguistics is the global science of verbal structure, poetics may be regarded as an integral part of linguistics’ (p. 1).
 R. Fowler, ‘Linguistics, Stylistics; Criticism?’, Lingua, 16 (1966), pp. 157-158.
 H. H. Vendler, book review of Essays on Style and Language: Linguistic and Critical Approaches to Literary Styles (1966), R. Fowler (editor), Essays in Criticism, XVI (1966), p. 457.
 R. Fowler, ‘Language and the Reader: Shakespeare’s Sonnet 73’, Style and Structure in Literature (1975), pp. 79-122.
 P. Barry, ‘The Limitations of Stylistics’, Essays in Criticism, XXXVIII (1988), p. 182.
 Ibid., p. 181.
 ‘The more detailed forms of practical criticism’ would presumably exclude the practical criticism of I. A. Richards, which ultimately privileges reader response over exhaustive textual analysis.
 M. Short, Exploring the Language of Poems, Plays and Prose (1996), p. 6.
 Ibid., p. 356.
 M.A.K. Halliday recognises that ‘the only ultimately valid unit for textual analysis is the whole text’, but he also acknowledges the difficulties inherent in this conclusion: ‘it takes many hours of talking to describe exhaustively even the language of one sonnet’. M.A.K. Halliday, ‘Descriptive Linguistics in Literary Studies’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor), p. 58.
 An example of such an essay is D. Hymes’ ‘Phonological Aspects of Style: Some English Sonnets’, Style in Language (1960), T. Sebeok (editor), pp. 109-131.
 Mikhail Bakhtin argued that ‘the linguistic analysis of a poetic work has no criteria for separating what is poetically significant from what is not’. M. Bakhtin, ‘The Formal Method in Literary Scholarship’ (1928), The Bakhtin Reader (1994), P. Morris (editor), p. 145.
 R. Fowler, ‘The New Stylistics’, Style and Structure in Literature (1975), p. 6. Fowler quotes D. Freeman, ‘The Strategy of Fusion: Dylan Thomas’s Syntax’, ibid., p. 21.
 K. Wales, ‘Stylistics’, Encyclopedia of Language & Linguistics (2006), p. 215.
 M. Short, Exploring the Language of Poems, Plays and Prose (1996), p. 358.
 P. Simpson, Language, Ideology and Point of View (1993), p. 3.
 J. L. Mey, ‘Literary Pragmatics’, Encyclopedia of Language & Linguistics (2006), p. 256.
 S. Fish, ‘What Is Stylistics and Why Are They Saying Such Terrible Things About It?’ (1972), Approaches to Poetics (1973), S. Chatman (editor), p. 114. Fish also notes on page 113 of the same article that ‘a serious defect in the procedures of stylistics [is] the absence of any constraint on the way in which one moves from description to interpretation, with the result that any interpretation one puts forward is arbitrary.’
 J. Russell, ‘From Style to Meaning in “Araby”: Comment and Rebuttal’, College English, (1966), p. 170.
 Ibid. I would like to add that though meaning may or may not be ‘the final goal of the analyst’, it is not the final goal of the teacher. The teacher is far more interested in imparting an understanding of how the text works, in order to enable the student to formulate readings of other literary texts independently of his teacher. This is an area perhaps where a stylistic approach is demonstrably more useful than a traditional analysis.
 For example, M.A.K. Halliday’s analysis of ‘Leda and the Swan’ in ‘Descriptive Linguistics in Literary Studies’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor); Halliday makes two separate points about the grammar of the poem, but he does not pursue these points any further with reference to literary effect, or the poem’s meaning.
 R. Fowler, Linguistics and the Novel (1977), pp. 48-55.
 G. Leech and M. Short, Style in Fiction (1981), chapter 3, pp. 74-118. The texts studied are Joseph Conrad’s The Secret Sharer, D. H. Lawrence’s Odour of Chrysanthemums, and Henry James’ The Pupil.
 G. L. Dillon, ‘Whorfian Stylistics’, Journal of Literary Semantics, (1982), p. 75.
 R. Fowler, ‘Style and the Concept of Deep Structure’, Journal of Literary Semantics, (1972), p. 5.
 R. Ohmann, ‘Generative Grammars and the Concept of Literary Style’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor), p. 259.
 Fish, among others, found fault with the concept of transformational grammar, and in 1972 he wrote: ‘It is possible…to salvage the game…by making it more sophisticated, by contextualising it. One could simply write a rule that allows for the different valuings of the same pattern by taking into account the features which surround it in context. But this would only lead to the bringing forward of further counterexamples and the continual and regressive rewriting of the rule. Eventually a point would be reached where a separate rule was required for each and every occurrence’. ‘What Is Stylistics and Why Are They Saying Such Terrible Things About It?’ (1972), Approaches to Poetics (1973), S. Chatman (editor), pp. 120-121.
 M. Short, Exploring the Language of Poems, Plays and Prose (1996), p. 6.
 While I acknowledge that there probably should be scare quotes around the words ‘literary’ and ‘literature’, I have not typed them because I feel that their constant appearance would be tedious for the reader; I prefer instead to state once and for all that I realise these are highly subjective terms.
 P. Simpson, Language, Ideology and Point of View (1993), p. 3.
 M. A. K. Halliday, ‘Descriptive Linguistics in Literary Studies’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor), p. 67.
 Bakhtin uses the term ‘speech genres’ for register in his essay ‘The Problem of Speech Genres’ (1952-1953). In her introduction to The Bakhtin Reader Pam Morris writes that ‘the most interesting new area in this essay is the suggestion that the form of utterances is not a matter of free choice on the part of any individual speaker…. Bakhtin argues that speech genres impose an order and form on everyday speech in ways we are largely unaware of, but which can have a considerable effect upon our speech flexibility and ease’. P. Morris, The Bakhtin Reader (1994), p. 17. In other words, we do not have as much freedom as we suppose in the construction of our individual utterances.
 Halliday also developed a method to help in the characterisation of register based on the field, tenor and mode of the text in question, where field, or domain, is the subject matter, tenor relates to the relationship between the participants and the formality of the situation, and mode concerns the type of channel –whether the text is written (a letter) or spoken (a radio play), for example – and the organisation of the text. Field, tenor and mode correspond approximately to Halliday’s three categories of language function: ideational, interpersonal and textual respectively. Halliday makes what I think is an important contribution to the form/content debate here. He notes that ‘ “an early announcement is expected”…and “apologies for absence were received”…are not simply free variants of “we ought to hear soon” and “was sorry he couldn’t make it”.’ (The Linguistic Sciences and Language Teaching (1964), p. 87, quoted in R. Fowler, Linguistic Criticism (1996), pp. 190-191.) If a different form triggers a different register, then the words spoken will surely be understood differently by the hearer/reader, in that the words of the message will be located in a different situation type. This, then, is one argument in favour of the inseparability of form and content.
 R. Fowler, Linguistic Criticism (1996), p. 204.
 R. Carter and W. Nash, ‘Language and Literariness’, Prose Studies (1983), p. 139.
 S. Smith, Novel on Yellow Paper (1936), p. 35.
 Fowler argues that when a text contains numerous registers, the point of association between the varieties has to be worked out, and therefore heteroglossia is intrinsically connected with defamiliarisation. R. Fowler, Linguistic Criticism (1996), p. 197.
 T. Eagleton, Literary Theory: An Introduction (1983), p. 2.
 Ibid., p. 6.
 G. Leech and M. Short, Style in Fiction (1981), p. 29: a text is ‘opaque in the sense that the medium attracts attention in its own right; and indeed, the interpretation of sense may be frustrated and obstructed by abnormalities in the use of the lexical and grammatical features of medium.’
 B. Burton and R. Carter, ‘Literature and the Language of Literature’, Encyclopedia of Language & Linguistics (2006), p. 267. Eagleton also recognises the truth of this argument: ‘Anything can be literature, and anything which is regarded as unalterably and unquestionably literature – Shakespeare, for example – can cease to be literature…. Literature, in the sense of a set of works of assured and unalterable value, distinguished by certain shared inherent properties, does not exist…. Just as people may treat a work as philosophy in one century and as literature in the next, or vice versa, so they may change their minds about what writing they consider valuable. They may even change their minds about the grounds they use for judging what is valuable and what is not.’ T. Eagleton, Literary Theory: An Introduction (1983), pp. 10-11.
 R. Fowler, Linguistic Criticism (1996), p. 205.
 M. Bakhtin, ‘The Problem of Speech Genres’ (1952-1953), The Bakhtin Reader (1994), P. Morris (editor), p. 82.
 Fish offers a similar argument to that proposed by Carter and Nash in his article ‘How Ordinary is Ordinary Language?’, New Literary History (1973).
 R. Carter and W. Nash, ‘Language and Literariness’, Prose Studies (1983), p. 130.
 B. Burton and R. Carter, ‘Literature and the Language of Literature’, Encyclopedia of Language & Linguistics (2006), p. 273.
 Cf. Wittgenstein and polythetic categories of ‘family resemblances’.
 David Lodge has since reassessed his position, however, and he notes in the afterword to the second edition of Language in Fiction (1984), p. 296, that he now considers himself an advocate of pluralism in the sense described by Leech and Short in Style in Fiction (1981), chapter one, pp. 29-34.
 C. Brooks, ‘The Heresy of Paraphrase’, The Well-Wrought Urn: Studies in the Structure of Poetry (1947), pp. 192-214.
 P. Barry, ‘The Enactment Fallacy’, Essays in Criticism (1980), p. 95.
 Mervyn Peake, 1911-1968; although mostly known for his work as an artist and illustrator Peake is also the author of the Gormenghast trilogy, a work best described as a Gothic fantasy.
 Bakhtin refers to the literary work as an ‘utterance’, regardless of the length of that work, and he asserts that ‘thematic content, style, and compositional structure…are inseparably linked to the whole of the utterance and are equally determined by the specific nature of the particular sphere of communication’. M. Bakhtin, ‘The Problem of Speech Genres’ (1952-1953), The Bakhtin Reader (1994), P. Morris (editor), p. 81. Thus ‘form’ is the whole text and ‘content’ is the form plus its context.
 R. Fowler, Linguistics and the Novel (1977), p. 13.
 R. Fowler, Linguistics and the Novel (1977), p. 20.
 An oft-quoted example of an ambiguous phrase is that of Hockett’s telegram, which reads ‘Ship sails today’. Is ship a noun (agent) or a verb (action, imperative)?
 R. Fowler, Linguistics and the Novel (1977), p. 11.
 Ibid., p. 22.
 R. Fowler, Linguistics and the Novel (1977), p. 17.
 Ibid., p. 11.
 D. Keyes, Flowers for Algernon (1966).
 R. Fowler, Linguistics and the Novel (1977), p. 11.
 R. Fowler, ‘Style and the Concept of Deep Structure’, Journal of Literary Semantics (1972).
 If the phrase structure component represents obligatory transformations (the fixed element), transformational rules are optional (the variable component).
 R. Fowler, ‘Style and the Concept of Deep Structure’, Journal of Literary Semantics (1972), p. 16.
 I shall comment further on the ramifications of individual word meaning in chapter four.
 R. Fowler, ‘Style and the Concept of Deep Structure’, Journal of Literary Semantics (1972), p. 8.
 R. Ohmann, ‘Generative Grammars and the Concept of Literary Style’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor), p. 267.
 Ibid., p. 264.
 R. Ohmann, ‘Generative Grammars and the Concept of Literary Style’ (1964), Linguistics and Literary Style (1970), D. C. Freeman (editor), p. 266.
 Ibid., p. 268.
 R. Ohmann, ‘Literature as Sentences’, College English (1966).
 D. H. Hirsch, ‘Linguistic Structure and Literary Meaning’, Journal of Literary Semantics (1972), p. 88.
 See Appendix C.
 B. McHale, ‘Free Indirect Discourse: A Survey of Recent Accounts’, Poetics and Theory of Literature (1978).
 Samuel Vimes is Commander of the City Watch of Ankh-Morpork and a character who appears regularly in the Discworld series.
 See Appendices E and F.