Brian McFarlane; Clarendon Press, 1996, Oxford
Part I Backgrounds, Issues, and a New Agenda
Conrad, Griffith, and 'Seeing'
Commentators in the field are fond of quoting Joseph Conrad's famous statement of his novelistic intention: 'My task which I am trying to achieve is, by the powers of the written word, to make you hear, to make you feel-it is, before all, to make to see'.1 This remark of 1897 is echoed, consciously
1 Joseph Conrad, Preface to The Nigger of the Narcissus ( J. M. Dent and Sons: London, 1945), 5.
or otherwise, sixteen years later by D. W. Griffith, whose cinematic intention is recorded as: 'The task I am trying to achieve is above all to make you see'.2 George Bluestone's all-but-pioneering work in the film-literature field, Novels into Film, draws attention to the similarity of the remarks at the start of his study of 'The Two Ways of Seeing' , claiming that 'between the percept of the visual image and the concept of the mental image lies the root difference between the two media?'.3 In this way he acknowledges the connecting link of 'seeing' in his use of the word 'image'. At the same time, he points to the fundamental difference between the way images are produced in the two media and how they are received. Finally, though, he claims that 'conceptual images evoked by verbal stimuli can scarcely be distinguished in the end from those evoked by non-verbal stimuli',4 and, in this respect, he shares common ground with several other writers concerned to establish links between the two media.
By this, I mean those commentaries which address themselves to crucial changes in the (mainly English) novel towards the end of the nineteenth century; changes which led to a stress on showing rather than on telling and which, as a result, reduced the element of authorial intervention in its more overt manifestations. Two of the most impressive of such accounts, both of them concerned with ongoing processes of transmutation among the arts, notably between literature and film, are Alan Spiegel Fiction and the Camera Eye5 and Keith Cohen Film and Fiction.6 Spiegel's avowed purpose is to investigate 'the common body of thought and feeling that unites film form with the modern novel',7 taking as his starting-point Flaubert, whom he sees as the first great nineteenth-century exemplar of 'concretized form', a form dependent on supplying a great deal of visual information. His line of enquiry leads him to James Joyce who, like Flaubert, respects 'the integrity of the seen object and . . . gives it palpable presence apart from the presence of the observer'.8 This line is pursued by way of Henry James who attempts 'a balanced distribution of emphasis in the rendering of what is looked at, who is looking, and what the looker makes of what she [i.e. Maisie in What Maisie Knew] sees',9 and by way of the Conrad--Griffith comparison. Spiegel presses this comparison harder than Bluestone, stressing that though both may have aimed at the same point--a congruence of image and concept--they did so from opposite directions. Whereas Griffith used his images to tell a story, as means to understanding, Conrad (Spiegel claims) wanted the reader to
2 Quoted in Lewis Jacobs, The Rise of the American Film (Harcourt, Brace: New York, 1939), 119.
3 George Bluestone. Novels into Film (University of California Press: Berkeley and Los Angeles, 1957), 1.
4 Ibid. 47.
5 Alan Spiegel, Fiction and the Camera Eye: Visual Consciousness in Film and the Modern Novel (University Press of Virginia: Charlottesville, 1976).
6 Keith Cohen, Film and Fiction: The Dynamics of Exchange (Yale University Press: New Haven, 1979).
7 Spiegel, Fiction and the Camera Eye, p. xiii.
8 Ibid. 63.
9 Ibid. 55.
'''see" in and through and finally past his language and his narrative concept to the hard, clear bedrock of images'.10
One effect of this stress on the physical surfaces and behaviours of objects and figures is to de-emphasize the author's personal narrating voice so that we learn to read the ostensibly unmediated visual language of the later nineteenth-century novel in a way that anticipates the viewer's experience of film which necessarily presents those physical surfaces. Conrad and James further anticipate the cinema in their capacity for 'decomposing' a scene, for altering point of view so as to focus more sharply on various aspects of an object, for exploring a visual field by fragmenting it rather than by presenting it scenographically (i.e. as if it were a scene from a stage presentation).
Cohen, concerned with the 'process of convergence' between art-forms, also sees Conrad and James as significant in a comparison of novels and film. These authors he sees as breaking with the representational novels of the earlier nineteenth century and ushering in a new emphasis on 'showing how the events unfold dramatically rather than recounting them'.11 The analogy with film's narrative procedures will be clear and there seems no doubt that film, in turn, has been highly influential on the modern novel. Cohen uses passages from Proust and Virginia Woolf to suggest how the modern novel, influenced by techniques of Eisensteinian montage cinema, draws attention to its encoding processes in ways that the Victorian novel tends not to.
Dickens, Griffith, and Story-Telling
The other comparison that trails through the writing about film and literature is that between Griffith and Dickens, who was said to be the director's favourite novelist. The most famous account is, of course, that of Eisenstein, who compares their 'spontaneous childlike skill for story-telling',12 a quality he finds in American cinema at large, their capacity for vivifying 'bit' characters, the visual power of each, their immense popular success, and above all their rendering of parallel action, for which Griffith cited Dickens as his source. On the face of it, there now seems nothing so remarkable in these formulations to justify their being so frequently paraded as examples of the ties that bind cinema and the Victorian novel. In fact Eisenstein's discussion of Dickens's 'cinematic techniques', including anticipation of such phenomena as frame composition and the close-up, is really not far removed from those many works which talk about film language, striking similar analogical poses, without giving adequate consideration to the qualitative differences enjoined by the two media.
Later commentators have readily embraced Eisenstein's account:
10 Ibid. pp. xi-xii.
11 Cohen, Film and Fiction, 5
12 Sergei Eisenstein, Film Form, ed. and trans. Jan Leyda ( Harcourt, Brace: New York, 1949), 196.
Bluestone, for instance, states boldly that: 'Griffith found in Dickens hints for every one of his major innovations',13 and Cohen, going further, points to 'the more or less blatant appropriation of the themes and content of the nineteenth-century bourgeois novel'.14 However, in spite of the frequency of reference to the Dickens--Griffith connection, and apart from the historical importance of parallel editing in the development of film narrative, the influence of Dickens has perhaps been overestimated and under-scrutinized. One gets the impression that critics steeped in a literary culture have fallen on the Dickens-Griffith comparison with a certain relief, perhaps as a way of arguing the cinema's respectability. They have tended to concentrate on the thematic interests and the large, formal narrative patterns and strategies the two great narrative-makers shared, rather than to address themselves, as a film-oriented writer might, to detailed questions of enunciation, of possible parallels and disparities between the two different signifying systems, of the range of 'functional equivalents'15 available to each within the parameters of the classical style as evinced in each medium.
As film came to replace in popularity the representational novel of the earlier nineteenth century, it did so through the application of techniques practised by writers at the latter end of the century. Conrad with his insistence on making the reader 'see' and James with his technique of 'restricted consciousness', both playing down obvious authorial mediation in favour of limiting the point of view from which actions and objects are observed, provide dear examples. In this way they may be said to have broken with the tradition of 'transparency' in relation to the novel's referential world so that the mode and angle of vision were as much a part of the novel's content as what was viewed. The comparisons with cinematic technique are clear but, paradoxically, the modern novel has not shown itself very adaptable to film. However persuasively it may be demonstrated that the likes of Joyce, Faulkner, and Hemingway have drawn on cinematic techniques, the fact is that the cinema has been more at home with novels from--or descended from--an earlier period. Similarly, certain modern plays, such as Death of a Salesman, Equus, or M. Butterfly, which seem to owe something to cinematic techniques, have lost a good deal of their fluid representations of time and space when transferred to the screen.
Adaptation: The Phenomenon
As soon as the cinema began to see itself as a narrative entertainment, the idea of ransacking the novel--that already established repository of narrative fiction--for source material got underway, and the process has continued
13 Bluestone, Novels into Film, 2.
14 Cohen, Film and Fiction, 4.
15 David Bordwell term, in The Classical Hollywood Cinema ( Routledge and Kegan Paul: London, 1985), 13.
more or less unabated for ninety years. Film-makers' reasons for this continuing phenomenon appear to move between the poles of crass commercialism and high-minded respect for literary works. No doubt there is the lure of a pre-sold title, the expectation that respectability or popularity achieved in one medium might infect the work created in another. The notion of a potentially lucrative 'property' has clearly been at least one major influence in the filming of novels, and perhaps film-makers, as Frederic Raphael scathingly claims, 'like known quantities . . . they would sooner buy the rights of an expensive book than develop an original subject'.16 Nevertheless most of the film-makers on record profess loftier attitudes than these. DeWitt Bodeen, co-author of the screenplay for Peter Ustinov Billy Budd (1962), claims that: 'Adapting literary works to film is, without a doubt, a creative undertaking, but the task requires a kind of selective interpretation, along with the ability to recreate and sustain an established mood'.17 That is, the adaptor should see himself as owing allegiance to the source work. Despite Peter Bogdanovich's disclaimer about filming Henry James Daisy Miller ('I don't think it's a great classic story. I don't treat it with that kind of reverence'18), for much of the time the film is a conscientious visual transliteration of the original. One does not find film-makers asserting a bold approach to their source material, any more than announcing crude financial motives.
As to audiences, whatever their complaints about this or that violation of the original, they have continued to want to see what the books 'look like'. Constantly creating their own mental images of the world of a novel and its people, they are interested in comparing their images with those created by the film-maker. But, as Christian Metz says, the reader 'will not always find his film, since what he has before him in the actual film is now somebody else's phantasy'.19 Despite the uncertainty of gratification, of finding audiovisual images that will coincide with their conceptual images, reader-viewers persist in providing audiences for 'somebody else's phantasy'. There is also a curious sense that the verbal account of the people, places, and ideas that make up much of the appeal of novels is simply one rendering of a set of existents which might just as easily be rendered in another. In this regard, one is reminded of Anthony Burgess's cynical view that 'Every best-selling novel has to be turned into a film, the assumption being that the book itself whets an appetite for the true fulfilment--the verbal shadow turned into light, the word made flesh.'20 And perhaps there is a parallel with that late
16 Frederic Raphael, "'Introduction'", Two for the Road (Jonathan Cape: London, 1967).
17 DeWitt Bodeen, "'The Adapting Art'", Films in Review, 14/6 (June-July 1963), 349.
18 Jan Dawson, "'The Continental Divide: Filming Henry James'", Sight and Sound, 43/1 (Winter 1973-4), 14; repr. in part as 'An Interview with Peter Bogdanovich' in G. Peary and R. Shatzkin (eds.), The Classic American Novel and the Movies (Frederick Ungar Publishing: New York, 1977).
19 Christian Metz, The Imaginary Signifier (Indiana University Press: Bloomingdale, 1977), 12.
20 Anthony Burgess, "'On the Hopelessness of Turning Good Books into Films'", New York Times, 20 Apr 1975, p. 15.
nineteenth-century phenomenon, described by Michael Chanan in The Dream that Kicks, of illustrated editions of literary works and illustrated magazines in which great novels first appeared as serials. There is, it seems, an urge to have verbal concepts bodied forth in perceptual concreteness.
Whatever it is that makes film-goers want to see adaptations of novels, and film-makers to produce them, and whatever hazards lie in the path for both, there is no denying the facts. For instance, Morris Beja reports that, since the inception of the Academy Awards in 1927-8, 'more than threefourths of the awards for "best picture" have gone to adaptations . . . [and that] the all-time box-office successes favour novels even more'.21 Given that the novel and the film have been the most popular narrative modes of the nineteenth and twentieth centuries respectively, it is perhaps not surprising that film-makers have sought to exploit the kinds of response excited by the novel and have seen in it a source of ready-made material, in the crude sense of pre-tested stories and characters, without too much concern for how much of the original's popularity is intransigently tied to its verbal mode.
The Discourse on Adaptation
On being faithful
Is it really 'Jamesian'? Is it 'true to Lawrence'? Does it 'capture the spirit of Dickens'? At every level from newspaper reviews to longer essays in critical anthologies and journals, the adducing of fidelity to the original novel as a major criterion for judging the film adaptation is pervasive. No critical line is in greater need of re-examination--and devaluation.
Discussion of adaptation has been bedevilled by the fidelity issue, no doubt ascribable in part to the novel's coming first, in part to the ingrained sense of literature's greater respectability in traditional critical circles. As long ago as the mid-1940s James Agee complained of a debilitating reverence in even such superior transpositions to the screen as David Lean Great Expectations. It seemed to him that the really serious-minded film-goer's idea of art would be 'a good faithful adaptation of Adam Bede in sepia, with the entire text read offscreen by Herbert Marshall'.22 However, voices such as Agee's, querulously insisting that the cinema make its own art and to hell with tasteful allegiance, have generally cried in the wilderness.
Fidelity criticism depends on a notion of the text as having and rendering up to the (intelligent) reader a single, correct 'meaning' which the filmmaker has either adhered to or in some sense violated or tampered with. There will often be a distinction between being faithful to the 'letter', an
21 Morris Beja, Film and Literature ( Longman: New York, 1979), 78.
22 Agee on Film (McDowell Oblonsky: New York, 1958), 216.
approach which the more sophisticated writer may suggest is no way to ensure a 'successful' adaptation, and to the 'spirit' or 'essence' of the work. The latter is of course very much more difficult to determine since it involves not merely a parallelism between novel and film but between two or more readings of a novel, since any given film version is able only to aim at reproducing the film-maker's reading of the original and to hope that it will coincide with that of many other readers/viewers. Since such coincidence is unlikely, the fidelity approach seems a doomed enterprise and fidelity criticism unilluminating. That is, the critic who quibbles at failures of fidelity is really saying no more than: 'This reading of the original does not tally with mine in these and these ways.'
Few writers on adaptation have specifically questioned the possibility of fidelity; though some have claimed not to embrace it, they still regard it as a viable choice for the film-maker and a criterion for the critic. Beja is one exception. In asking whether there are 'guiding principles' for film-makers adapting literature, he asks: 'What relationship should a film have to the original source? Should it be "faithful"? Can it be? To what?'23
When Beja asks 'To what' should a film-maker be faithful in adapting a novel, one is led to recall those efforts at fidelity to times and places remote from present-day life. In 'period' films, one often senses exhaustive attempts to create an impression of fidelity to, say Dickens's London or to Jane Austen's village life, the result of which, so far from ensuring fidelity to the text, is to produce a distracting quaintness. What was a contemporary work for the author, who could take a good deal relating to time and place for granted, as requiring little or no scene-setting for his readers, has become a period piece for the film-maker. As early as 1928, M. Willson Disher picked up the scent of this misplaced fidelity in writing about a version of Robinson Crusoe: 'Mr Wetherell [director, producer, writer and star] went all the way to Tobago to shoot the right kinds of creeks and caves, but he should have travelled not westwards, but backwards, to reach "the island", and then he would have arrived with the right sort of luggage.’24 Disher is not speaking against fidelity to the original as such but against a misconstrued notion of how it might be achieved. A more recent example is Peter Bogdanovich's use of the thermal baths sequence in his film of Daisy Miller. 'The mixed bathing is authentically of the period', he claims in an interview with Jan Dawson.25 Authentically of the period, perhaps, but not so of Henry James, so that it is only a tangential, possibly irrelevant fidelity that is arrived at. The issue of fidelity is a complex one but it is not too gross a simplification to suggest that critics have encouraged film-makers to see it as a desirable goal in the adaptation of literary works. As Christopher Orr has noted: 'The concern
23 Beja, Film and Literature. 80.
24 M. Willson Disher, "'Classics into Films'", Fortnightly Review, NS 124 ( Dec 1928), 789.
25 Dawson, 'The Continental Divide' , 14.
with the fidelity of the adapted film in letter and spirit to its literary source has unquestionably dominated the discourse on adaptation.'26 The issue is inevitably raised in each of the succeeding Case-Studies, and is the object of a Special Focus in the study of Daisy Miller, which offers revealing insights into the limits of fidelity, especially from the film-maker's point of view.
Obscuring other issues
The insistence on fidelity has led to a suppression of potentially more rewarding approaches to the phenomenon of adaptation. It tends to ignore the idea of adaptation as an example of convergence among the arts, perhaps a desirable--even inevitable--process in a rich culture; it fails to take into serious account what may be transferred from novel to film as distinct from what will require more complex processes of adaptation; and it marginalizes those production determinants which have nothing to do with the novel but may be powerfully influential upon the film. Awareness of such issues would be more useful than those many accounts of how films 'reduce' great novels.
Modern critical notions of intertextuality represent a more sophisticated approach, in relation to adaptation, to the idea of the original novel as a 'resource'. As Christopher Orr remarks: 'Within this critical context [i.e. of intertextuality], the issue is not whether the adapted film is faithful to its source, but rather how the choice of a specific source and how the approach to that source serve the film's ideology.'27 When, for instance, MGM filmed James Hilton 1941 bestseller, Random Harvest, in the following year, its images of an unchanging England had as much to do with Hollywood antiisolationism with regard to World War II as with finding visual equivalents for anything in Hilton. The film belongs to a rich context created by notions of Hollywood's England, by MGM's reputation for prestigious literary adaptation and for a glossy 'house style', by the genre of romantic melodrama (cf. Rebecca, 1940, This Above All, 1942), and by the idea of the star vehicle. Hilton's popular but, in truth, undistinguished romance is but one element of the film's intertextuality.
Some writers have proposed strategies which seek to categorize adaptations so that fidelity to the original loses some of its privileged position. Geoffrey Wagner suggests three possible categories which are open to the film-maker and to the critic assessing his adaptation: he calls these ( a ) transposition, 'in which a novel is given directly on the screen with a minimum of apparent interference';28 ( b ) commentary, 'where an original is taken and either purposely or inadvertently altered in some respect . . . when there has
26 Christopher Orr, "'The Discourse on Adaptation'", Wide Angle, 6/2 (1984),
28 Geoffrey Wagner, The Novel and the Cinema (Fairleigh Dickinson University Press: Rutherford, NJ, 1975), 222.
been a different intention on the part of the film-maker, rather than infidelity or outright violation';29 and ( c ) analogy, which must represent a fairly considerable departure for the sake of making another work of art'.30 The critic, he implies, will need to understand which kind of adaptation he is dealing with if his commentary on an individual film is to be valuable. Dudley Andrew also reduces the modes of relation between the film and its source novel to three, which correspond roughly (but in reverse order of adherence to the original) to Wagner's categories: 'Borrowing, intersection, and fidelity of transformation'.31 And there is a third comparable classification system put forward by Michael Klein and Gillian Parker: first, 'fidelity to the main thrust of the narrative'; second, the approach which 'retains the core of the structure of the narrative while significantly reinterpreting or, in some cases, deconstructing the source text'; and, third, regarding 'the source merely as raw material, as simply the occasion for an original work'.32 The parallel with Wagner's categories is clear.
There is nothing definitive about these attempts at classification but at least they represent some heartening challenges to the primacy of fidelity as a critical criterion. Further, they imply that, unless the kind of adaptation is identified, critical evaluation may well be wide of the mark. The faithful adaptation (e.g. Daisy Miller or James Ivory Howard's End, 1992) can certainly be intelligent and attractive, but is not necessarily to be preferred to the film which sees the original as 'raw material' to be reworked, as Hitchcock so persistently did, from, say, Sabotage (1936) to The Birds (1963). Who, indeed, ever thinks of Hitchcock as primarily an adaptor of other people's fictions? At a further extreme, it is possible to think of a film as providing a commentary on a literary text, as Welles does on three Shakespearian plays in Chimes at Midnight (1966), or as Gus Van Sant does in My Own Private Idaho (1992), drawing on both Shakespeare and Welles. There are many kinds of relations which may exist between film and literature, and fidelity is only one--and rarely the most exciting.
REDEFINING ISSUES AND A NEW APPROACH
The Centrality of Narrative
The more one considers the phenomenon of adaptation of novel into film the whole history of the reliance on the novel as source material for the fiction film--the more one is drawn to consider the central importance of
29 Ibid. 224.
30 Ibid. 226.
31 Dudley Andrew, 'The Well-Worn Muse: Adaptation in Film History and Theory' , in Syndy Conger and Janice R. Welsch (eds.), Narrative Strategies ( West Illinois University Press: Macomb, Ill., 1980), 10.
32 Michael Klein and Gillian Parker (eds.), The English Novel and the Movies ( Frederick Ungar Publishing: New York, 1981), 9-10.
narrative to both. Whatever the cinema's sources--as an invention, as a leisure pursuit, or as a means of expression--and whatever uncertainties about its development attend its earliest years, its huge and durable popularity is owed to what it most obviously shares with the novel. That is, its capacity for narrative. By the time of Edwin Porter The Great Train Robbery (1903), in which scenes set in different locations are spliced together to tell a story, the cinema's future as a narrative art was settled, and no subsequent development of its techniques has threatened the supremacy of that function.
Christian Metz, discussing film narrativity, writes: 'Film tells us continuous stories; it "says" things that could be conveyed also in the language of words; yet it says them differently. There is a reason for the possibility as well as for the necessity of adaptations.'33 He goes on to consider the 'demand' for the feature-length fiction film, 'which was only one of the many conceivable genres',34 but which has dominated film production. 'The basic formula, which has never changed, is the one that consists in making a large continuous unit that tells a story and calling it a "movie". "Going to the movies" is going to see this type of story.'35 Whatever other uses the cinema might have found, it is, as Metz suggests, as a story-teller that it found its greatest power and its largest audience. Its embourgeoisement inevitably led it away from trick shows, the recording of music halls acts and the like, towards that narrative representationalism which had reached a peak in the classic nineteenth-century novel. If film did not grow out of the latter, it grew towards it; and what novels and films most strikingly have in common is the potential and propensity for narrative. And narrative, at certain levels, is undeniably not only the chief factor novels and the films based on them have in common but is the chief transferable element.
If one describes a narrative as a series of events, causally linked, involving a continuing set of characters which influence and are influenced by the course of events, one realizes that such a description might apply equally to a narrative displayed in a literary text and to one in a filmic text. Nevertheless, much of the dissatisfaction which accompanies the writing about films adapted from novels tends to spring from perceptions of 'tampering' with the original narrative. Words like 'tampering' and 'interference', and even 'violation', give the whole process an air of deeply sinister molestation, perhaps springing from the viewer's thwarted expectations relating to both character and event. Such dissatisfactions resonate with a complex set of misapprehensions about the workings of narrative in the two media, about the irreducible differences between the two, and from a failure to distinguish what can from what cannot be transferred.
33 Christian Metz, Film Language: A Semiotics of the Cinema, trans. Michael Taylor ( Oxford University Press: New York, 1974), 44.
35 Ibid. 45.
To begin with the last point: there is a distinction to be made between what may be transferred from one narrative medium to another and what necessarily requires adaptation proper. Throughout the rest of this study, 'transfer' will be used to denote the process whereby certain narrative elements of novels are revealed as amenable to display in film, whereas the widely used term 'adaptation' will refer to the processes by which other novelistic elements must find quite different equivalences in the film medium, when such equivalences are sought or are available at all.
Narrative Functions: Novel and Film
Roland Barthes has defined the essence of a narrative function as 'the seed that it sows in the narrative, planting an element that will come to fruition later--either on the same [narrative] level or elsewhere, on another level',36 going on to claim that, 'A narrative is never made up of anything other than functions: in differing degrees, everything in it signifies.' He distinguishes two main groups of narrative functions: distributional and integrational and, though he is not concerned with cinema in this discussion, this distinction is valuable in sorting out what may be transferred (i.e. from novel to film) from that which may only be adapted. To distributional functions, Barthes gives the name of functions proper ; integrational functions he calls indices. The former refer to actions and events; they are 'horizontal' in nature, and they are strung together linearly throughout the text; they have to do with 'operations'; they refer to a functionality of doing. Indices denotes a 'more or less diffuse concept which is nevertheless necessary to the meaning of the story'.37 This concept embraces, for instance, psychological information relating to characters, data regarding their identity, notations of atmosphere and representations of place. Indices are 'vertical' in nature, influencing our reading of narrative in a pervasive rather than a linear way; they do not refer to operations but to a functionality of being.
The most important kinds of transfer possible from novel to film are located in the category of functions proper, rather than that of indices, though some elements of the latter will also be seen to be (partly) transferable. Barthes further subdivides functions to include cardinal functions (or nuclei) and catalysers. Cardinal functions are the 'hinge-points' of narrative: that is, the actions they refer to open up alternatives of consequence to the development of the story; they create 'risky' moments in the narrative and it is crucial to narrativity ('the processes through which the reader . . . constructs the meaning of the text'38) that the reader recognizes the possibility of
36 Roland Barthes, 'Introduction to the Structural Analysis of Narratives' ( 1966), in Image-Music-Text, trans. Stephen Heath ( Fontana/Collins: Glasgow, 1977). 89.
37 Ibid. 92.
38 Orr, 'The Discourse on Adaptation' , 73.
such alternative consequences. The linking together of cardinal functions provides the irreducible bare bones of the narrative, and this linking, this 'tie between two cardinal functions, is invested with a double functionality, at once chronological and logical'.39 These cardinal functions, or, in Seymour Chatman's terms, kernels ('narrative moments that give rise to cruxes in the direction taken by events'40), are, as I shall show, transferable: when a major cardinal function is deleted or altered in the film version of a novel (e.g. to provide a happy rather than a sombre ending), this is apt to occasion critical outrage and popular disaffection. The film-maker bent on 'faithful' adaptation must, as a basis for such an enterprise, seek to preserve the major cardinal functions.
However, even if the latter are preserved in the filming process, they can be 'deformed' by varying the catalysers which surround them. Catalysers (in Chatman's term, satellites) work in ways which are complementary to and supportive of the cardinal functions. They denote small actions (e.g. the laying of the table for a meal which may in turn give rise to action of cardinal importance to the story); their role is to root the cardinal functions in a particular kind of reality, to enrich the texture of those functions: 'their functionality is attenuated, unilateral, parasitic: it is a question of a purely chronological functionality', 41 in Barthes's words. Unlike the 'risky moments' created by cardinal functions, the catalysers 'lay out areas of safety, rests, luxuries';42 they account for the moment-to-moment minutiae of narrative.
In so far as these functions, whether cardinal or catalysing, are not dependent on language, in the sense that they denote aspects of story content (actions and happenings) which may be displayed verbally or audio-visually, they are directly transferable from one medium to the other. Among the integrational functions, which Barthes subdivides into indices proper and informants, only the latter may be directly transferred. Whereas the former relate to concepts such as character and atmosphere, are more diffuse than the functions proper, and are therefore more broadly open to adaptation rather than to the comparative directness of transfer, informants 'are pure data with immediate signification'.43 They include 'ready-made knowledge' such as the names, ages, and professions of characters, certain details of the physical setting, and, in these senses and in their own ways, share the authenticating and individuating functions performed in other respects by catalysers, and they are often amenable to transfer from one medium to another. What Barthes designates as cardinal functions and catalysers consti-
39 Barthes, 'Introduction to the Structural Analysis of Narratives' , 94.
40 Seymour Chatman, Story and Discourse: Narrative Structure in Fiction and Film ( Cornell University Press: Ithaca, NY, 1978), 53.
41 Barthes, 'Introduction to the Structural Analysis of Narratives' , 94.
42 Ibid. 95.
43 Ibid. 96.
tutes the formal content of narrative which may be considered independently of what Chatman calls 'its manifesting substance' (e.g. novel or film), and informants, in their objective name-ability, help to embed this formal content in a realized world, giving specificity to its abstraction. Perhaps informants may be seen as a first, small step towards mimesis in novel and film, the full mimetic process relying heavily on the functioning of the indices proper, to which I shall return shortly.
I should note at this point that Barthes has subsequently modified the structural taxonomy set up here with his designation of the five narrative codes which structure all classical narrative in S/Z, his reading of Balzac Sarrasine. For my purposes, the earlier distinction between distributional and integrational functions, with the metaphors implied in their characterization, provides a more accessible and usable taxonomy in establishing what may be transferred from a long, complex work in one medium to a long, complex work in another. Barthes was not, of course, concerned with cinematic narrative when he wrote his 'Structural Analysis' essay, but, in Robin Wood's words, 'the critic has the right to appropriate whatever's/he needs from wherever it can be found, and use it for purposes somewhat different from the original ones'.44
Kinds of Narration and their Cinematic Potential
The distinctions to be drawn between various narrational modes as they appear in the novel are difficult to sustain in film narrative. The novels chosen as case-studies in this book exhibit notably different approaches to the question of narrative point of view: for example, first-person, omnis-cient, a mixture of both, the use of 'restricted consciousness'. However, these different approaches are considerably elided in the narrational procedures adopted by the films, a matter to be investigated in detail in Part II of this study. It is sufficient to draw attention at this point to the varying amenability to cinematic practice of these kinds of literary narration.
The first-person narration
There is only a precarious analogy between the attempts at first-person narration offered by films and the novel's first-person narration, comprising the individual discourses of each character surrounded by a continuing (generally past-tense) discourse which is attributed to a known and named narrator who may or may not be an active participant in the events of the novel. These attempts will usually be one of two kinds.
44 Robin Wood, 'Notes for a Reading of I Walked with a Zombie' , CineAction! , 3-4 ( Winter 1986), 9.
(a) The subjective cinema The subjective cinema on the scale of The Lady in the Lake (1946) has scarcely been tried since, in mainstream film-making at least, and has the status of a curiosity rather than of a major contribution to screen practice. Of its more localized manifestations (e.g. the point-of-view shot or succession of shots), screen narration has clearly made much use, as in Alan Bridges' film version of Rebecca West's novella The Return of the Soldier (1982), in which the first-person narration of the original is reduced to allowing the novel's narrator a preponderance of point-of-view shots. However, a 'preponderance' is by no means equivalent to the continuing shaping, analysing, directing consciousness of a first-person narrator. Further, as Thomas Elsaesser has noted, 'The subjective perception --what the characters themselves see and how they experience it--is integrated with an objective presentation of these individual points of view and what they signify inside the same narrative movement and the continuous action.'45 While cinema may be more agile and flexible in changing the physical point of view from which an event or object is seen, it is much less amenable to the presentation of a consistent psychological viewpoint derived from one character.
(b) Oral narration or voice-over The device of oral narration, or voice-over, may serve important narrative functions in film (e.g. reinforcing a sense of past tense) but, by virtual necessity, it cannot be more than intermittent as distinct from the continuing nature of the novelistic first-person narration. (Woody Allen Radio Days (1987) is one of the few films that would be incomprehensible without its voice-over.) In usual film practice, the narrating voice-over may be dropped for sequences at a time: in fact, a sustained, non-diegetic oral accompaniment to visually presented action is scarcely feasible in relation to the feature films with which this study (like most cinema audiences) is concerned. Those words spoken in voice-over accompany images which necessarily take on an objective life of their own. One no longer has the sense of everything's being filtered through the consciousness of the protagonist-speaker: even in a film such as David Lean Great Expectations, which goes to unusual lengths to retain the novel's 'first-person' approach, the grotesques who people Pip's world are no longer presented to the viewer as an individual's subjective impressions. One now sees everything the camera 'sees', not just what impressed itself on the hero-narrator's imaginative responsiveness. In relation to those films which employ the voice-over technique, one's sense of the character to whom it is attributed is more likely to be the product of his involvement in the action directly presented than of his occasional comment upon it, whereas this is frequently not the case in the first-person novel.
45 Thomas Elsaesser, 'Film and the Novel: Reality and Realism of the Cinema' , Twentieth Century Studies' 9 ( Sept 1973), 61.
The omniscient novel
The narrative in such a novel is conveyed through two kinds of discourses: those attributed to various characters in direct speech (Colin MacCabe's 'object language'46) and that of the narrative (I should prefer 'narrating') prose, the apparently authoritative 'metalanguage' which surrounds them. It is this latter which guides our reading of the direct speech of the characters. MacCabe goes on to construct an analogy between these two kinds of discourse as they appear in the novel and in the film:
The narrative prose achieves its position of dominance because it is in the position of knowledge and this function of knowledge is taken up in the cinema by the narration of events. Through the knowledge we gain from the narrative we can split the discourses of the various characters from their situation and compare what is said in those discourses with what has been revealed to us through narration. The camera shows us what happens--it tells the truth against which we can measure the discourses.47
'The camera shows us what happens . . .'. David Bordwell has taken issue with MacCabe's hierarchy of discourses, both for its oversimplification of the classic realist novel and for the way in which it 'reduce[s] the range of filmic narration'. He particularly challenges MacCabe's 'privileging of camera work . . . over other film techniques', claiming that 'all materials of cinema function narrationally--not only camera but speech, gesture, written language, music, color, optical processes, lighting, costume, even offscreen space and offscreen sound'.48 There is a certain captiousness in Bordwell's response since, it seems to me at least, MacCabe's use of 'the camera' is a shorthand way of referring to all those narrational materials which the camera can show or imply: that is, all from Bordwell's list except those which relate to soundtrack, which can, of course, initiate a tension with the visual image. Clearly, certain functions of the narrating prose, such as establishing setting and physical appearance of characters, can be achieved through the film's mise-en-scčne. Other functions, such as those which enable us, through the writer's tone, to evaluate a character's speech, seem less immediately amenable to the camera's eye. The camera in this sense becomes the narrator by, for instance, focusing on such aspects of mise-en-scčne as the way actors look, move, gesture, or are costumed, or on the ways in which they are positioned in a scene or on how they are photographed: in these ways the camera may catch a 'truth' which comments on and qualifies what the characters actually say.
It is, however, too simple to suggest that the mise-en-scčne, or its deployment by the cinematic codes (notably that of montage), can effortlessly
46 Colin MacCabe, 'Realism and the Cinema: Notes on Some Brechtian Theses' , Screen 15/2 ( Summer 1974), 10.
48 David Bordwell, Narration in the Fiction Film ( Methuen: London, 1985), 20.
appropriate the role of the omniscient, inaudible narrator, or that the camera (to interpret MacCabe more narrowly), by 'show[ing] what happens', replaces such a narrator. For one thing (and a very obvious one) the camera here used metonymically to denote its operator and whoever is telling him what to aim it at, and how--is outside the total discourse of the film, whereas the omniscient narrator is inextricably a part of the novel's. Or perhaps it is truer to say that the omniscient narration is inextricably part of the novel's total discourse, as much so as the spoken words of the characters. (However, whereas the latter--the spoken words--can, if a film-maker wants them to, be rendered word for word by the characters in the film, clearly no such possibility exists for the narration: for that narrational prose to which, in most novels, we allow a privileged position of knowledge about characters, periods, places; knowledge which may in fact be concealed from characters in the novel.) By exercising control over the mise-en-scčne and soundtrack or through the manipulations of editing, the film-maker can adapt some of the functions of this narrational prose. The latter may indicate adverbially the tone of voice in which a remark is made by a character; the camera, on the other hand, may register a similar effect through attention to the actor's facial expression or posture (i.e. aspects of the mise-en-scčne), or by cutting so as to reveal a response to such a remark (i.e. through montage) which will guide the viewer's perception of the remark, as well as through the actor's vocal inflection (i.e. through sound-track).
In regard to rendering on film the descriptive functions of narrating prose, relating to places, objects, activities, there is perhaps a stronger possibility of the new (cinematic) reality at once displacing the earlier (verbally created) reality; in matters relating to character and to the psychological action involving characters, the situation is more complex. There is, in film, no such instantly apparent, instantly available commentary on the action unfolding as the novel's narrating prose habitually offers. In the omniscient novel, in which this prose is not 'suspect' in the sense of belonging to a first-person narrator, the continuing mediation between the reader and the action of the novel is, by virtue of its privileged status as 'knowledge', the reader's guarantee of the 'truth' of the proceedings. In a sense, all films are omniscient: even when they employ a voice-over technique as a means of simulating the first person novelistic approach, the viewer is aware, as indicated earlier, of a level of objectivity in what is shown, which may include what the protagonist sees but cannot help including a great deal else as well.
The mode of 'restricted consciousness'
In broad terms, it appears that neither first-person nor omniscient narration is, of its nature, amenable to cinematic narrative. Both seem always to know too much, or at least to know more than we feel is known in advance by the
more directly experienced film narrative; and this sense of foreknowledge is no doubt intimately connected to the characteristic past-tense rendering of the prose narrative as opposed to the perceptual immediacy of the film. The novelistic form of the restricted consciousness (as in Daisy Miller) perhaps approximates most closely to the cinematic narrative mode. Cohen, in discussing the techniques of Conrad and James, and making comparisons with impressionist painters, writes:
The indirect approach of these novelists [Conrad and Ford Madox Ford] is not fully comprehensible without reference to their unconventional handling of point of view. . . . The reader, one might say, is constantly forced to pass through several foregrounds before he can make out clearly what is looming in the background. . . . The same basic mechanism is operative with James's 'central reflectors' through whom all or nearly all of the action takes place.49
Such 'central reflectors'--for example, Strether in The Ambassadors, Winterboume in Daisy Miller --provide a point of identification for the reader, not necessarily in the affective sense but as a more or less consistently placed vantage-point from which to observe the action of the narrative. One is conscious always that there is a more comprehensive point of view than that available to such protagonists; that there is, as it were, a narrator looking over their shoulder, in the way that the camera may view an action over the shoulder of a character in the foreground of a shot, giving the viewer both the character's point of view and a slightly wider point of view which includes the character. The Jamesian concept of the 'centre of consciousness', by no means to be confused with narrational omniscience or the latter's obliteration by first-person narration, is perhaps the nearest that film may come in the direction of either first- or third-person narration. Its use will be examined in more detail in the case-study of Daisy Miller.
A Note on Terminology
The foregoing distinctions considered under the headings of 'Narrative Functions' and 'Kinds of Narration' may be summarized as those between a series of events sequentially and consequentially arranged and the modes (more easily distinguishable in literary terms) of their presentation. This distinction between narrative and narration finds rough parallels in that between story and discourse. The latter pair-- histoire and discours in modern French poetic--derives from the Russian Formalist distinction of the 1920s 'between fabula --the story-material as pure chronological sequence--and suzet, the plot as arranged and edited by the shaping of a story-teller, i.e. the finished narrative work as we experience it in a text; no longer pure story but
49 Cohen, Film and Fiction, 35.
a selective narrative act',50 in Roger Fowler's words. Histoire and discours he goes on to define as 'story-matter and its manner of delivery'.51 In the proliferating terminology of film theory, a further parallel frequently subsumes the categories referred to above in discussions of enunciated and enunciation. Of these terms, originating with the linguist Émile Benveniste, the former designates the 'utterance' (l'énoncé) as manifested in 'a stretch of text'52 (David Bordwell's phrase), as a coherent set of events enacted in a series of syntagmatic units, as the sum of its narrative functions. The latter, the enunciation (l'énonciation), characterizes the process that creates, releases, shapes (I am aware of groping for exactly the right word) the 'utterance'. Enunciation, that is, refers to the ways in which the utterance is mediated, and, as such, obviously shares common ground with narration, suzet, and discourse. Neither film nor novel is 'transparent', however much either seeks to suppress signs of its enunciation, such 'suppression' being much more marked in the case of film. Film may lack those literary marks of enunciation such as person and tense, but in the ways in which, for example, shots are angled and framed and related to each other (i.e. in matters relating to mise-en-scčne and montage) the enunciatory processes are inscribed. The institutional codes and their often highly individual deployment by different film-makers can either minimize or foreground the processes of cinematic enunciation but they cannot eradicate them, even when, as Metz writes, 'films give us the feeling that we are witnessing almost a real spectacle'.53 Film enunciation, in relation to the transposition of novels to the screen, is a matter of adaptation proper, not of transfer. In the case-studies offered below, it will be considered in relation to how far the films concerned exhibit the interaction of cinema-specific and extra-cinematic codes, and to what extent they provide--or seek to provide--equivalences for the enunciatory procedures of the novels on which they are based. An essential function of this study will be to distinguish between:
i. those elements of the original novel which are transferable because not tied to one or other semiotic system--that is, essentially, narrative ; and
ii. those which involve intricate processes of adaptation because their effects are closely tied to the semiotic system in which they are manifested--that is, enunciation.
I have preferred 'enunciation', finally, to 'narration' because the latter is too often attached, in a limiting way, to matters of person and tense. By enunciation, I mean the whole expressive apparatus that governs the presentation and reception--of the narrative.
50 Roger Fowler, Linguistic and the Novel ( Methuen: London, 1977), 78-9.
51 Ibid. 79.
52 Bordwell, Narration in the Fiction Film, 21.
53 Metz, Imaginary Signifier, 4.
What is Understood by Adaptation
The 'distinctive feature' of adaptation, asserts Dudley Andrew, is 'the matching of the cinematic sign system to a prior achievement in some other system'. He claims that 'Every representational film adapts a prior conception . . . [but that] Adaptation delimits representation by insisting on the cultural status of the model . . . in a strong sense adaptation is the appropriation of a meaning from a prior text'.54 The 'matching' and the 'appropriation' referred to are in the interests of replacing one illusion of reality by another. Whatever claims of fidelity and authenticity are made by film-makers, what these essentially amount to are the effacement of the memory derived from reading the novel by another experience--an audio-visual-verbal one-which will seem, as little as possible, to jar with that collective memory. It seeks, with one concretized response to a written work, to coincide with a great multiplicity of responses to the original. Its aim is to offer a perceptual experience that corresponds with one arrived at conceptually. The kinds of complaints directed at film adaptations of classic or popular novels, across a wide range of critical levels, indicate how rarely the 'appropriation of meaning from a prior text' is fully achieved--even when it is sought. Underlying the processes suggested here, in the manufacture of the more or less faithful film version at least, are those of transferring the novel's narrative basis and of adapting those aspects of its enunciation which are held to be important to retain, but which resist transfer, so as to achieve, through quite different means of signification and reception, affective responses that evoke the viewer's memory of the original text without doing violence to it.
The preceding paragraph of course suggests (wrongly) that the film adaptation will reach only viewers who are familiar with the novel. The very fact that this is not the case ought to be a deterrent to the fidelity-seeking critics, indicating that there is a varying, but large, segment of the audience to whom an adaptation is of no more consequence or interest as such than any other film. The stress on fidelity to the original undervalues other aspects of the film's intertextuality. By this, I mean those non-literary, non-novelistic influences at work on any film, whether or not it is based on a novel. To say that a film is based on a novel is to draw attention to one--and, for many people, a crucial--element of its intertextuality, but it can never be the only one. Conditions within the film industry and the prevailing cultural and social climate at the time of the film's making (especially when the film version does not follow hot upon the novel's publication) are two major determinants in shaping any film, adaptation or not. Among the former (i.e. conditions within the industry) one might include the effect of certain star personae, or, in the days of the studios' dominance, a particular studio's 'house style', or a director's predilections or genre conventions, or the pre-
54 Andrew, "'Well-Worn Muse'", 9.
vailing parameters of cinematic practice. As to the latter (i.e. the climate of the times) it is difficult to set up a regular methodology for investigating how far cultural conditions (e.g. the exigencies of wartime or changing sexual mores) might lead to a shift in emphasis in a film as compared with the novel on which it is based. However, it is necessary to make allowance in individual cases of adaptation for the nature of such influences, and this matter will be looked at more closely in the Special Focus section of the chapter on Cape Fear.
Perhaps, indeed, it is just because questions of narrativity can be formalized that so much attention is paid to the original text's contribution to the film. And certainly, in raising the issue of intertextuality, I am not denying how powerfully formative the source work is in shaping the response of many people to the film version. Consequently two lines of investigation seem worthwhile: ( a ) in the transposition process, just what is it possible to transfer or adapt from novel to film; and ( b ) what key factors other than the source novel have exercised an influence on the film version of the novel? For those who know and/or value the novel, the process of narrativity in regard to the film version will necessarily differ from that of the spectator unfamiliar with it: in either case, a true reading of the film will depend on a response to how the cinematic codes and aspects of the mise-en-scčne work to create this particular version of the text.
What Kind of Adaptation?
While the fidelity criterion may seem misguided in any circumstances, it is also true that many film-makers are on record as being reverently disposed towards reproducing the original novel on film. It is equally clear, however, that many adaptations have chosen paths other than that of the literalminded visualization of the original or even of 'spiritual fidelity', making quite obvious departures from the original. Such departures may be seen in the light of offering a commentary on or, in more extreme cases, a deconstruction ('bring[ing] to light the internal contradictions in seemingly perfectly coherent systems of thought'55) of the original. While I do not wish to propose a hierarchy of valuableness among such approaches, it does seem important in evaluating the film version of a novel to try to assess the kind of adaptation the film aims to be. Such an assessment would at least preclude the critical reflex that takes a film to task for not being something it does not aim to be. Given the precariousness of the concept of fidelity in relation to novels made from films, it seems wiser to drop terms like 'violation', 'distortion', 'travesty', and those others which, like them, imply the primacy of the printed text.
55 John Sturrock, "'Introduction'" to Sturrock (ed.), Structuralism and Since ( Oxford University Press: Oxford, 1979), 14.
AGENDA FOR FURTHER STUDY
Nothing is likely to stop the interest of the general film-viewer in comparing films with their source novels, usually to the film's disadvantage. The aim of the present study is to use such concepts and methods as permit the most objective and systematic appraisal of what has happened in the process of transposition from one text to another. Given the prevalence of the process, and given that interpretations and memories of the source novel are powerful determining elements in the film's intertextuality, there is little value in merely saying that the film should stand autonomously. So it should, but it is also valuable to consider the kinds of transmutation that have taken place, to distinguish what the film-maker has sought to retain from the original and the kinds of use to which he has put it.
Transfer and Adaptation Proper
This distinction, elaborated earlier in this chapter, is central to the procedures of the following case-studies and, I believe, to any systematic study of what happens in the transposing of novel into film.
In considering what can be transferred from novel to film, one begins to lay the theoretical basis for a study of the phenomenon of turning novels into films as well as a basis for what has been transferred in any particular case (i.e. how far the film-maker has chosen to transfer what is possible to do so). In broad terms, this involves a distinction between narrative (which can be transferred) and enunciation (which cannot, involving as it does quite separate systems of signification). Some potentially valuable strategies for considering the idea of transfer are outlined below.
The story/plot distinction
Terence Hawkes, drawing on Viktor Shklovsky's work on the nature of narrative, makes the following distinction: '"Story" is simply the basic succession of events, the raw material which confronts the artist. Plot represents the distinctive way in which the "story" is made strange, creatively deformed and defamiliarized.'56 Novel and film can share the same story, the same 'raw materials', but are distinguished by means of different plot strategies which alter sequence, highlight different emphases, which--in a word-defamiliarize the story. In this respect, of course, the use of two separate systems of signification will also play a crucial distinguishing role.
56 Terence Hawkes, Structuralism and Semiotics ( Methuen: London, 1977), 65--6.
The distinction between 'distributional' and 'integrational' functions
As discussed earlier, Barthes's 'distributional functions', those which he designates as 'functions proper', are those most directly susceptible to transfer to film. This classification is further subdivided into cardinal functions, those narrative actions which open up alternatives with direct consequences for the subsequent development of the story ('the risky moments of a narrative' in Barthes's term), supported, given a richer texture, by elements characterized by a different order of functionality. This 'different order' may be either 'lesser' (in the case of catalysers) or 'vertically functioning' (in the case of distributional functions) as opposed to the essential horizontality of the cardinal functions. The first level of 'fidelity' in relation to the film version of a novel could be determined by the extent to which the film-maker has chosen to transfer the cardinal functions of the precursor narrative.
Identification of character functions and fields of action
If we take V. Propp's notion 'that the all-important and unifying element is found . . . in the characters' functions, the part they play in the plot',57 that these functions58 are distributed among a limited number of 'spheres of action',59 and that the 'discernible and repeated structures which, if they are characteristic of so deeply rooted a form of narrative expression, may . . . have implications for all narrative60 (i.e. not just for folk-tales), then we may see a further way of systematizing what happens in the transposition of novel into film. No doubt the character functions are more clearly displayed in a Russian folk-tale than in a complex nineteenth-century English novel or a feature-length film; nevertheless, some of Propp's formulations point to underlying, transferable components of narrative. (Barthes's concept of cardinal functions is partly based on Propp's work, as Barthes acknowledges.61) To Propp, 'Function is understood as an act of character, defined from the point of view of its significance for the course of the action.'62 It is not that he fails to allow for other narrative elements as having their roles to play; he, in fact, pays special attention to the question of motivations, which 'often add to a tale a completely distinctive, vivid colouring';63 but he finds those elements other than character functions and their connectives 'less precise and definite'.64
57 Terence Hawkes, 68.
58 V. Propp, Morphology of the Folktale (1927), trans. Laurence Scott (University of Texas: Austin, 1968), 22-63.
59 Ibid. 79.
60 Ibid. 21.
61 Barthes, "'Introduction to the Structural Analysis of Narratives'", 92.
62 Propp, Morphology of the Folktale, 21.
63 Ibid. 75.
64 Ibid. 43.
I am suggesting here that, in considering what kind of adaptation has been made, one might isolate the chief character functions of the original and observe how far these are retained in the film version. (Peter Wollen's Proppian analysis of Hitchcock North by Northwest65 suggests that a sophisticated narrative is susceptible to procedures and categorizations derived from the study of much simpler modes.) By observing these functions, distributed among seven 'spheres of action' (named for their performers-'villain', 'helper', etc.), one could determine whether the film-maker has aimed to preserve the underlying structure of the original or radically to rework it. Such a study would give a firmer basis for comparison by sorting out what functions are crucial to the narrative: i.e. to the plot which organizes the raw materials of the story.
Identification of mythic and/or psychological patterns
In relation to those myths which encapsulate in narrative form certain universal aspects of human experience, Lévi-Strauss has claimed that 'the mythical value of the myth is preserved even through the worst translation . . . [Unlike poetry, its] substance does not lie in its style, its original music, or its syntax, but in the story it tells'.66 By extension, then, it is not too much to expect that the mythic elements at work in a novel seem likely to be transferable to the screen since their life is independent of whatever manifestation they are found in, resistant as they are to even 'the worst translation'. Intimately connected with the idea of myth, such Freudian concepts as the Oedipus complex so profoundly underlie human experience, and, therefore, the narrative renderings of that experience, that their nature remains unchanged through varying representations. The denotative material which provides the vehicle for these patterns may change from novel to film without affecting the connotations of the mythic and psychological motifs themselves. It is clear that these patterns exercise a powerfully organizing effect on narratives: one could propose, for example, that the Freudian notion that: 'An action by the ego is as it should be if it satisfies simultaneously the demands of the id, of the super-ego and of reality'67 provides a way of classifying the narrative elements and their motivations in a story--whether on page or screen. What the approaches outlined above have in common are:
a. they all refer to elements which exist at 'deep levels' of the text;
b. they address narrative elements which are not tied to a particular mode of expression (i.e. those which may be found at work in verbal or other sign systems); and
65 Peter Wollen, "'North by North-West: A Morphological Analysis'", Film Form, 1 ( 1976), 20-34.
66 Claude Lévi-Strauss, Structural Anthropology ( Penguin Books: Harmondsworth, 1972), 210.
67 David Stafford-Clark, What Freud 'Really' Said ( Penguin Books: Harmondsworth, 1967), 112.
c. all are susceptible to that more or less objective treatment that eludes less stable elements (e.g. character motivation or atmosphere).
They relate to the level of narrative, to areas in which transfer from one medium to another is possible, and to isolate them is to clear the way for examination of those elements that resist transfer and call for adaptation proper.
Those elements of the novel which require adaptation proper may be loosely grouped as (in Barthes's term) indices, as the signifiers of narrativity, and as the writing, or, more comprehensively, as enunciation, to use the term now commonly employed in film theory. The film version of a novel may retain all the major cardinal functions of a novel, all its chief character functions, its most important psychological patterns, and yet, at both micro- and macrolevels or articulation, set up in the viewer acquainted with the novel quite different responses. The extent to which this is so can be determined by how far the film-maker has sought to create his own work in those areas where transfer is not possible. He can, of course, put his own stamp on the work by omitting or reordering those narrative elements which are transferable or by inventing new ones of his own: my point is that, even if he has chosen to adhere to the novel in these respects, he can still make a film that offers a markedly different affective and/or intellectual experience. Some key differences which need to be considered in relation to areas of adaptation proper are summarized below. Essentially they refer to distinctions between enunciatory modes.
Two signifying systems
The full treatment of such a topic is of course beyond the scope of this study; at this point I want merely to draw attention to some matters centrally important here. The novel draws on a wholly verbal sign system, the film variously, and sometimes simultaneously, on visual, aural, and verbal signifiers. Even the apparently overlapping verbal signs (the words on the novel's page, the written or printed words used in the film, e.g. letters, street signs, newspaper headlines), while they may give the same information, function differently in each case. In the examples given, the letter, the street sign, and the newspaper headline will each resemble their real-life referents in ways that are customarily beyond the novel's capacity for iconic representation. And this semi-exception to the rule of difference between the two systems points to a major distinction between them: the verbal sign, with its
low iconicity and high symbolic function, works conceptually, whereas the cinematic sign, with its high iconicity and uncertain symbolic function, works directly, sensuously, perceptually. Such a distinction is all but axiomatic but failure to concede its pervasive importance leads to a good deal of impressionistic, dissatisfied, and unsatisfying comparison of novel and film. Comparisons of this kind grow out of a sense of the film-maker's having failed to find satisfactory visual representations of key verbal signs (e.g. those relating to places or persons), and of a sense that, because of its high iconicity, the cinema has left no scope for that imaginative activity necessary to the reader's visualization of what he reads. In the study of adaptation, one may consider to what extent the film-maker has picked up visual suggestions from the novel in his representation of key verbal signs--and how the visual representation affects one's 'reading' of the film text.
The novel's linearity and the film's spatiality
We construct meaning from a novel by taking in words and groups of words sequentially as they appear on the page. In order, say, to grasp a scene, a physical setting, we have no choice but to follow linearly that arrangement of arbitrary symbols set out, for the most part, in horizontal rows which enjoin the linearity of the experience. The relentless linearity associated with the usual reading of a novel favours the gradual accretion of information about action, characters, atmosphere, ideas, and this mode of presentation, of itself, contributes to the impression received. At first/glance, it may seem that the relentless movement of film through the projector offers an analogy to this situation. (And, of course, classic narrative cinema is posited on a powerful, forward-thrusting linearity, the product of causality and motivation.) However, though viewing time (and, thus sequentiality) is controlled much more rigorously than reading time, frame-following-frame is not analogous to the word-following-word experience of the novel. There are at least two significant differences to be noted: (i) the frame instantly, and at any given moment, provides information of at least visual complexity (sometimes increased by the input of aural and verbal signifiers) beyond that of any given word because of the spatial impact of the frame; and (ii) the frame is never registered as a discrete entity in the way that a word is. We do not ordinarily view a film frame by frame as we read a novel word by word.
The fact that we are always being exposed to the multiplicity of signifiers contained within the space of a frame or series of frames has implication for the adaptation of verbal material; for example, as it relates to the representation of characters and settings. What we receive as information from the mise-en-scčne may be less susceptible to the film-maker's control (because of the strongly spatial orientation of film and because of the simultaneous
bombardment by several kinds of claims on our attention) than what we receive from the linear presentation of words on the page. Dickens, for instance, may force us to 'see' Miss Havisham in the interior of Satis House in the order he has chosen in Great Expectations; as we watch her visual representation in David Lean's film we may be struck first, not by the yellow-whiteness of her apparel, but by the sense of her physical presence's being dwarfed by the decaying grandeur of the room. In the form which stresses spatiality rather than linearity, the eye may not always choose to see next what, in any particular frame, the film-maker wants it to fasten on. The challenge to the film-maker's control of the mise-en-scčne is obvious.
Cinematic enunciation has two other approaches, specific to its medium, relating to the disposition of space and, hence, to the generation of aspects of narrative in ways closed to the novel. They are: (i) Noel Burch's theory of a dialectic between on-screen and off-screen space (he identifies six 'segments' of off-screen space, four determined by the borders of the frame, the others 'an off-screen space behind the camera' and 'the space existing behind the set or some object in it'68); and (ii) Raymond Bellour's proposal of alternation69 (e.g. between long shot and close-up, between seeing and being seen) as a key cinematic practice, operating on levels of both code and diegesis. Both of these concepts advert to enunciatory techniques peculiar to the unfolding of cinematic narrative, and both will be considered in relation to particular case-studies in this book. Neither has any real equivalent in the verbal narrative, except in the much broader sense of alternation offered by a novel's moving between two major strands of narrative. The 'spacelessness' of the novel's linear procedures precludes the setting up of spatial tension achieved by (i) and the spatial mobility required by (ii).
If film, unlike verbal language, has no vocabulary (its images, unlike words, are non-finite), it also lacks a structuring syntax, instead of which it has conventions in relation to the operation of its codes. In so far as these codes enable us to 'read' film narratives, in so far as we learn to ascribe meanings to them (e.g. to assume that 'fade out/fade in', as an editing procedure, denotes a major lapse of time), it is through frequent exposure to their deployment in a particular way, without there being any guarantee that they will always be used in this way. There is, for instance, nothing corresponding to the comparatively fixed usage of full-stop and comma as punctuational signs denoting the longest and shortest pauses respectively, or to those rules which signify tenses, in the written work.
68 Noel Burch, The Theory of Film Practice ( Cinema Two/ Secker and Warburg: London, 1973), 20.
69 Janet Bergstrom. "'Alternation, Segmentation, Hypnosis: Interview with Raymond Bellour'", Camera Obscura, 3-4 ( 1979), 76-8.
Further, in 'reading' a film, we must understand other, extra-cinematic codes as well. These include:
a. language codes (involving response to particular accents or tones of voice and what these might mean socially or temperamentally);
b. visual codes (response to these goes beyond mere 'seeing' to include the interpretative and the selective);
c. non-linguistic sound codes (comprising both musical and other aural codes);
d. cultural codes (involving all that information which has to do with how people live, or lived, at particular times and places).
In a sense, the cinematic codes may be seen as integrating the preceding four in ways that no other art-form does. When we witness a film, we share with the film's maker a basic assumption that we know the codes: i.e. a general cinematic code which, as Christian Metz has shown, can be broken down into subcodes, such as those to do with editing, or those to do with particular genres, and the extra-cinematic codes referred to above. Failure to recognize--or, at least, to pay adequate attention to--the differences between the operation of these codes in film and the novel's reliance on the written representation of language codes has been a key element in accounting for the fuzzy impressionism of so much writing about adaptation.
Stories told and stories presented
In moving from novel to film, we are moving from a purely representational mode to 'an order of the operable',70 to use Barthes's distinction (which has not, to my knowledge, been pursued in film studies but which offers a broad statement of intermedial disparity). This distinction relates partly to earlier points about:
i. differences between two 'language' systems, one of which works wholly symbolically, the other through an interaction of codes, including codes of execution;
ii. tense: film cannot present action in the past as novels chiefly do; and
iii. film's spatial (as well as temporal) orientation which gives it a physical presence denied to the novel's linearity.
Another aspect of the distinction between telling and presenting is located in the way in which the novel's metalanguage (the vehicle of its telling) is replaced, at least in part, by the film's mise-en-scčne. In a sense, the film's story does not have to be told because it is presented. Against the gains in immediacy, the loss of the narrational voice may, however, be felt as the chief casualty of the novel's enunciation.
70 Barthes, "'Introduction to the Structural Analysis of Narratives'", 80.
The enunciatory matters discussed above refer to crucially important novelistic elements which offer challenges to the film-maker, especially if he does not wish the experience of his film to shatter a pre-existing reality (i.e. of the novel) but, rather, to displace it.