JAC Volume 4

Teach writing as process not product. Teach writing across the curriculum. These slogans highlight two important concerns of teachers of writing during the past decade. Another concern—the relation of speaking to writing—hasn’t been voiced quite so loudly, but it’s been a continuing interest of researchers and teachers of writing. People have contrasted speaking and writing to explain why writing is so difficult to teach and to learn. Since these contrastive analyses of the modes are usually based on the work of linguists and sociolinguists, it seems useful to survey this work and consider how it has influenced the teaching of writing.

The investigation of speaking and writing has gone through three stages. In the first stage linguists assumed that writing was simply speech written down. “Writing,” Leonard Bloomfield argued, was “not language, but merely a way of recording language by means of visible marks.”¹ During this period linguists assumed that spoken language was primary and they based their analyses on a corpus drawn from the spoken language of informants. For example, the structural description of English sentences which Charles Fries presented in his very influential book The Structure of English was derived from his analysis of telephone conversations.² American structuralists like Fries, however, took little interest in the things that have recently intrigued researchers about conversation: its semantic abbreviation; how, in conversation, text is produced collaboratively by two or more participants; its reliance on shared knowledge to convey meaning. In short, they weren’t interested in those aspects of conversation which distinguish it from writing.

There are several reasons for this lack of interest. I will mention only one. Most people who contrast speaking and writing feel compelled to talk about the different ways meaning is conveyed in the two modes. But linguists during this first stage weren’t interested in meaning—at least in what we usually think of as meaning. They attempted to describe what Fries called linguistic meaning—meaning signaled by the structural arrangement of words in sentences and by the lexical meaning of words; they were not interested in what Fries called social meaning. Fries makes this prophetic distinction, prophetic for our purposes because social meaning becomes very important for later investigators of speaking and writing, at the end of The Structure of English. Throughout his book, he says, he has described only linguistic meaning, but this is only one part of total meaning. Then he gives as an example the sentence “John Smith can swim a hundred yards in forty-five seconds” and says while its linguistic meaning is clear, you won’t be able to understand this sentence unless you comprehend its social meaning, “unless you can fit the linguistic meaning . . . into a social frame of organized information—unless you know. . . that this time is four seconds faster than the world record for this distance” (p. 295).

In summary, during this first stage the influence of linguistics on composition theory was minimal both because linguists were interested in spoken language and because they were interested only in the linguistic not the social meaning of spoken language. Considering social meaning would have brought linguistics closer to rhetoric, the study of how one adapts discourse to make it appropriate for people who have different “frames” of “organized information.” Linguistics during this first stage discouraged new approaches to the analysis of texts because linguists staked out the sentence, not the text, as the domain for investigation. They did not go outside the sentence to consider intersentence relations in a paragraph or the ways a text relates to social and physical context. Because they weren’t interested in texts, linguists during this stage provided no fresh insights to lure teachers away from the “static abstractions” of traditional rhetoric, from the Four Modes of Discourse (narration, description, exposition, persuasion), for example, and the Trinity of coherence, unity, and emphasis.³ Instead they encouraged a handbook approach to composition instruction in which sentence-level exercises on subject-verb agreement, faulty predication, and diction were emphasized.

The second stage in the investigation of speaking and writing originates with the work of Lev Vygotsky and owes much to the work of Basil Bernstein and M. A. K. Halliday and Ruqaiya Hasan.⁴ In Thought and Language, translated into English in 1962, Vygotsky argued that conversation, at least a conversation between two people who know each other well, is closer than writing to the inner speech of thought. Like inner speech it is semantically abbreviated, containing sentences with the subject removed and only the predicate remaining. In conversations friends can speak in code words and elliptical phrases because the speakers can rely on shared knowledge to fill in the semantic spaces (pp. 138-148). Such speech does not require the “deliberate structuring of the web of meaning” that writing demands (p. 100).

Later Bernstein, who has admitted his debt to Vygotsky, was to call speech which relies on shared knowledge restricted; and speech which doesn’t assume shared knowledge elaborated. Bernstein argued that British working class speakers used this restricted speech in certain situations because they were socialized in a particular way, because their parents used this restricted speech in what he called “critical socializing contexts”—when they disciplined them, for example, and when they taught them skills (p. 198).

After Hasan wrote a paper on cohesion in which she distinguished endophoric and exophoric reference,⁵ Bernstein was able to explain in more linguistic detail how these two variants differed. Endophoric reference was reference to something in the text; exophoric reference was to something in the context of situation. If a speaker describes a city and then concludes by saying "This is the capital,” the reference is endophoric. If the speaker points to a city, or to a map of a city, and says “This is the capital,” the reference is exophoric. Drawing on Hasan, Bernstein called a restricted speech variant context-dependent and an elaborated speech variant context-independent. The examples above are of endophoric and exophoric reference in speech. This is appropriate because Bernstein used Hasan’s insights to distinguish the speech of different social classes in Britain, but socio-linguists and composition theorists—E. D. Hirsch, for example—have used Bernstein’s terms to distinguish typical spoken from typical written discourse: speech was context-dependent and writing was context-independent.⁶ By context­dependent most researchers did not mean only discourse containing exophoric reference. They meant discourse (1) which relies on paralinguistic (gestures, intonation) as well as linguistic devices to convey meaning; and (2) which can only be understood if one shares a great deal of background knowledge with the speaker or writer.⁷

Vygotsky and Bernstein have impressed composition theorists—James Moffett, James Britton, and Linda Flower, for example—with the difficulties writers encounter in transforming private inner speech into public written texts.⁸ Vygotsky stressed the similarities between thought and informal conversation. In his view, thought was inner speech and inner speech was related genetically and structurally to informal conversation. It was related genetically because it developed from verbal exchanges with parents and friends which gradually became interiorized as thought. “[T]he child starts conversing with himself as he has been doing with others,” Vygotsky explained (p. 19). It was related structurally because inner speech, like external speech between close friends, contained predicates without subjects, personal code words, and other features leading to implicit not explicit meanings. Because Vygotsky linked thinking and speaking and stressed how greatly both these activities differed from writing, he encouraged composition theorists to see the process of “putting thoughts down on paper” as a process not of transcribing but of transforming talk. Students, composition theorists came to believe, had to abandon strategies and structures used in conversation (exophoric reference, gestures, abbreviated syntax, code words with personal meanings) and adopt strategies and structures more appropriate for writing (endophoric reference, explicit syntax). To help students with this transformation, Moffett designed a structural curriculum which moves from types of writing which resemble thought and speech (interior monologues, correspondence) to more public writing (reports, essays) which demand communicative strategies and styles very different from those used in conversation.

In their very influential report the Development of Writing Abilities (11-18), James Britton and his co-researchers argued that “what children write in the early years should be a form of written-down expressive speech”; later they should learn how to transform this writing into less personal, more explicit “transactional” writing (p. 82). Flower has also been interested in helping writers move from personal to public modes of expression, but unlike Moffett and Britton she does not view this transformation from a developmental perspective. She sees it as a stage in the process of writing, not as an episode in the life of a writer. Many adult writers, she says, will write “writer-based prose”—the “adult written analogue” of egocentric speech—because they can’t juggle all the constraints that writing demands (p. 22). Writers, Flower argues, must produce “reader-based” texts with the web of meaning fully structured and explicit but they also have to dredge up information from memory, make sure their texts achieve their rhetorical purpose, and give their writing a personal voice. Inexperienced writers can’t deal simultaneously with all these demands, Flower suggests, and thus we should let them produce writer-based prose first and then teach them how to transform it into reader-based writing.

Flower describes writer-based prose as a way-station on the road to more explicit reader-based writing. For her it’s a means to an end. Both Moffett and Britton find more inherent value in writer-based, or expressive, writing than does Flower: Moffett says students should “learn to play freely the whole symbolic scale”—practice and enjoy reading and writing texts that fall at many points on his continuum (p. 28); Britton says a writer never outgrows his need for expressive language and quotes two letters of Keats to impress upon us the value of “mature” expressive writing (p. 84). Both Moffett and Britton, however, set up continuums from expressive writing which is close to speech to explicit transactional writing which is far removed from it and people tend to see continuums as developmental progressions. As a result of the work of Vygotsky and his followers, composition theorists began, during this second stage, to view writing as a long uphill struggle away from semantically abbreviated conversation toward the beautiful explicitness of “formal essayist prose.”

This last phrase comes from an article by David Olson called “From Utterance to Text: The Bias of Language in Speech and Writing,” an article that has reinforced the linear view of progress in writing ability emerging from the work of Vygotsky and his followers.⁹ According to Olson, both historically and developmentally (individually) there is movement from utterance—context-dependent statements—to text—explicit, context-independent statements. There are two aspects to the historical movement from utterance to text. First, a more explicit writing system—the Greek alphabet—developed from less explicit logographic and syllabic systems. These latter systems, Olson argues, were more dependent on knowledge shared by reader and writer because they had no signs for vowels; words were indicated by consonant signs only and thus texts were open to different interpretations depending on context. Alphabetic systems, which unlike syllabaries had signs for vowels as well as consonants, enabled writers to produce more explicit texts. This according to Olson, was an advance at the level of the grapheme. There was also a second advance at the semantic level. According to Olson, written texts gradually became more explicit not only because words in sentences could be recorded unambiguously using the new alphabetic writing but also because writers began to want to produce texts that could stand on their own, i. e., that could be interpreted by people who didn’t know the author or share his background knowledge. Writers wanted to produce “formal essayist prose,” a kind of writing which, Olson says, was encouraged by the invention of printing which enabled writers to communicate with more “distant” readers—readers whom they didn’t know personally and who might not share their personal premises and knowledge.

In saying that developmentally there is movement from utterance to text Olson means that as children grow up they learn to encode more of their meaning in the text itself and rely less on context. For example, a young child may say “Mommy sock” to refer to his mother putting on his sock or to his mother’s sock: the phrase is ambiguous out of context. According to Olson, it is the function of schools to wean children from such context-dependent utterances and get them to produce context-independent texts (p. 278).

Olson has influenced composition theory. Flower uses his distinction between utterance and text to explain the constraints imposed by writing which one must “juggle” in the act of composing. Writers have to make their writing “fully contextualized” as they tend to a host of other things, a constraint that doesn’t burden speakers.¹⁰ In his theoretical work on composition and style, George Dillon says Olson’s categories “somewhat modified and amplified, give us a way of describing where the student in a freshman composition class is coming from and going toward.”¹¹ We should see the problems of student writing, he says, “not as crudities or diseases but as manifestations of an incomplete substitution of the conventions of text for those of utterance” (p. 49). Before the work of Vygotsky, Bernstein, and Olson became well known, composition teachers had already begun to concentrate on only one mode—exposition. Because in Olson’s scheme the explicit, unambiguous expository text is the ideal, the end point, of historical and individual development, he has reinforced this narrowing of the goals of composition instruction. If we listen to Olson our purpose becomes to teach students how to write texts that readers can decode without making inferences and without drawing on any personal knowledge they may share with the writer.

These second-stage analyses of speaking and writing have also influenced instruction in another way: by becoming the theoretical foundation for the language deficit hypothesis, the notion that the language of members of some subcultures of the U.S. is an inferior medium of communication and thus the people who speak these deficit languages need compensatory education. How did this come about? First, researchers connected experience with context-independent discourse to the development of abstract thought. Patricia Greenfield makes this connection very explicitly in an article entitled “Oral and Written Language: the Consequences for Cognitive Development in Africa, the U. S., and England.” “Abstraction,” says Greenfield, “is, as the etymology of the word suggests, a ‘separation from.’" “ It is “the mental separation of an element from the situation or context in which it is embedded.”¹² Therefore speakers from an oral culture who produce and listen to texts that are context dependent, which are not separated from situation, are not developing their capacity for abstract thought. In Piagetian terms, their cognitive development becomes arrested at the period of concrete operations: they can perform certain mental operations—classifying, for example—on objects that they can physically manipulate, but they perform these same operations less well than schooled children and when the problem is presented verbally, they may not be able to solve it all. Greenfield, for example, found that unschooled Wolof children in Senegal could not categorize objects as skillfully as schooled children. She concludes that because schooling, particularly writing, provides “practice in the use of linguistic contexts as independent of immediate reference,” it leads to the development of abstract thought (p. 174).

Robert Hess and Virginia Shipman explored this context-dependent vs. context-independent distinction in the speech of mothers interacting with their children.¹³ In explaining how to do a sorting task, lower class mothers, they found, used context-dependent discourse while middle class mothers used context-independent discourse. Hess and Shipman conclude with a definition of language deprivation: ‘The meaning of deprivation,” they argue, “is deprivation of meaning” (p. 885). The meaning is “deprived,” according to Hess and Shipman, because it is conveyed by gestures and elliptical phrases rather than by elaborate verbal cues.

Composition theorists such as Thomas Farrell and Andrea Lunsford argue for a similar deprivation of meaning in the essays of basic writers.¹⁴ Farrell merges the historical and developmental arguments. Many basic writers, he says, lack the cognitive skills to write well because they come from residually oral culturals which do not value writing, or talk which is like writing (analytical, detached from situation). Therefore they are in the same situation the whole Western world was in before writing became widespread. Britton et al. apply this “Ontogeny recapitulates phylogeny” argument to all developing writers, not just basic writers from minority cultures. They suggest that practice writing transactionally will advance students to formal operational thought just as the arrival of literacy improved the mental ability of preliterate peoples (p. 202). Lunsford is less interested than Farrell and Britton et al. in the causes of the alleged cognitive deficit of basic writers but like them she believes that basic writers write badly because they “are operating well below the formal-operations or true concept stage of cognitive development” (p. 260).

This belief that writing has cognitive consequences, one of the legacies of these second-stage analyses of speaking and writing, also fuels the writing across the curriculum movement. A key principle of this movement is that writing to learn and learning to write go hand in hand. Students will become better thinkers and writers if they write more—in all subjects, not just English. In “Writing is a Mode of Learning,” the seminal article for the writing across the curriculum movement, Emig argued that writing is a good mode of learning because, among other things, it allows students to practice moving from private inner speech to public written speech; because, unlike talking, it “restrains dependence upon the actual situation”; and because it requires that “deliberate structuring of meaning” which Vygotsky described.¹⁵ Here, as in other second-stage analyses, writing is seen as completely different from speaking and the distinguishing feature emphasized is writing’s independence from context.

A rejection of dependence on context as the distinguishing feature of speaking characterizes the third stage in the exploration of speaking to writing, a stage which has just begun. Researchers have pointed out, for example, that an oral speech—a formal lecture, for example—can be very explicit and context­independent and a personal letter very dependent on shared understanding. People have begun to see that earlier comparisons of speaking and writing weren’t comparisons of speaking and writing but of two different genres: casual conversation and formal essayist prose. Formal speech and formal writing are not very different. Linguists and rhetoricians are now abandoning the speaking-writing distinction in favor of other oppositions: dialogue-monologue, unplanned-planned discourse, involvement-detachment, and oral strategy-written strategy.¹⁶ The first member of these pairs is usually associated with speaking, the second with writing, but advocates of these new oppositions carefully point out that they cut across the modes: a monologue may be oral, an oral text may be planned, an oral text may exhibit detachment, and a literate text may reflect oral strategies.

Deborah Tannen’s discussion of oral and literate strategies is a study typical of this third stage.¹⁷ Tannen defines oral strategies as “aspects of discourse which make maximal use of context, by which maximal meaning and connective tissue are implied rather than stated”; written strategies, on the other hand, are aspects of discourse “by which maximal background information and connective tissue are made explicit.” In addition, oral strategies “depend for effect on paralinguistic and non­verbal channels, while literate strategies . . . depend on lexicalization to establish cohesion” (p. 3). Although this sounds like a typical “stage two” distinguishing of the two modes based on relation to context, it’s not this at all because Tannen is talking about strategies not modes. According to Tannen, both oral and literate strategies are available to both speakers and writers. For example, an oral storyteller may adopt a literate strategy and make explicit the point of his story, or how the different parts of a story relate, by including such statements as “Boy, was I scared!” or “And this was the best part.” Or an oral story teller may adopt an oral strategy and present his narrative in such a way so that the point of the story has to be inferred (p. 4). A speaker, asked to provide an oral summary of a silent film he has just seen, may adopt an oral strategy and refer to a male character mentioned earlier in his summary by using a special high rise-fall intonational contour: “the man . . . the man came out the tree”; or he may adopt a literate strategy and make the cohesive tie by using a restrictive relative clause: ‘They walked by the man who was picking the pears.”¹⁸

According to Tannen, oral strategies, whether they occur in spoken or written texts, encourage the audience to become involved in the process of creating meaning. She and Margaret Rader object to Olson’s linear view of language development from implicit utterance to explicit essayist prose because it leaves out imaginative fiction, a genre which is sophisticated, written, and relies on contributions from the reader. “Nothing intrinsic to the medium of writing,” says Rader, “dictates that no contribution should come from the reader.”¹⁹ Good writers, in fact, encourage reader involvement and take advantage of the opportunity to revise provided by writing—but not by speaking—to achieve it. An oral story teller signals how we are to take a story by facial expressions, intonation and gesture; deprived of these props, writers must carefully craft and recraft their sentences, choosing words with just the right connotations, selecting details which will trigger the appropriate associations. The trick, says Rader, is first to get readers imagining and then constrain that imagining so it proceeds in the desired direction.

Rader discusses imaginative literature but concludes her article by questioning whether any texts—even expository essays—are fully explicit. Even Olson admitted that the fully explicit text was probably only a desirable goal, not a reality. Rader says it’s a dream: How could we understand an article in a scientific journal, she asks, if we possessed only a grammar and a lexicon of the language? (p. 197). Marilyn Cooper also believes that written texts are, like oral texts, dependent on context; in fact, she says, there’s really no difference between communicating in speaking and communicating in writing: the "fundamental communicative process" is the same in both modes.²⁰ Cooper’s position is more extreme than that of most researchers. The common practice has been to argue that typically oral texts do exhibit a particular feature—dependence on context, for example—but that this feature is found in written texts as well. To support her assertion that communicating in speaking is the same as communicating in writing Cooper analyzes oral conversation and then some written texts and shows that in neither is the meaning carried by the syntax and lexicon alone. To understand spoken and written texts one has to know (tacitly) certain “maxims” for conversation identified by Grice²¹ because these maxims are purposely “flouted” to convey a particular mcaning. Consider a child, Cooper says, who arrives home from school two hours late. When he says “Hi, Dad,” his father replies, “Where have you been?” The father “flouts” a maxim (Greetings should be reciprocated) to implicate that he is angry (p. 116).

Linguists have for some time, however, agreed that meanings in oral conversation are implied, not explicit. To prove her more controversial point that the meanings of written texts are equally implicit Cooper describes several written texts including a Winston cigarette ad which claims “Nobody does it better.” The ad, Cooper says, flouts Grice’s maxim of manner (Avoid obscurity of expression) by not providing a clear antecedent for “it.” Readers must supply antecedents and thus the text is not autonomous; all its meaning is not in the text (p. 126).

This example is similar to the sentence about the swimmer who swam 100 yards in 45 seconds that Fries used to distinguish linguistic from social meaning. We can’t fully understand either the cigarette ad or Fries’ sentence unless we can relate it to what Fries called “a social frame of organized information.” Fries didn’t think linguists should investigate these frames, but linguists now—in what I’m calling stage three—are convinced that they must. In investigating them they are clarifying what is meant by context. Stage-two linguists used context rather imprecisely to refer both to physical setting and background knowledge related to the topic of discourse. Recently linguists have described context more precisely and in the process have increased our understanding of speaking and writing. Margaret Himley, for example, discusses five contexts all of which, she says, influence how meaning is conveyed in texts.²² She distinguishes physical context from psychological context—the knowledge structure of the participants. An exit sign conveys meaning, for example, when it is placed in a particular physical context—over the side door of a movie theater. Other texts convey meanings by evoking “frames” or “scripts” which are part of an individual’s knowledge structure. Van Dijk defines frames as “knowledge representations about the ‘world’ which enable us to perform such basic cognitive acts as perception, action, and language comprehension."²³ Frames have been investigated by researchers in artificial intelligence interested in devising programs that would enable computers to comprehend natural language.²⁴ These researchers found that texts require readers to make inferences to fill in the semantic spaces between propositions and that readers can make these inferences because they know the relevant frames. A computer ignorant of frames and programmed to respond only to syntax and lexicon will never be able to understand natural language texts. For example, Cooper says the cigarette ad mentioned above (“Nobody does it better”) conveys meaning by evoking three frames: a product frame, which encourages the inference that no one produces better cigarettes; a worker frame evoked by the line and the accompanying picture of a hard-hatted workman; and a “sexual relations” frame evoked by the allusion to the popular song from which the line comes (p. 126).

How can these third-stage analyses of speaking and writing assist composition teachers? One way is by reminding us that written texts do not have to be fully explicit to succeed. In fact, Dillon, who generally accepts Olson’s view that it’s our job to move students from context-dependent utterances to context-independent texts, points out that the essay from Montaigne to the present has never been a completely context-independent form. He questions Olson’s lumping of the essay tradition of Montaigne with the objective, scientific language movement of the Royal Society and Bishop Sprat. As the essay developed, he says, it began to include puns, aphorisms, paradoxes, and figurative language. Context-independent prose wasn’t the goal at all. Dillon suggests that we should be true to this tradition and “value writing which shows richness, resonance, and a sense of larger implications.” Unfortunately, he says

To relate these theoretical disputes even more closely to practical problems, consider a student in a typical English 101 course which begins with personal writing and ends with expository writing. Early in the term the student is told to write a personal narrative, which she does, taking pains throughout to tell the reader exactly how she felt at different moments; she ends with an explicit statement of the meaning of the experience for her. Her teacher tells her to show, not tell, how she felt and to give readers more freedom to grasp the point of her narrative for themselves. “Don’t insult your readers by doing their thinking for them,” her teacher tells her. Later in the course, in commenting on an expository paper, the teacher tells the same student to spell things out more clearly and to make more explicit how she got from point A to point B. The student has a right to be confused, but if teachers, encouraged by third-stage analyses such as those of Tannen and Himley, realize that different written texts may exhibit different degrees of context­dependence, then they may be able to turn such confusion into a profitable discussion of different strategies appropriate for different kinds of texts.

In discussing context-dependence and independence and oral and literate strategies with students, teachers may find the continuum a more useful conceptual tool than the binary opposition. In other words, instead of speaking of discrete categories like utterance and text and expressive writing and transactional writing it may be more useful to speak of varying degrees of context dependence. Himley suggests a continuum ranging from exophoric (context-dependent) to endophoric (context-independent) texts (p. 20). She arranges these sample items along it: an exit sign, a letter to a friend, a business letter, a prepared statement at a public meeting, an essay, a lecture, and a journal article. In working with a student, we can place her paper on this continuum and then discuss whether the degree of context-dependence it exhibits is appropriate given her purpose and audience. Because the kinds of writing done within organizations exhibit varying degrees of context-dependence, Himley’s continuum should prove useful to teachers of technical and business writing.²⁵ A project report written to a superior familiar with the project, for example, could be more exophoric than a report written for someone in another division or firm.

What I’m suggesting is that we should see context-dependence as a flexible feature that can be adjusted to achieve a particular rhetorical purpose. We should not view it, as some stage-two analysts did, as a sign of arrested cognitive development. In certain situations, when one is describing a picture present to both speaker and hearer, for example, context-independent discourse would be unnatural and inefficient. It is misleading to call discourse produced in such situations restricted or deprived simply because not all the meaning is conveyed through the verbal channel. In writing as well as in speaking it is sometimes more effective and efficient to keep one’s text implicit.

Fortunately scholars are beginning to see context­dependent discourse not as a sign of arrested cognitive development but as a response to pressures in the rhetorical situation, as a manifestation of a natural rhetorical competence. James Collins summarizes arguments against explanations of context-dependent discourse based solely on cognitive deficit and urges us to view writing development within a "developmental-functional framework" which allows for consideration of rhetorical, educational, and social contexts.²⁶ An article by Richard Ohmann in College English exemplifies this new direction.²⁷ Ohmann presents a transcript of two interviews, one with two working class people and one with the mayor of a small Connecticut mill town. The speech of the workers is very restricted in the Bernsteinian sense, the speech of the mayor very elaborate. But Ohmann explains the difference by referring not to social class or cognitive ability but to the different rhetorical situation of the workers and the mayor during the interview when the videotape was made.

Michael Gregory and Susanne Carroll discuss another continuum that teachers should find useful.²⁸ It begins with language written to be spoken as if not written (drama, for example), moves to language written to be spoken (a speech, for example), and ends with language written not necessarily to be spoken. Under this last heading they include: language written to be read as spoken (dialogue in a novel, for example), language written to be read as thought (interior monologue, for example), and language written to be read (a telephone book, a dictionary, for example). Himley relates Gregory and Caroll’s continuum to the exophoric-endophoric continuum she proposes. As one moves from language written to be spoken as if not written to language written to be read as if not spoken, the texts become increasingly endophoric, i. e., more and more information is explicitly realized in the text (p. 14).

Gregory and Carroll’s analysis is useful because few students have thought about how writing relates to speech: they assume the two modes are completely unrelated. I’ve found that asking questions chosen to elicit some of the insights revealed by Gregory and Carroll’s scheme leads to a lively and useful class discussion. Here are some questions I have asked:

This last question is important. Although students know they aren’t supposed to write so their essays read like a telephone book, many seem to think all writing should read like a dull, voiceless government document. Many teachers, however, want students to write their essays so they will read like speech—so they will have an authentic voice. Discussing Gregory arid Carroll’s continuum and representative texts which would fall at different points along it might help close this gap between what students assume and teachers expect.

Robin Lakoff explains why it’s important for teachers and students to understand how writing relates to speech. She calls our attention to an important change in values: before, she says, we prized a “literary-based model” of communication; now we prize a model based on oral discourse. We now expect most written documents to be written so they will read like speech, to be “couched in forms imitative of the oral mode.”²⁹ If this is so—if, to be effective, modern prose should read as if it were heard—then we should teach students how to make their writing read this way. But how? What else can we do besides learn to think in terms of continuums instead of binary oppositions? Linguists help us by providing theoretical insights but they rarely offer solutions for instructional problems. Tannen, for example, suggests that “oral strategies may underlie successful discourse production and comprehension in the written as well as the oral mode,”³⁰ but she makes no pedagogical recommendations. Her definitions of oral and literate strategies become confusing when one tries to apply them to the process of writing. The word strategy suggests a conscious intention within the mind of the speaker/writer—a plan that might be realized by different devices; but her strategies are also “aspects of discourse,” a phrase which suggests features intrinsic to the discourse. Thus one isn’t sure whether her strategies apply to the process or product of composing. Reading Tannen leaves one conscious of the need to teach oral and literate strategies but uncertain as to just what these strategies are.

Mina Shaughnessey and others have pointed out that students often produce dull and syntactically snarled writing because they are trying too hard to write in a bookish style.³¹ To get these writers to write more clearly and less stuffily, should we encourage them to write as they talk, have them do a lot of expressive writing (the mode which according to Britton is closest to speech), introduce them to Studs Terkel’s works and have them transcribe and edit tapes of colorful speakers speaking? Or should we have them read as much good writing as they can, recommending, in particular, works by writers who have achieved an authentic voice?

Probably we should do both these things: encourage students to realize that lively, interesting talk is a good model for lively, interesting writing; and urge them to read good writers in the hope that they will pick up techniques of imitating speech in writing. Britton hints that such a two-pronged attack would be appropriate when he says that as students learn to adapt their multi-purpose expressive writing into transactional, poetic, and mature expressive writing, “their mode of doing so will be a kind of shuttling back and forth between their speech resources on the one hand and the written forms they meet on the other” (pp. 82-83).

I say students should “pick up” ways of imitating speech in writing because I’m not sure these techniques can be taught as conscious strategies. An informal, “conversational” writing style is not achieved by simply transcribing talk. This is so because we cannot capture in writing all the phonological features of speech and because we need to pack into our written sentences words that will help our readers imagine the context. In other words, writers who employ what Tannen calls oral strategies may not want to make their prose maximally explicit, but they will have to make it more explicit than conversation and in doing so they will make their writing something other than a direct transcription of talk. Norman Page points out that writers usually exploit a few devices to create the illusion of speech; they don’t try to write what a person would actually say. He quotes a passage from Alan Sillitoe’s The Ram’s Daughter and demonstrates that Sillitoe uses certain lexical items— coppers, nicker, no-good—to give the impression of lower-class speech. The rest of the passage—the syntactic patterns of the sentences, for example—are typical of standard written English.³² In short, an informal style imitative of oral discourse is a complicated blend of oral and literate strategies: it’s a tough act to pull off successfully. Maybe experienced writers are able to achieve this style deliberately. Lakoff argues that Tom Wolfe in The Right Stuff uses a variety of devices--"italics, quotation marks, capitalization and other aberrant punctuation devices, ellipses, fragments, expletives, dialectal and colloquial forms, and much, much more" to capture the nuances of speech (p. 255). But inexperienced writers can probably acquire an informal style most easily by reading and imitating writers who have achieved it.

Page helps us understand the complexity and variety of ways speech can be represented in writing by discussing a piece of dialogue from Dickens’s Martin Chusslewit. Grammar books usually discuss only direct and indirect speech, he says, but there are many other ways of presenting dialogue. Here is part of the dialogue as Dickens wrote it in direct speech: "There are some happy creeturs,” Mrs. Gamp observed, “as time runs back’ards with, and you are one, Mrs. Mould . . ." And here are some other methods of speech presentation that Page identifies:

Here we also have a continuum, one that can be related to Gregory and Carroll’s. Because they would occur in a novel (and not a play, for example), all of these versions can be considered language written not necessarily to be spoken. But some are much closer to the direct speech on which they are based than others. As one moves from version 6 to 1, one moves from lan­guage written to be read more or less as spoken to language written to be read more or less as written.

Even the above analysis oversimplifies the situation, however, because not only dialogue but also the language of the narrator may be imitative of oral speech. In considering how the language of the narrator relates to speech we need to examine what we mean by imitation in literary works. Since Aristotle’s Poetics imitation has been seen as a distinguishing feature of literature: creators of literature, critics have explained, do not give us real objects and events but imitations of these things. Barbara Herrnstein Smith, however, points out that poets (understood in the broad sense, as creators of literature) imitate in another, more important, way: they imitate verbal acts.³³ Sometimes they imitate speech acts, sometimes text acts. Lyric poets, says Smith, imitate the act of speaking a personal utterance; novelists usually imitate the act of writing. In David Copperfield, for example, Dickens imitates the act of a person writing his autobiography.³⁴ In the opening lines to this story, the reference to “pages,” the use of the verb “record,” and the bookish sentence structure (long periodic sentences, parallelism, for example) alert readers that they are supposed to accept the fiction that they are reading a written, not a spoken, account of David’s life:

Lakoff reminds us that we now expect many written documents, including personal essays and autobiographies, to be written in a language which imitates the oral mode. This imitation is different from, but related to, the imitation that Smith associates with poetry and novels. A personal essay or an autobiography is what Smith calls a “natural utterance”: a discourse produced by a “historically real person, occasioned by a historically real universe” (p. 20). In this sense it is not an imitation; it is the real article (in both senses of the word). What may be imitated in a personal essay or an autobiography (or informal letter or memo) is not a verbal act but a verbal style: we may try to imitate the way we speak (or think we speak, or wished we spoke). This imitation of verbal style is related to Smith’s imitation of verbal acts. When we read modem autobiographies we feel as if the narrator is talking to us, not writing to us. Therefore a novelist writing a novel told in the first person may make her narrator sound as if he is speaking in order to create the illusion that her book is an actual personal account of a man living in the twentieth-century. Holden Caulfield, the narrator in Salinger’s Catcher in the Rye, is supposedly writing his story while in an institution, but Salinger makes Holden’s writing read like talk:

Teachers can discuss with students the different ways we imitate speech in writing. Since the issues are complex, however, I don’t recommend extended discussion—just enough to alert students to the various ways writers can imitate verbal acts and styles. Every discussion should include examples. Teachers can also have students try their hand at imitating some professional imitations of verbal acts and speaking styles.³⁶ Another useful assignment is to invite students to compare natural utterances with their literary imitations—to compare, for example, a transcript they’ve made of their own inner speech with the interior monologues of characters from novels by Virginia Woolf or James Joyce; or their own personal letters with the letters in epistolary novels like Alice Walker’s The Color Purple; or a taped informal conversation with a literary dialogue.

In charting the developmental relationships between speaking and writing Barry Kroll discusses a consolidation phase during which children learn that writing is talk written down; then a differentiation phase in which they learn that many written texts must be more explicit and context-independent than informal spoken discourse; then a phase Kroll calls systematic integration in which students learn that good writing is in many ways imitative of oral discourse.³⁷ Kroll’s developmental progression parallels the development of linguistic analyses of spoken and written texts that I have just sketched, and the lesson, it appears, is that research insights recapitulate ontogeny. In any event, I do believe we are arriving at a more mature appreciation of the relationship of speaking and writing, one which betters accounts for the variety and complexity of texts in both modes.

Kroll’s analysis suggests that helping students imitate speech in their writing is a proper goal for an advanced composition class. I agree. The production of context-independent, explicit texts is too narrow a goal. Advanced writers should be able not only to make their writing explicit but also, when they wish, to make it resonate with larger implications (to echo Dillon); and they should know how, when they wish, to give their writing voice so it reads not like a dictionary or a bad computer manual but like a person, talking.