Teachers and administrators looking for ways to use microcomputers in writing instruction will benefit from holding firmly to the principle that the heart of the matter is to teach writing, not word processing.

Keeping the emphasis on writing instruction helps us make the right decisions. When choosing software, for example, we can eliminate the popular program which makes it harder to double space than to merge a form letter with a list of names and addresses; such software is fine for some business applications, but has little to do with the ways writers write. Emphasizing sound writing instruction also can keep us from using microcomputers inappropriately. One published article, for example, advocates putting stock responses to writing on disk and then inserting them into student papers; the author adds that you probably shouldn't let your students know what you're doing. He's right, since they might be angry about canned responses to their writing.

Stock responses don't teach writing, and keeping the emphasis on writing instruction, of course, means understanding what does. The publication of Braddock, Lloyd-Jones and Schoer's Research in Written Composition in 1963 is judged by many writing researchers to mark the beginning of serious scholarship in writing and the teaching

of writing. Before its publication, popular approaches to the teaching of writing held correctness in mechanics and usage to be a dominant concern. Students were often asked to write in imitation of literary models, and they usually studied texts which separated logic from rhetoric. When they wrote, they sat isolated and quiet, produced 500 word themes, and all wrote at the same pace and about the same subjects. Students turned in typed copies of first drafts, and teachers marked them up and passed them back about a week later. The audience for the writing was almost always the teacher, and the purpose of the writing was almost always to examine students. Writing was roughly equivalent to being tested, and collaboration on writing was usually tantamount to cheating.

Since then, research has taught us much about writing and the teaching of writing. We've learned that writing is a process, in the sense of a continuum of overlapping and recursive stages. Studies of writing processes show "stages"--such as generating ideas, shaping and connecting thought and language, revising by adding and deleting and developing, and editing for correctness in usage and mechanics--and the studies also show writers moving back and forth rather constantly between these stages.

Gone are the notions that writers produce typed copies of first drafts and that teachers look at the writing only after it has been turned in for a grade. Instead, we now believe writers produce several drafts (except when the writing is routine and automatic) and that teachers can help during the process of forming these many drafts of the writing. We've learned also that

thinking and writing are inextricably connected during writing processes. Indeed, we now believe that writing is a mode of thinking and learning.

Studies of cognitive processes involved in writing show writers actively involved in constructing meaning. From the first dim stirring of an idea to the publishing of a full and well-formed text, writing involves discovering, organizing, developing thought. Thought is shaped and made communicative by language, and language is learned as it is used to give form to thinking. This is quite the opposite of the "logic and rhetoric" position of teaching them to think, then teaching them to write. Thinking and writing are inseparable.

We've learned that workshop and tutorial methods are more efficient than teacher- or textbook-dominated instruction in writing. Oral response during the writing process, conferring with individual writers and their writing while the writing is produced, is more helpful than theme annotation after the writing is done. Collaboration is not cheating; indeed it is a good synonym for learning. Most of what We learn is learned in cooperation with others, and writing is no exception. The most useful thing teachers can do for writers, regardless of their ages or abilities, is to provide audiences for their writing, audiences who will read and respond in supportive and helpful ways.

We've learned also that correctness does not have to be a major and initial concern in the teaching of writing. Most errors in writing are systematic rather than careless; they show up for good reasons. The writer, for example, who confuses to, too, two is relying on sound rather

than sense as a guide to spelling. Similarly, the writer who confuses the four "add a final s rules" which operate in standard written English (s for subject verb agreement, s for plural formation, s' for possession, and 's for contraction) does so because she or he has an idiosyncratic sense of the rules and because she speaks a nonstandard dialect. In both examples, the transformation of thought into written language has stopped at the level of spoken language. Now, there is nothing wrong with that; we all tend to resort to talk when the writing gets tough. Talk can be a way out of writing difficulties, and at the editing stage of writing we can help by conferring with writers, talking with them about one problem at a time.

The teacher's role in this process-oriented, collaborative approach to the teaching of writing is an active one. In a metanalysis of 72 experimental studies, George Hillocks found that teachers of writing are most effective when they structure writing instruction in three ways: by setting clear and specific objectives, by providing materials and problems to get writers interacting with each other in specifiable processes important to writing, and by providing activities centered on problem solving through peer interaction concerning specific tasks. Hillocks found also that in addition to such structuring of classroom activities, teachers should make specific assignments, help students learn criteria for judging writing, provide exercises in manipulating syntax, and engage students in examining data.

Preliminary research shows that microcomputers can be very valuable in our curricula if we bear in mind four important points:

1. The writing teacher is indispensable as a collaborator and audience, as facilitator and assignment-maker. microcomputers alone cannot teach writers why revision is important or how to bring a first draft to full meaning. Nor can currently available software read and respond to student writing on any satisfactory level. As technology evolves this will continue to be true; computers will never replace teachers.

2. Writers learn best when writing is taught as a process in decentralized classrooms. In doing so, the collaborative method of instruction is most valuable as a primary mode of instruction. Computer-assisted instruction can help, but cannot take over the central roles played by writers and respondents.

3. Microcomputers are counter-productive when used in a theoretical vacuum. We need to use great care when we integrate microcomputers into our classrooms. This means we avoid software which concentrates exclusively upon subskills or isolates them prematurely. We also avoid software which neglects or fragments the processes involved in writing. And very importantly, we avoid software which

teaches mechanical subskills instead of teaching writing.

We reject these unsound microcomputer uses for the same reasons we reject other unsound teaching practices. They don't teach writers how to write. For example, Alderman, Appel, and Murray's study of four thousand community college students using PLAT0, a comprehensive collection of drill-and-practice programs, concluded that PLAT0 had no definite positive effects on learning. Teachers need to be critical of the many drill-and-practice programs now on the market, and we should use them with discrimination, if at all.

Microcomputers do have exciting possibilities as writing tools if they are used well. The problem is separating the many ineffective uses from the good ones, and once teachers know what to look for, this isn't difficult. Good computer-assisted instruction programs are integrative, interactive and individualized. They are easy to operate and user-friendly.

Many writing teachers and researchers have focused on word processing as one of the most valuable uses of microcomputers in the writing classroom. High school teacher Gail Womble observes five important results of word processing:

1. Students often develop into more fluid writers;

2. Revision is more intensive and varied, and is sustained over a longer period of time;

3. Illegible handwriting is no longer an obstacle; writers can focus on the more important business of revision:

4. Since they no longer have to recopy, students are more willing to revise.

Editing programs are readily available to English teachers and our students. In fact, most programs available today are drill-and-practice programs teaching punctuation, usage, spelling or grammar. Unfortunately, these programs don't really teach writers how to write.

More sophisticated editing programs emphasize text analysis instead of drilling writers. For example, programs can gauge the maturity of a writer's word choice; find certain kinds of mistakes; point out omitted material in journalism articles; and generate statistics about the writer's text. The Writer's Workbench, for example, is comprised of software programs providing many kinds of text feedback, but it is not designed primarily for the composition classroom.

Some programs attempt to help writers learn how they make mistakes, taking a holistic view of writing, and these are the most promising. Glynda Hull and William Smith, researchers at the University of Pittsburgh, give writers the responsibility of locating and identifying errors, instead of identifying errors for them. Writers are given text to correct, and the program is able to determine whether the response is accurate. The program also highlights mistakes the student does not see.

Glynda Hull and William Smith are developing

programs which will help students detect errors in their own texts, since students seem to learn to edit best in the context of their own papers. They offer several important guidelines for teachers selecting from among currently available commercial programs:

1. Select programs which teach students to make corrections rather than programs which make automatic corrections.

2. Separate questions of style from questions of style from questions of correctness.

3. Avoid programs which use grammatical terminology to explain why something is wrong. Students, for example, usually don't know what the past perfect is, and telling them doesn't help them correct their mistakes.

Microcomputers are no panacea. In some ways computers make the teaching of writing harder, not easier. Teachers need to learn how to use microcomputers, select software, train students, plan for writers not using microcomputers, and teach writers to respect one another's privacy. And not a single bit of research tells us yet that writing quality improves when word processing is used for instruction purposes. These points are worth some serious thought and further research.

"Microcomputers and Writing" was reprinted by permission of the authors from Faculty of Education Studies Newsletter, Spring 1985, SUNY-Buffalo.

JIM COLLINS and ELIZABETH SOMMERS are co-editors of Writing On-Line: Using Commuters in the Teaching of Writing, published by Boynton/Cook. They teach in SUNY-Buffalo.