This book collects twenty articles by thirty nine authors report­ing on composition research of the last several years. It is the first volume in a projected series by Guilford titled Perspectives in Writing Research. Forthcoming titles include Writing Blocks, Protocol Analysis, and Writing In Nonacademic Situations. The series editors, Linda S. Flower and John R. Hayes, remark in a foreword to the present volume, “In this series, we will attempt to present the best new research on reading and writing, drawn from diverse fields, in a form that is useful to teachers as well as researchers. We hope that the volumes in the series can serve as core readings in upper-class and graduate courses.” The question of the book’s utility seems to me worth addressing. To whom is it likely to prove useful and for what purposes? First, however, a description of the book’s contents and methods is in order.

New Directions in Composition Research¹ is divided into four sections, each presenting five essays. The sections are "Research Methods," "The Composing Process," "The Writing Situation," and “The Instructional Context.” These divisions are not unreasonable, but they are more for convenience than to signal any very clear distinctions between the essays in the various sections. The book might just as easily have been divided along other lines, for example, the age of the writers in the writing situation studied. The truth of the matter is that these articles were written individually, several originally for contexts other than this book, and illuminate each other only by chance. The editors have written introductions to each section which attempt to unite the concerns of the articles under the heading they have chosen, while adumbrating the contents of each essay. They perform the latter function admirably, but such unity as is achieved is entirely the result of the editors’ artifice. The separateness of the essays is not a matter of great concern; but it does mean that the reader is confronted not with a single authorial perspective or with four, but with twenty. That fact makes reading the entire book a rather fragmented experience. It also makes the reviewer’s task a difficult one. To comment on every essay in a reasonable space is not possible. But “representative” selections in such a diverse offering are also not really possible. For example, comments on Kenneth J. Kantor’s article “Classroom Contexts and the Development of Writing Intuitions: An Ethnographic Case Study” will in no way inform the reader of the nature of Marilyn M. Cooper’s essay ‘The Pragmatics of Form: How Do Writers Discover What to Do When?” even though both are in the section on research methods. Generalizations which are accurate to all the essays are likely to be banal, but they are often what the reviewer is reduced to.

Before descending (or ascending) to that level I will try to give some idea of contents, though it must be a condensed once. Seven of the essays report on research undertaken with college students. Four report on research with secondary schools students and four on primary school students. One essay deals with writing done on the job by adults, and the remaining four deal with subjects not specific to any age group: teacher training, protocol analysis, word processors in composition, and the discovery of form. The essays dealing with college writers tend to focus on the development of specific skills. Three have to do with revision and one with anxiety about writing. The other three focus on evaluation, one examining the effect of teachers’ comments, one comparing teachers’ evaluations of college and professional writers performing the same task (The professionals did not fare uniformly better.), and one contrasting students who did well on a writing placement style examination with those who did poorly. The essays on secondary school students are more general in their concerns. Two consider the processes by which students succeed or fail in writing correct, clear, and meaningful sentences. One describes the development of “writing intuitions” and the fourth considers the attitudes of secondary school teachers toward writing instruction. The essays on primary school students are the most gen­eral of all in their focus, aiming at describing and understanding how students begin to learn to write by relating writing to play, to oral language, to reading, and to “social cognitive” ability.

Although these articles do represent new directions in compo­sition research in that they exemplify avenues of exploration fairly recently applied to composition, they are all copiously referenced and all parts of an ongoing research effort in the fields they represent. None of them is new in the sense that the authors have here revealed for the first time an approach to thinking about writing. Further, most of the essays represent an application to composition research of perspectives from other academic fields, especially linguistics, psychology, anthropology, and communication. Two exceptions to this tendency are the pieces on protocol analysis an don the importance of emphasizing process over product in teacher training. Although these are perspectives developed in studying composition, neither of them is likely to be unfamiliar to students of composition in 1984. Thus, while the book’s title is not misleading, it may promise more novelty than the volume actually delivers.

Still, this book does chart a trend in composition research and in doing so it provides the reader with some valuable information. The trend it charts, however, is not so much the set of multiple directions which the title implies as a single direction with multiple expressions. Part of that direction is indicated by the editors of this volume when they write in their introduction,

The editors clearly imply that the model of composition re­search is taken from the social sciences and, if there is any doubt, the articles quickly allay it. The rhetoric of these pieces is nearly uniform and it is the prescribed rhetoric of the scientific paper: statement of the subject, review of literature to establish what has been done and that the present study is a needed addition to the field, description of the study, report and analysis of the data, discussion and interpretation of the results, conclusions and implications for further research. Exceptions to this pattern, where they exist, are not exceptions to the general adoption of scientific rhetoric but variations of it. They simply report on and collate other research, they describe in detail the setting up of a specific kind of study, they propose a model for classifying material, or they describe observations without statistically manipulating them. Overall, if this book can be believed (and I think it can), the direction of composition research is toward the social sciences and away from the humanities. Further, it is in the direction of empiricism and away from a priori theory, toward the collection and statistical manipulation of quantitative data and away from the anecdotal impressions of writing teachers and critics.

To many this is doubtless good news and to some extent whether one takes it as good news or not is merely a matter of personal taste in one’s reading and research. However, there are further implications than the gratification of one set of tastes over another, and even the matter of taste is not an insignificant one.

In the first place, the decision to pursue research in any field according to the scientific, quantitative model implies important assumptions about the nature of the material being studied. Without reciting the now-familiar litany of the inappropriateness of subsum­ing new fields to the scientific model in an age of scientific self-doubt, it is surely appropriate to ask whether the assumptions implied by the decision to study composition by the collection and analysis of nu­merical data are validated by experience or useful. If they are, a reasonable case can be made for applying them. But if they are not, it seems only reasonable to doubt that the new direction in composition research is one that will serve us well or that many will follow.

One assumption of quantitative research is that in order for a concept to be studied it must be quantifiable. Quantification involves an operational definition and an instrument capable of measuring that definition reliably. To take a typical example from this volume, “situational writing anxiety” (the anxiety felt about writing which is specific to a particular situation as opposed to anxiety felt about writing in general) is defined as “transitory in nature and depend[ing] on the particular characteristics of a writing situation” (p. 260). In order to operationalize this definition, the authors had to find a way to vary the situation or certain parts of it and then to measure the differences in anxiety as the situation varies. Their solution was to administer a test in which college students read a description of a writing situation in which one of five anxiety producing factors was varied. The students then answered a questionnaire aimed at meas­uring their anxiety about the prospective writing task. The question­naire was in the form of a semantic differential so that students responded to a series of statements like “I feel panicky about the writing project” or “I feel calm about the writing assignment” by agreeing or disagreeing on a scale of five possible degrees.

A little reflection on this experiment will show that the only source of information about anxiety in this case was the response to the questionnaire. This instrument used some seventeen items each containing one key words as a synonym or antonym to anxiety, for example, “terrified,” “comfortable,” “apprehensive,” “uneasy,” “se­cure,” “self-confident,” or “relaxed.” By varying the words used, the designers of the test aim to elicit responses which reflect their feelings in the general area defined as anxiety. Further, the test assumes that this area is more-or-less the same for everyone. Both of these assumptions are open to question since the matter of synonyms is not clear and simple. I may for example feel “tense” about something but still secure” while another person may find these states incompatible. I may be “nervous” without being “jittery” or “worried” but still “self­confident.” Sensitivity to language and care in considering it are by no means the same for every respondent nor is it possible to find synonyms which simply repeat meaning in different forms. These have been complaints of humanists about the quantitative research of social scientists for many years, and to find them ignored or glossed over by people who are presumably in the profession of being sensitive and careful about language use is disquieting.

Even if we grant these assumptions, research of this type makes further assumptions about its data. Note that in the present example, respondents are not explaining what they feel or felt about an actual situation, they are “imagining” how they “would feel in that situation” (p. 265). They have done no writing nor have they been in the situation described to them. Thus, even if we assume that their responses to certain words accurately reflect feelings in a general way, we must also assume that they can accurately imagine feelings in a situation not experienced. Moreover, we must assume that the descriptions of the situation which are to guide their imagination are realistic enough to permit the accurate imagination we have already assumed. Anyone who has ever taken tests of this type must doubt this assumption on general principle. (When asked, for example, whether one would rather attend a concert, go to a museum, or walk in the woods, can anyone feel that it is possible to imagine accurately the situation?) Lest I seem unfair to these authors, here is an example of the difference in situations as they describe them: "The assignment you are given is a familiar one. You’ve done many of the same type before. The subject of the assignment is something you are familiar with." vs. ‘The assignment is not a familiar one. You’ve never done anything like it before. The subject of the assignment is something you are unfamiliar with” (p. 263). Note that, as with the concert, museum, walk example, the description is at a high level of abstraction. A symphony concert is not a rock concert, nor is an assignment to write a poem an assignment to explain a Supreme Court decision. Do we assume that people will feel the same about both assignments, merely because they are unfamiliar, any more than we assume that people will have the same preferences about the concerts regardless of their content? What respondents are being asked to do is not to imagine any situation, for they haven’t sufficient information to do that. They are being asked to construct a situation given a few generalities, or they are being asked to respond to certain abstract concepts with a predicted emotion. Though one might with some effort grant the assumptions of synonymity, of similar response to words, and of ability to imagine accurately feelings in a hypothetical situations, it seems too much to ask that we should also grant that, being asked to construct those situations from abstractions or respond to mere abstractions, the respondents can give answers which measure or predict feelings in any set of comparable situations.

Perhaps in spending so much time on this example, I may seem to be taking aim at a weak spot; but I can see little difference between this example and the operationalization and quantification of “cohesion” or “acceptable” style or numerous other concepts that form the basis of various essays in the volume. My conclusion about these essays and the new direction of composition research in general is that while the researchers have become immensely accom­plished at manipulating their data, they have failed to ask with any sophistication what their data actually measure or what assumptions must be granted to agree that the data measure what the researchers wish to study. A sign of the interest in the manipulation of data, seemingly for its own sake, is the number of elaborate statistical work­ups performed on tiny samples. The essay on writing in nonacademic situations analyzes in depth two samples, one of five subjects and one of six. The book’s first essay, on the writing abilities of the freshman class at a large university, collected over 400 essays, but publishes its elaborate statistical tables for a sample of only ten of those. Another analyzes a sample of nineteen fourth-graders. I don’t think that one need lay claim to much statistical expertise to question the value of statistics drawn from samples of this size.

Even where statistical methods are not at issue questionable assumptions are made by the writers. In the chapter on protocol analysis, for example, the authors note that verbalization may “change the way one thinks” (p. 55). But after referring to a study which they claim “indicates that the kind of verbalizing which protocol subjects are asked to do does not alter the writer’s focus of attention, although it often slows it down,” they conclude that protocols offer “access to a deep and broad pool of information about the writing process without unduly distorting it” (pp. 55-56). Leaving aside the question of how much distortion is not undue, the assumption that “slowing down” the writer’s focus of attention does not constitute a distortion is a major one. After all, it is partly the ability to consider rapidly a number of possibilities and directions in one’s writing which allows for the flexibility and fluency of accomplished writers. Or so I assume.

There is, throughout the essays in this book, a tendency to reify constructs whether they are used for statistical purposes or not. Thus, “reading-like behaviors,” “commissives,” “lexical cohesion” (to pick three random examples) are all treated as though they were actual things instead of the invented categories of the researchers created for the purposes of a given study. Once again, this sort of naiveté on the part of scholars presumably familiar with the arbitrary nature of linguistic categories is disconcerting.

However, even constructs which have no more than an arbi­trary existence may prove useful. The assumptions of these writers, if they have not been validated by experience or accurately recognized by their creators, might still have some utility. Do these essays tell us things which we can use in the classroom? Do they offer new insights into the minds of writers? Unfortunately, they do no. The majority of their conclusions are of two kinds: Further research in this area is warranted and will possibly unlock the mysteries this study has described; and: Now we have a taxonomy and a method for studying a phenomenon we have long been aware of but never studied system­atically before. The former of these is, of course, the well-known and time-honoured plea for recognition of a study’s value and more research money.

The latter is an attempt to stake a claim as a pioneer in a field. Since these are such standard parts of the scientific rhetoric which these essays have adopted, it would perhaps be most appropriate to overlook them. There are two reasons for not doing so, however. The first of these is that, because this volume is attempting to establish the validity of a new scientific field, it warrants for further research deserve scrutiny. This scrutiny is especially important because much of the research in this book was supported by the United States Department of Education. (Four of the essays acknowledge grants from the DOE). Given the government’s limited support for education, it seems crucial that this support be used in genuinely useful ways. A second reason comes from comparison with the other social sciences which the new discipline of composition research seeks to emulate. How have their promises to unlock the mysteries they describe fared? What has been the utility of taxonomies and empiri­cal approaches to studying long familiar problems? Answers to these questions should influence our willingness to grant these researchers the further credibility and support they request.

Beyond the general conclusions sketched above, the essays seem to come to startlingly obvious conclusions or to verify com­monly acknowledged ideas. For example, one essay suggests that children who are read to and come from homes where books and writing are valued are likely to develop faster in reading and writing. Another points out that most on-the-job writers have rhetorically based reasons for their writing choices. A third explains that as young writers develop they acquire more elaborate intuitions about writing. In other essays we learn that students make more mistakes in long sentences than in short ones because they have difficulty holding the more complex structures of the long sentence in mind to its conclu­sion; that students who are given instruction in self-assessment improve their abilities to do some kinds of revisions; and that ambiguity, novelty, conspicuousness, evaluation, and prior experience influence the degree of anxiety writers feel. It is difficult for me to imagine the teacher of even minimal experience who has not figured most of these notions out for herself; but, if there are such people, they would probably find this book a slow and tedious way to learn them. If the book could claim to have proved them in some definitive way, the demonstration of the proof might be of interest; but, as I have argued above, the nature of the evidence offered in this book demands as much or more faith in social-scientific methods of investigation as the bare notions demand in one’s common sense and powers of observation.

Of course there are those who have little faith in their powers of observation and who doubt the value of anecdote as a means of knowledge. There are also those with a taste for reading the sort of essay which this volume offers. As I mentioned earlier, though, matters of taste are not without significance. In this case, the obvious significance is pedagogical. Do we want students and future teachers taught by people who admire the scientific report as a form of literature? Are the people who contributed to this book and who wish to distinguish their discipline from literary theory reliable critics of student writing? Are the processes of reasoning to which they evidently adhere models for students? Ultimately, of course, the question is whether the teaching of writing is a part of the humanities or not. I do not presume to answer that question for everyone; but, for those who see writing as the essence of the humanities, a taste for the social sciences in language or in reasoning is a matter of importance and concern.

I can best explain this position, as well as fulfill my promise to consider this book’s usefulness, by offering that most unscientific thing, an anecdote. Recently I happened to teach a course titled ‘Teaching Writing” to a group of students including both prospective and experienced teachers. During the course of the term I brought up for discussion several of the essays in this book. In each case, the students were impatient with the material. ‘Where is the conclusion that we can use?,” they wanted to know. “What is the point of all this jargon? What’s the point of all the statistics?” The most telling criticism came from a student who has several years experience teaching high-school English and who also has a master’s degree in computer science. The problem is, he explained, these people want to quantify writing, but writing just isn’t something that lends itself to quantification very well. What works for one writer isn’t what works for another, and you can’t teach by the numbers anyway. What the students wanted, without exception, were techniques and experi­ences that they could apply to their own teaching and a chance to discuss what the ends of composition instruction were. I cannot imagine any of my colleagues finding New Directions in Composition Research the slightest bit useful for those purposes; but if they did I have every confidence their students would soon disabuse them.

What this book will no doubt be very helpful for is learning how to design the sort of study these researchers are engaged in and how to get research support for it. If courses are offered in those subjects, no doubt they will adopt this book. As a resource for teachers of writing, however, the book is destined to be little noted nor long remembered. As an indication of the direction of composition re­search or the establishment of a new discipline, it is probably accurate; but, to humanists and literary critics who also teach writing, to those who believe in their own common sense and experience, to those who seriously doubt that any conclusions about “the actual production of written language” can be arrived at without considering intentions and effects, and to any one concerned with how our limited financial resources for education are to be used, it will be a depressing indication indeed.