A Framework for Writing

Context and Purpose: Where, for Whom, and Why?

A hypothetical list of distinctive types of writing, each with its unique purpose and unique form, would presumably be limited only by our patience. We could discuss writing according to the place where it is done—writing done at home, in school, in the workplace, in community and government institutions. Such broad contextual categories suggest some obvious categories of writing. A teenager's self-sponsored home writing might consist of text messaging, letters, and/or a diary, whereas school-sponsored writing would include book reports, essay test questions, science lab reports, and so forth. Writing can be a solitary activity, with absolute silence from start to finish, or for a first grade child it could be a social activity, done in the context of talking, playing, and drawing with classmates. Something is written to be read by an audience. The audience may be the writer (e.g., a locked personal journal), only one other person (a report for a teacher), a group of known people (the minutes of a faculty meeting), or a group of unknown people. Mature writers "write to an audience"; as they write they are inside the mind of the eventual reader, constructing the reader's response and adjusting their writing accordingly. In addition to the obvious communicative purposes of writing, texts also store and preserve information. A newspaper article about a family member is cut out for a scrapbook; a difficult article is put aside for later re-reading. Writing is all of these variations and many more. The contexts and purposes for writing change dramatically over the course of elementary and secondary schooling and beyond.

Linguistic Form

For purposes of this chapter, text is defined broadly as a piece of writing done for a particular purpose. For a young child, a text might be a few words that accompany a drawing. For the tenth grader, the five-paragraph theme supporting a point of view is a text. A written genre is a distinctive type of text—for example, narrative text, persuasive text, factual text, and so forth. To write in a particular genre is to conform to a particular set of linguistic constraints at the text level (e.g., narratives start with a setting) and also at the sentence level (e.g., narratives usually employ simple past tense forms of verbs). Conveniently, then, writers do not have to invent a new form each time the same situation arises. Genre acquisition cannot be reduced to learning a specified set of skills; however, writers must gain at least some control over major formal features of genre (Popken, 1996). Text and sentence-level features of narrative and expository (informational, factual)¹ genres important in school writing have been described elsewhere by Scott (1988, 1994, 1995) and Westby (1994).

In addition to linguistic constraints imposed by genre, written sentences have distinctive grammatical properties that arise from modality—the fact that they are written rather than spoken (Perera, 1984, 1986; Scott, 1988, 1994, 1995, 2002). Without direct teaching, children's writing takes on this distinctive "written" grammatical flavor at an early age.

Writing requires other types of linguistic knowledge; words must be spelled and sentences must be punctuated. Whereas spelling and punctuation are frequently described as lower level "mechanical" activities, learning how to spell and punctuate are more recently seen as cognitive-linguistic activities of considerable dimension. For example, punctuation used by novice writers is said to reveal much about children's developing metasyntactic and metatextual knowledge (Kress, 1982; Simone, 1996).

The Process of Writing

Among those who model the writing process, there seems to be unanimous agreement that it is a complex mental process (Bereiter & Scardamalia, 1987). Compared to speaking, writing requires a high level of abstraction, elaboration, conscious reflection (Gombert, 1992), and self-regulation (Sexton, Harris, & Graham, 1998; Singer & Bashir, 1999). Models of the writing process are not concerned with very casual sorts of writing, for example, dashing off a note to a friend or a quick reminder memo. Rather, models of writing attempt to explain the composition process—how we would proceed to write an essay, or a report, or a story of some length. Another term for this type of writing is epistemic writing—the type that both advances the writer's knowledge of a topic and is credible to the reader (Bryson & Scardamalia, 1991). One well-known model conceives of writing as a problem-solving activity with three overlapping and even recursive stages² (Hayes & Flower, 1980, 1987).  Writers first develop an internal representation of the problem and establish goals (e.g., "write a paper that effectively describes deforestation of the American Northwest, is at least five pages, and earns an A from the teacher"). In the planning phase, writers select information from their knowledge base and organize that information for an effective presentation. Subprocesses of planning include (1) generating relevant information by retrieving it from long-term memory, (2) organizing the retrieved information, (3) setting goals for the text and criteria for its evaluation, and (4) developing "en route" strategies for completing the paper (Black, 1981; Graham & Harris, 1993). Mature writers draw on their knowledge of text structure during the planning stage (e.g., "I need to follow a point-counterpoint structure for each of the three points"). Topic knowledge alone does not necessarily ensure clear writing.

In the generation phase, pen is put to paper and text is produced. Writers must now choose the words and structures that encode the meanings they wish to convey. Hayes and Flower (1987) reported that ideas in an outline are expanded by mature writers on the average by a factor of eight as text is actually generated. Writers work by producing a part of a sentence, pausing, generating the next part, pausing, all in a left-to-right manner. By studying the types of errors writers make, researchers have gained insight into the nature of the text generation process (e.g., Daiute, 1984).

Revising is the final phase of the composition model. In an attempt to improve the text, writers make changes that range from changing a word, adding a comma, to reorganizing or adding/deleting major portions. Research challenges the view of revision as an end-stage process, stressing instead the recursive nature of revising (Witte, 1983). Older writers and expert writers devote proportionally more time to revising and make changes involving larger stretches of discourse and text meaning. Revisions of younger writers and novice writers are more frequently devoted to the word or sentence level and are less apt to change the meaning.

This three-stage model of writing was devised in response to protocol analysis, a research paradigm in which writers are asked to think aloud (Hayes & Flower, 1980, 1987) as they write. Verbatim transcripts of what writers say, along with observations of what they do (e.g., analyses of pause behavior), make inherently private cognitive processes more accessible for study. Recently, think-alouds have been used to study how writing processes change with development and whether different processes are used by children with learning disabilities.

Whereas the Hayes and Flower model was designed as a model of mature, or expert writing, several researchers in children's writing have suggested modifications tailored to developmental writing. Berninger, Fuller, and Whitaker (1996) suggested eight modifications that better account for beginning and developing writing. For example, they suggested that the generating (translating) phase of writing be divided into two components: transcription (the translation of language representations in working memory into written symbols) and text generation (transforming ideas into language in working memory). This division better accounts for observed disassociations in developmental writing—children who can generate ideas but not transcribe these onto the page—and less often, the opposite—children who can transcribe but have little to "say." A second developmental perspective was provided by

Bereiter and Scardamalia (1987) who distinguished between novice and mature writing in their work. The novice writer uses a knowledge-telling model to generate a text. Like the expert writer, the novice writer sees composition as a problem to be solved, but a different kind of problem—namely one of accessing enough relevant information to satisfy length and genre requirements for the writing assignment or task. Bryson and Scardamalia (1991, p. 45) illustrated this model for the hypothetical writing assignment "Is television a good influence on children?" Two types of cues are used to generate text: topic identifiers such as "television shows" and "children," and discourse knowledge (e.g., "say what you think and then give reasons"). These cues are the entree to long-term memory where information is called up and transcribed in "think-say" cycles. The composition moves forward in a linear, sometimes associative, manner in which the mention of one point can trigger the mention of an associated point (e.g., "It's good for children to watch comedy shows. My favorite comedy show is ... ").

Expert writing, on the other hand, is described in a knowledge-transforming model (Bereiter & Scardamalia, 1987). This writer is guided by a "discover what I know" rather than a "tell what I know" mandate. Presumably, the discoveries would not have come about without the act of writing. Differences between the novice and expert models are evident in longer start-up times and more extensive note-taking. Think-aloud analysis reveals an active "dialectic" between content (what the writer knows and believes) and rhetorical issues (how the writer should best say it); in the course of this internal conversation, thinking can evolve in new directions (Bryson & Scardamalia, 1991). The knowledge-transforming model is stressed increasingly in current writing pedagogy literature, particularly in response to what some perceive as an overemphasis of "vacuous" writing process instruction (Writing and Thinking, Interview with Leif Fearn, 1996). Some writers continue to use a knowledge-telling model throughout their school years and beyond.

Two additional views of the nature of writing have surfaced in recent years. First, the social-interactive model of writing (Ny strand, 1989) sees writing as fundamentally an interaction of minds—that of the writer and the hypothesized or real reader. Writing is an act of thinking as the reader would think and making text adjustments that result in a better communication. Whereas problem-solving models focus on cognitive processes of the writer, the social-interaction model emphasizes two sets of cognitive processes (the writer and the reader) and their interaction. The details of the social-interaction model are beyond the scope of this overview, but adopting the model does have implications for descriptions of writing development and for writing pedagogy (Fitzgerald, 1992). A second trend is to view writing within a broader framework of theories of self-regulated learning (e.g., Zimmerman, 1989). Basically, these models categorize a variety of strategies and feedback mechanisms whereby writers move forward in the composing process. For example, a writer could impose a minimum number of words that must be written before something less taxing (e.g., going to the movies) is done. Or, it might be necessary to rearrange the environment (e.g., move the computer to another room). Zimmerman (1989) stressed three strategies in particular—self-observation, self-judgment, and self-reaction—as critical in writing and learning. Graaham and colleagues (Graham & Harris, 1994, 1999; Graham, Harris, & Troia, 2000) have drawn extensively on self-regulation models in designing intervention approaches for poor writers.

Writing in the Future

It is not uncommon to hear discussion about writing and its uses in the future. We hear and read that the electronic/technological age, with its thirst for speed in the delivery of information, will turn us into information consumers with no patience for writing, which, by nature, takes more time. Perhaps speech recognition technology in computers will impact writing. When Donald Graves was asked about the future of writing (Writing Process: In Retrospect, 1996), he predicted an increase in the volume of writing:

Computers and writing are a perfect match. I see greater opportunities for publishing and getting your work out. Certainly the mechanical aspects are much easier than ever before, so I see the computer as a wonderful friend. (p. 6)

Stephen and Susan Tchudi (Why Write? 1995) share Graves's optimism. They point out that e-mail, for instance, has actually revived interest in print literacy. Whereas letter writing declined when long distance telephoning became affordable, many people who once corresponded by phone are once again writing their messages on e-mail and actually prefer this method to the phone. In an age of information bombardment, perhaps people will write more, not less, because writing provides an opportunity to slow down, reflect, and process information more thoroughly. Technology may actually "increase the need to think critically about reading and writing [italics added] as interpretive, critical acts" (Why Write? 1995, p. 5). The Tchudis also responded to a question about whether writing is really necessary for success in the world of work. They responded that, although there are a few truly successful people who can't write, this is the exception rather than the rule.

In discussions about the future of writing, others have concentrated on the form that writing will take. Kress (2003) argued recently that communication by screen is fundamentally different than communication by page. For hundreds of years, writing in books had little competition for the reader's attention, the occasional illustration notwithstanding. The written word was the entire message. On the screen, however, meaning is shared by writing and image, as well as (increasingly) movement and sound. The result, according to Kress, is that writing is undergoing changes in sentence and text structure, moving away from clausal embeddedness and complex nominal structures toward a simpler sentence structure that more closely resembles the clausal chaining structure of speech. The advent of reliable speech recognition technology may hasten the trend. Just as there is a logic of a written page, there is also a logic (Kress uses the term "grammar") of the visual image—the design of the screen. The screen is divided into blocks according to meaning categories. Whereas the reader of the page can start at the top and keep going in a linear fashion across and down the page to the bottom, a relatively new job for readers of screens is to construct a (nonlinear) reading "path." The organization of the screen, according to Kress, has already affected the organization of the page. A comparison of textbooks written fifty years ago with those of today reveals many changes. Sentences are shorter with fewer clauses. Textbooks have come to resemble the screen—a series of worksheets organized around a topic and "put between more or less solid covers" (Kress, 2003, p. 7).

I thought about Kress's ideas in relation to recent experience visiting a fifth-grade classroom. I listened and watched as students presented their reports on U.S. presidents, all composed on Microsoft Power Point. The students' turn-in written report was their Power Point slides rather than more traditionally composed text. Tufte (2003) writes that the cognitive style of Power Point bullets leave critical relationships unspecified. The question of whether and how screen literacy on the Internet and the use of slideware such as Power Point will impact the development of writing, or its eventual practice, is unknown, but educational linguists such as Kress (2003) argue that the potential for change is strong.

Notes

¹The terms expository, informational, and factual text will be used interchangeably.

²Hayes and Flower (1980, 1987) emphasized the recursive nature of operations in their model-a characteristic that distinguishes the model from previous sequential stage models of writing and underscores the constructivist problem-solving nature of writing (Fitzgerald, 1992). Planning and revision operations may occur at any point in the composing process and may cycle back to earlier portions of the text.