Mapping the Textual Genome: Post-Scripts, Post-Structuralism and Chandler's Poodle Springs
To borrow from Churchill's old line: creative writing is 'a riddle wrapped in a mystery inside an enigma' (Churchill 694). The riddle is whether or not it can, in the truest sense of the word, be 'taught'.
The mystery, it seems to me, is that many of the most prominent sceptics are actually teachers within the discipline itself. Comments such as those made by poet-critic Louis Simpson are indicative of the general view:
The enigma is that the majority of the 'untalented' masses (certainly not those attending creative writing classes) would be liable to agree with him. And if a hint of vanity may be detected behind Simpson's invocation of talent, the same charge can be levelled at these others as well. According to Nietzsche:
For the purpose of what is to follow, then, the proposition that writing cannot be taught will be conceived of as a literary 'Creation Myth'.
The predictable alternative is a theory of artistic evolution, such as that proposed by Harold Bloom: 'No one "fathers" or "mothers" his or her own poems, because poems are not "created," but are interpreted into existence, and by necessity they are interpreted from other poems' (Bloom 1982: 244). And, as in nature, a writer's evolution proceeds, according to David Lodge, by a series of 'kick-start[s] - that is, they begin by imitating and emulating the literature that gives them the biggest kicks' (Lodge 1996: 171). Lodge acknowledges his own indebtedness to the influence of Graham Greene ('how to use a few selected details to evoke character or sense of place'), Evelyn Waugh ('how to generate comedy by a combination of the familiar and the incongruous') and James Joyce ('how to make a modern story re-enact, echo or parody a mythical or literary precursor-narrative') (Lodge 1996: 172). What is clear in any case is that the study of admired authors is a valuable, perhaps even an essential component in the teaching of creative writing.
The question is what form this should take. What Lodge appears to be describing is the literary equivalent of an art student's detail, whereby an isolated section of a past master's work is reproduced and studied outside the context of the surrounding image. This approach is not without its perils however, as Roland Barthes discovered during his short-lived apprenticeship in classical drawing: 'I copy and naïvely connect detail to detail; whence unexpected "conclusions": the horseman's leg turns out to be perched right on top of the horse's breastplate, etc' (Barthes 1975: 93-94).
The peril of studying literary details is that students will lose sight of the 'big picture'; that is, the context (plot, point of view, etc) in which, say, Greene evokes character or a sense of place. Taken out of their frame, it is questionable whether such details can be understood at all. Indeed, the seeds of this objection are by no means new, having been planted in the earliest stages of the creative writing debate, in Henry James's The Art of Fiction (1884):
This leads me to conclude that the solution is to retrace the text in its entirety, which is not to say that James would agree: 'The critic who over the close texture of a finished work shall pretend to trace a geography of items will mark some frontiers as artificial, I fear, as any that have been known to history' (James 1884: 34). The same is true for Lodge: 'Even a single sentence in a novel is a complex product of innumerable chains of cause and effect which reach deep into the writer's life and psyche. To distinguish, analyse and retrace them all would be impossible' (Lodge 1996: 178). Furthermore, he argues, 'a creative writer who tries to make his students clones of himself [or any other writer, presumably] is doing them the worst possible service' (Lodge 1996: 176, emphasis mine).
By invoking the spectre of cloning, Lodge in effect casts writing teachers as literary geneticists who to a greater or lesser degree mediate the 'natural' evolution of their students' craft. Or perhaps the analogy just appeals to me because it is my position that, like genes, details from the writings of Greene, Waugh or Joyce cannot be isolated until the 'genome' of their corresponding texts has first been mapped, i.e. until they are understood within the context of the whole. I would not dispute James' view that this geography is artificial, however I will attempt to demonstrate that its artificiality is not inherently problematic. For the moment it may suffice to note that the alternative, as exemplified by Barthes' horseman, can produce its own undesirable mutations.
Of course, it might also be noted that Lodge's warnings against the dangers of literary cloning transcend even critical consensus and exist more in the realm of commonsense. This does not mean that they are warranted. According to Bloom:
This notion of freshness, arguably writing's highest value (in critical circles, though not necessarily in the popular market), is perhaps less a matter of saying something new than of arranging what has already been said in new ways. Even if I were to convince myself that, between its quoted passages from Bloom et al, sections of this article are wholly original, the sum of its readers will undoubtedly detect echoes of past theorists in every thought and phrase. That I cannot name these precursors is testament only to my relative ignorance; that I am aware of my ignorance commends me in that at least I am not naïve. The advantage of textual mapping/rewriting is not that it prevents student writers from reinventing the wheel; however, by enacting this reinvention consciously, it does increase the possibility that they will find new ways of 'spinning' it.
The supposed disservice of this approach is that it will prevent student writers from developing their own voice, but in fact there is evidence to suggest that the opposite is true. After all, if writing in another's voice (style, structure, thematic preoccupations, etc) is indeed a form of literary constraint, at once a self-imposed handicap and a crutch, it is essentially no different to writing in a particular poetic form or meter, or learning to walk gracefully by balancing a book on your head. Grace is achieved not only because of the book, but also in spite of it, and this becomes even more evident once the book has been discarded. Moreover, no one would be likely to suggest that discarding the book might somehow present a problem (if students of deportment suffered from the same lack of self-determination so commonly ascribed to writers, the history of millinery might have taken an altogether bizarre turn). In fact, writers have successfully adopted and discarded literary constraints for years. As Colin Symes remarks:
For a writer such as Georges Perec, whose experiments in self-imposed handicaps included writing a three-hundred page novel without using the letter 'e' (La Disparition [A Void, 1979]), the use of constraints actually stimulated his creativity to the extent that 'many of his texts almost wrote themselves' (Symes 1999). It is worth observing, after Symes, that this bears a nominal resemblance to 'inspired automatic writing' (Symes 1999).
Automatic writing is also one of the lesser miracles of the Creation Myth, but pales into insignificance compared to the central myth of the artistic vision or gestalt, whereby a complete and self-contained composition appears in the artist's mind while they might otherwise be preoccupied reading the newspaper, weeding the garden, or in their bed asleep. But is it a myth? On the contrary, Mike Sharples contends that this phenomenon is not even the province of a gifted few:
This phenomenon can also be simulated. For the less 'talented' or 'visionary' student writer, Sharples offers the following pragmatic alternative:
in James's words, an 'artificial geography', and it may be suggested that an example of such a geography can be found in Barthes' S/Z [1970a]. Briefly summarised, S/Z is a post-structural analysis of Honoré de Balzac's short story 'Sarrasine' (1830), in which the text is divided up into 561 fragments or 'lexias' and catalogued according to five semantic codes, with the intention of providing not a single, authoritative (authorial) interpretation, but rather a polysemous, 'plural' reading (Barthes 1970a: 16). Renouncing the idea of a structural model to which all texts could then be applied, the gradual, 'step-by-step' unfolding of its analysis instead traces the order of the reading itself (left to right, top to bottom, page by page), producing what Barthes describes as a 'structuration' or 'writing-reading' (Barthes 1991: 73). In short, S/Z is an externalised representation of the reading process: 'the decomposition (in the cinematographic sense) of the work of reading: a slow motion, so to speak' (Barthes 1970a: 12).
As a model for mapping a text's genome, S/Z is ideally suited for two reasons. The first of these reasons can be found in the underlying theory of its related precursor-text. Despite being the object of several decades of (often posturing) debate, the central thesis of 'The Death of the Author' (1967) (note 1) is simply, to agree with James, that a text is not composed of a series of blocks, but rather a tapestry of semantic threads, woven together by the reading subject: 'a text's unity lies not in its origin but in its destination' (Barthes 1967: 148). What is mapped in the text is not the mind of its 'Author-God' (Barthes 1967: 146), or a template for all texts, or even the monolithic sum of all its possible readings (one reading for all); what it traces instead are the multitude of readings contained within a single individual (all for one). The goal of an S/Z-style analysis, accordingly, is not so much interpretative as it is corporeal, by investing the reader 'in the production, not the product' (Barthes 1991: 189). It allows the reader to 'speak' the text (Barthes 1991: 73) rather than to say any one thing about it, because while conventional criticism 'rests on the notion that the text contains insignificant elements', the proliferation of meaning that it exemplifies 'does not support the separation of foundation and design, insignificant and significant...: everything signifies something' (Barthes 1970a: 51). Finally, then, and most importantly within the context of this article: 'if one or another of these [meanings] are sometimes permitted to come forward, it is in proportion' (Barthes 1970a: 6).
The other advantage of the S/Z model is that it 'permits the rewriting of texts' (Barthes 1991: 110). The problem for student writers is not that they are unable to weave together the threads of their reading, but rather that these threads are fed from the narrative spool in a very specific order, i.e. the order of the reading itself. Barthes' concept of rewriting is therefore significantly dependent upon a corresponding concept of rereading. To illustrate this significance, he offers an example from the text of 'Sarrasine,' in which the title-character becomes the victim of a case of mistaken sexual identity (his mistake, that is):
As with his concept of rewriting (which actually refers to the production of meaning), Barthes' concept of rereading is essentially symbolic. Rather than literally rereading the text (a difference between readings), this symbolic rereading is achieved through the application of S/Z's codes, 'producing' the text in its plurality (a difference within a single reading: 'a difference of which each text is the return' (Barthes 1970a: 3)). His claim therefore is not that a foreknowledge of events in the text will multiply the signifiers contained in its analysis; it is, more radically, that the multiplication of the signifiers will provide the reader with a foreknowledge of events within the text.
If the details drawn from an S/Z-style analysis do indeed exist in proportion to each other as a coherent structure 'of ideas, stimuli, associations', then certainly it would seem possible 'to read while looking up'. Here the predictive aspect of simultaneous rereading becomes an acid test for the efficacy of the model as a whole: it is one thing to read the outcome of a story in its earliest stages when one can peek ahead anyway, but would this still be the case with the safety net removed? If discernible patterns really do exist, it follows that it should also be possible to extend the structure of an unfinished text, tracing its semantic trajectory in the same way that a child joins the dots on an otherwise blank page, filling it with pictures.
The remainder of this article will therefore be devoted to proving this hypothesis through an analysis of 'The Poodle Springs Story', the beginnings of Raymond Chandler's last Philip Marlowe novel, comprising only four chapters upon his death in 1959. This will demonstrate the model's other practical application (as a tool for completing unfinished texts), but I would stress that my objective in doing so here is merely to prove (insofar as this is possible) that it works. In terms of writing courses, its purpose may more usefully be the mapping / symbolic rewriting of completed texts, as the basis for a literal rewriting of one part or another, the model then being reapplied to this revised section as a way of analysing/editing the results.
First things first, however. To begin with, what little there is of Chandler's novel can be outlined as follows: Marlowe arrives in Poodle Springs with his new bride Linda Loring (who proposed to him at the end of his previous adventure, Playback ) at a house she has rented for the season. As the daughter of millionaire businessman, Harlan Potter, Linda is extremely wealthy, a fact which doesn't sit well with her husband. Despite her attempts to dissuade him, Marlow is determined to return to his work as a private detective and drives into town to look for an office. Whilst there he is approached on the street by a man named Manny Lipshultz. Lipshultz is the owner of a place on the outskirts of town known as 'The Agony Club'. He claims his life is in danger. Unwilling to commit to anything before he gets settled, Marlowe gives Lipshultz his number and checks in at the police station, whereupon he comes to realise that his recent marriage has made him something of a local celebrity. On the way home he is accosted by a couple of men in Lipshultz's employ. Overcoming them, he arrives at the house and tells Linda of his confrontation. She seems unconcerned or, at best, unable to relate to the story.
Is it possible to derive a plot from these four chapters, whose sum total consists of little more than 4000 words? Renowned crime writer (and the person considered by most to be the heir-apparent to Chandler's throne), Robert B. Parker, evidently thought so, going on to complete the novel in 1989 as Poodle Springs (Chandler and Parker 1989). Others were more circumspect, with Ed McBain's review in the New York Times providing a typical example of the critical consensus. Like most of his peers, McBain sees the novel as a mixed success, reserving the majority of his criticism for Chandler, particularly with regard to the (ultimately superfluous) subplot of Marlowe's marriage to Linda, but also for his writing in general: 'It is difficult to believe that the first four chapters of "Poodle Springs" were intended by Chandler to be finished work' (McBain 1989) (note 2). Towards Parker, McBain is more forgiving:
Having 'set the scene', an S/Z-style analysis of Poodle Springs' first four chapters can now be attempted. Given that this analysis in its entirety is longer than Chandler's text itself (comprising in total no less than 122 annotated lexias), it is necessary in the context of this article to reduce this to a brief representative sample, with definitions of the five codes and summarised results to follow:
Beginning in the order they arise, the proairetic code is the code of actions (ACT) within a text. Consisting of a guideword (in this case, 'House'), all entries under this heading are numbered and unfolded into a series of supplementary definitions, such as to criticise the house, to enter through the front door or, as above, to inspect the house.
The hermeneutic code (HER) relates to questions posed by a text and to the answers it provides to those questions. Once identified, each enigma is catalogued under a series of sub-headings, including but not limited to: proposal, formulation, promise of/request for an answer and snare. The heading here is thematisation: 'an emphasizing of the subject [eight to ten million dollars] which will be the object of the enigma ["What is the source of Linda's wealth?"]' (Barthes 1970a: 209-210).
The semic or semantic code (SEM) is the code of connotations. These may apply both to people and to places, but most frequently the former. Semes predominantly relate to character, with objectification constituting Augustino's only entry under the code. This is hardly surprising: in the story, as in the house, his chief purpose is to serve as part of the scenery.
In an analysis published the year before S/Z as 'The Structural Analysis of Narrative: Apropos of Acts 10-11' (1969), Barthes established as many as twelve codes, including the narrative, topographic, historical, rhetorical, chronological and meta-linguistic, all of which later became subsets of the cultural code. In short, the cultural code catalogues references (REF) to any body of knowledge contained within a text, whether common, specialised, or esoteric (e.g. REF. Nineteenth-century parlour games).
Finally, the symbolic code or 'field' as Barthes calls it (SYM) encompasses a variety of symbolic languages: medieval, psychological, rhetorical and so on. In the example from Poodle Springs, the code relates to the last of these, representing one half (the 'A' term) of a series of antitheses, which shape the structure of the novel as a whole.
Results drawn from the application of these codes to Chandler's four chapters can be summarised as follows:
In light of these results, it is possible to construct the framework of a basic plot, which takes the following form: Lipshultz is a mobster in trouble with his 'godfather', Harlan Potter. According to the typology of American crime syndicates (REF), for this trouble to be fatal, it most likely involves either embezzled (mob) money, or else a desire to leave the syndicate altogether. Perhaps both: wanting to leave the syndicate, Lipshultz 'cooked' his nightclub's books in collaboration with a mob accountant, who then disappeared with the funds, leaving him 'holding the bag'. Marlowe would be hired to find the accountant, but in any case won't be able to stop Potter having Lipshultz killed, because it would take a crime as serious as murder to compel him to turn his father-in-law over to the police. Previously unaware of her father's occupation, Linda would nevertheless do her best to talk Marlowe out of it, finally seducing him into bed. Afterwards, Marlowe would change his mind and take Potter in anyway, effectively bringing an end to the marriage.
This plot is decidedly different from Parker's (which involves bigamy, pornography, and a $100,000 IOU); nevertheless, it is the similarities between the two that offer the most telling contrasts. In Parker's version (which, I should add, I hadn't read until my analysis was complete), the main antagonists are also a crime-boss and his daughter. Instead of using Potter and Linda however, he creates veritable 'clones' of these characters in the form of Clayton and Muriel Blackstone. As a result, Potter never makes it to town at all, and Linda is reduced to a narrative 'loose-end'. Similarly, Parker sets most of his version in Los Angeles, thereby all-but abandoning the allegorical setting of Poodle Springs. It is for this reason that critics like McBain are led to view the preoccupations of Chandler's chapters (Marlowe's marriage, establishing the setting) as mere filler. Even Marlowe's search for an office (Chapter 3) is resolved in Parker's version in the opening line of his first chapter. This is not to suggest that Parker's version is 'bad' (in fact, it is well worth reading) but it does demonstrate the importance (and, when ignored, the negative effects) of context. While managing to get many of the details right, Parker's version ultimately loses sight of the (Chandler's) 'big picture'.
Of course, it is precisely the details that are missing from my analysis as it is presented here, which includes a variety of unmentioned narrative and thematic observations (such as instances in the text of the metaphor of poker) and, perhaps more significantly, entries relating to style (Chandler's idiosyncratic use of similes, for example). These I hope to discuss in a future article, but for now it is enough to emphasise the way in which the reading 'holds together' (Barthes 1970a: 156), its codes cohering with and reinforcing one another, existing in proportion to each other as an externalised representation of one reader's (my) mental structures, i.e. as the 'big picture' from which such details can, at that later time, be 'drawn'.
Having 'read while looking up', the next step would be to extend Chandler's unfinished text in a piece of creative writing. The presentation of that piece of writing will also have to be left aside for the time being, because to include even a small sample in its proper, annotated context (in its production, not simply as a product) would eclipse the scope of the current article. What can be mentioned is that its four chapters (5-8), comprising over 4000 words, were written in a period of only four days. In light of the speed at which I normally write (this article, for example, which is half as long again, required a total of six weeks), this was nothing short of a 'miracle'. The difference was that in spite of the constraints of trying to write in another's voice, I found that my creative decisions were remarkably unconstrained, purposeful - indeed, almost 'automatic'. Over those four days, in effect, I was like a man possessed.
Barthes once wrote: 'I have a disease: I see language' (Barthes 1975: 161). This disease, in any case analogous to the gestalt visions that underpin the literary Creation Myth, is conceived of as such because, after S/Z, Barthes began to move away from conventional criticism toward increasingly fictive (hence scandalous or by his definition, 'novelistic') writings (Barthes 1975: 119-120). This vision is not introspective, as likewise the mapping of a textual genome traces not a (self-) portrait, but a (literary) landscape. It is, again, a map of that landscape, distanced from it, allowing the student writer to be no more inside the writing (unable to see the forest for the trees) than the writing is inside them (as the myth would have it: a 'God-given talent'). It provides a sense of perspective, which characterises the relationship between all readers/writers and texts: 'To know that writing is precisely there where you are not - this is the beginning of writing' (Barthes 1977: 100).
2) McBain's hunch is correct. As Chandler remarked
in one of his letters, 'You never quite know where
your story is until you have written the first draft of it. So I always
regard the first draft as raw material' (Chandler 2000:
78). Return to text
3) For these passages in their unannotated form, see Chandler and Parker 1989: 13-14. Return to text
Barthes, Roland 1991. The Grain
of the Voice: Interviews 1962-1980. Tr. Linda Coverdale. California:
Univ of California Press. Return to text
Barthes, Roland 1977. A Lover's
Discourse: Fragments. Tr. Richard Howard. London: Penguin, 1990.
Return to text
Barthes, Roland 1975. Roland Barthes
by Roland Barthes. Tr. Richard Howard. California: Univ of California
Press, 1994. Return to text
Barthes, Roland 1970a. S/Z.
Tr. Richard Miller. Oxford: Blackwell, 1996. Return
Barthes, Roland 1970b. 'Writing Reading'.
The Rustle of Language. Tr. Richard Howard. Oxford: Blackwell,
1986: 29-32. Return to text
Barthes, Roland 1969. 'The Structural Analysis of Narrative: Apropos of Acts 10-11'. The Semiotic Challenge. Tr. Richard Howard. California: U of California P, 1994. 217-245. Return to text
Barthes, Roland 1967. 'The Death of the Author'. Image Music Text. Ed. and Tr. Stephen Heath. London: Fontana, 1977: 142-8. Return to text
Bloom, Harold 1982. Agon: Towards a Theory of Revisionism.
New York: Oxford UP, 1983. Return to text
Bloom, Harold 1976. Poetry and
Repression: Revisionism from Blake to Stevens. New Haven: Yale UP.
Return to text
Brophy, Kevin 1998. Creativity: Psychoanalysis, Surrealism
and Creative Writing. Melbourne: Melbourne UP.
Chandler, Raymond 2000. The Raymond
Chandler Papers: Selected Letters and Non-Fiction 1909-1959. Eds.
Tom Hiney and Frank MacShane. London: Hamish Hamilton. Return
Chandler, Raymond and Robert B. Parker
1989. Poodle Springs. New York: Putnam. Return
Churchill, Winston S. 1939. 'The First
Month of War'. Radio Broadcast, London. 1 October 1939. In Robert Rhodes
James (ed.) Churchill Speaks: Winston S. Churchill in Peace and War,
Collected Speeches, 1897-1963. Leicester: Winward, 1981: 694-697.
Return to text
James, Henry 1884. 'The Art of Fiction'.
In Leon Edel (ed.) The House of Fiction. London: Rupert Hart-Davis,
1957: 23-45. Return to text
Lodge, David 1996. 'Creative Writing:
Can it/Should it be Taught?' The Practice of Writing. London:
Secker & Warburg: 170-178. Return to text
McBain, Ed 1989. 'Philip Marlowe is
Back, and in Trouble: Review of Poodle Springs by Raymond Chandler
and Robert B. Parker'. New York Times Book Review 15 October:
Return to text
Nesbit, Molly 1987. 'What Was An Author?'
Yale French Studies 73: 229-257. Return to text
Nietzsche, Friedrich 1878. Human,
All Too Human. Tr. Marion Faber and Stephen Lehmann. London: Penguin,
1994. Return to text
Sharples, Mike 1999. How We Write:
Writing as Creative Design. London and New York: Routledge. Return
Stitt, Peter 1985. The World's Hieroglyphic
Beauty: Five American Poets. Athens, Georgia: Georgia University
Press. Return to text
Symes, Colin 1999. 'Writing by Numbers: OuLiPo and the Creativity of Constraints'. Mosaic (Winnipeg) 32, 3 (September). http://web1.infotrac.galegroup.com.simsrad.net.ocs.mq.edu.au/itw/infomark/378/998/69012755w1/purl=rc1_EAIM_0_A56750205&dyn=4!xrn_7_0_A56750205?sw_aep=macquarie Return to text
Gareth Beal recently completed a PhD in Creative Writing at Macquarie University.
Vol 9 No 2 October 2005
Editors: Nigel Krauth & Jen Webb