00001.tif

PDF

Download

Open document

Flip pages

Copy asset link

Request this asset

Transcript (if available)

Content SYNTAX AND SUCCESS: STYLISTIC FEATURES OF SUPERIOR FRESHMAN ESSAYS by Celest Ann Martin A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (English) September 1979 UMI Number: DP23074 All rights reserved INFORMATION TO ALL U SERS T he quality of this reproduction is d e p en d en t upon the quality of the copy subm itted. In the unlikely event that the author did not sen d a com plete m anuscript and th ere are m issing pag es, th e se will be noted. Also, if m aterial had to be rem oved, a note will indicate the deletion. Dissertation Publishing UMI D P23074 Published by P roQ uest LLC (2014). Copyright in the D issertation held by the Author. Microform Edition © P roQ uest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United S ta te s C ode P roQ uest LLC. 789 E ast E isenhow er Parkw ay P.O . Box 1346 Ann Arbor, Ml 4 8 1 0 6 -1 3 4 6 UNIVERSITY O F S O U T H E R N CA LIFO RN IA THE GRADUATE SC H O O L UN IV ER SITY PARK LOS A N G ELES. C A L IFO R N IA 9 0 0 0 7 This dissertation, written by ufrdrerr"ihe direction of hjl/l/. Dissertation Com mittee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fulfillment of requirements of the degree of D O C T O R OF P H I L O S O P H Y Dean DISSERTATION COMMITTEE ....... s j y-n Chairman f M - j r *i— n m r % 0 M37^ f t: • * . - . - . 1 1 TABLE OF CONTENTS Page LIST OF TABLES. ........................ 1 CHAPTER INTRODUCTION .................... II. REVIEW OF THE LITERATURE ..... ... 23 Descriptive Studies. . . . . . The Loban Study. ....... Experimental Studies ..... Calculating Syntactic Density Scores. ..... Practical Programs ...... ... 29 III. METHODS AND RESULTS.............. Hypothesis .......... The Design . ............. . Test Questions......... Reading.................. Dependent Variable . ........ Independent Variables. .... Results and Interpretations. . ... 7# IV. DISCUSSION, COMPARISONS AND CONCLUSIONS..................... Some Exceptions* ....... Limitations of the Study . . . ... 9 4 - . . . 103 ¥. IMPLICATIONS I PEDAGOGICAL-:,AND THEORETICAL.......... .......... Implications for Teaching: After Sentence-Combining, What Next? ........ Implications for Research. . . BIBLIOGRAPHY. ........................... . . .1 :ri21 APPENDIX: Counting Procedures. ...... iii LIST OF TABLES Table Page 1. Means and Standard Deviations for the Variables................................... 79 2. Percentage of Essays Above Mean in Variables. .............................. 81 3. Chi Square Values for Variables Characteristic of Quality................. 82 4. Free Modifiers: A Comparison with the Hold Study..................... 87 5. Comparison of Present Findings to Previous Studies.. .................... 89 1 INTRODUCTION It's June of 1978; I've been a graduate student and a teaching assistant now for five years# and I am writing a dissertation on Freshman Composition* Presumably, then, I have an interest in writing, should have more than an adequate knowledge of my subject, and should, by now, know what writing is all about. Yet as I sit in this straight chair on a hot summer day, nothing seems so true to me as a statement once made by John Ciardi when he was talking with students about writing: "The major problem in writing is to go into a room, close the door, and put your butt firmly on the chair and leave it there."'*’ If after all these years of practice, I still find this one simple dictum so difficult to follow, how can I, or anyone else for that matter, realistically expect eighteen year olds who have five classes and a social life to keep up to take a sixteen week writing course— and end up liking to write? I wonder sometimes. Yet as one professor puts it, teaching Freshman Composition is, after all, a tremendous responsibility: I have taught English Composition for twenty- five years, almost always two classes totalling fifty students every semester, fall and spring, back to back, year in and year out. And I have taught' from the conviction that teaching freshmen 2 to write is the single most important job an English teacher can do. Although this attitude may be an anomaly in the profession, let me say that I share Professor Lindberg's conviction, and out of it has grown this dissertation. I have taught five years, not twenty-five, and out of those five, two have been spent in the Writing Lab at the University of Southern California. Hot until I began reading my colleagues* comments on student papers did I realize I too had been guilty of what Professor Lindberg describes as "an exercise in slogans" when evaluating themes. He cites such comments as "If you haven't written clearly that means you haven't thought clearly"; and "sloppy diction means sloppy self-discipline." Any teacher can add his or her favorite aphorism here, but the net effect is the same. The student reads the comment, sighs, folds up the paper, and despairs of ever writing well. Small recompense for the amount of effort it takes to get words down on paper, however well or poorly. And English teachers perpetuate this cycle when they continue to use terms like "style," "clarity" or "awkwardness" in the margins while failing to provide their students with some concrete features of these terms they can learn to manipulate in their writing. Of course, we cannot teach effectively what we cannot measure effectively. As long as five English teachers 3 together would come up with at least three different grades for the same paper, they are not likely to agree on abstract terms. Paul Diderich suggests that this phenomenon occurs far more frequently than we would care to acknowledge. Teachers who have never graded a set of papers that have previously been graded by another teacher seldom realize how commonly and seri ously teachers disagree in their judgments of writing ability. The term "style,** for example, is one that causes as much disagreement among teachers as it has among literary critics for centuries. I would like to include some discussion of style here, because the notion of what constitutes style in Freshman Writing has been the impetus for this dissertation, and will be one of the questions for which I hope to provide a quantita tively measurable answer. A concern for specificity and quantitative evidence in the study of style has appeared in countless articles in the last two decades; perhaps two of the most well-known of these sure Michael Riffterre’s "Criteria for Style Analysis" (1959) and Nils Erik Enkvist's "On Defining Style" (1964), both of which emphasize the importance of acknowledging the frequency of certain linguistic items within a given context. In short, an element which had long been ignored as a way of exact measurement began receiving attention. That element was syntax. 4 W. Nelson Francis and Louis Milie are only two of the critics who were dismayed with the lack of atten tion given to syntax. Said Francis in a 1962 article: "explicators and critics . . . are likely to disregard syntax almost wholly . . . This neglect of one of the principal keys to plain sense can be perilous."4 And while Milic echoes Francis' concern in his introduction 5 Stylists on Style, he qualifies the study of style with a distinction that has perhaps not received enough attention among those who strive to give their students a workable concept: "A distinction may need to be made, also, between the best theory of style for teaching composition, and the best for analyzing literary works."® In many composition classes, the notion still prevails that students will somehow "intuit" the con cept of style, if they are only made to analyze enough literary texts. Unfortunately in this analysis, it is frequently unclear to students what they are looking for, and even less clear what they are to do with it should they find it. The ability to talk about an author's style (particularly if that author has been Cliff-noted) is no guarantee that the qualities dis cussed will carry over into the student's writing. This is well-supported by the fact that a number of 5 known literary critics have a most impenetrable style themselves. E. D. Hirsch has recently introduced the notion of style as relative readability in his book The Philosophy of Composition. Readability refers to the easiness with which a reader understands a text, while •'relative** explicitly concedes the fact that easiness must vary with different semantic intentions.7 Joseph Williams, whose work is the basis for a major part of this dissertation, has isolated various structures as characteristic of a style that fits well with this definition. In his article "Defining Complexity," Williams acknowledges that clarity is not all. Most writers strive for, and readers look for, a little grace as well. But even when it is always clear, that plain style finally wears just a bit thin. It lacks the kind of texture that we associate with appropriately mature writing, writing that is not just clear but vigorous, not just direct, but forceful. Both notions of style have some bearing here. When a critic speaks of Hemingway's style, as compared to Joyce's, in a sense he is speaking of the readability of each author's work. It is my hypothesis that syn tactic density has some bearing on a work's readability, to the extent that use of certain syntactic structures by writers may increase or decrease the work's read 6 ability. In "Literature as Sentences," Richard Ohmann established the importance of syntactic density in literary response: Since a reader must register these multiple functions in order to understand the sentence, it is reasonable to suppose that the very process of understanding concentrates his attention on centers of density. Syntactic density, I am suggesting, exercises an « important influence on literary comprehension. If it is a significant factor in literary response and evaluation, then syntactic density should also play an important role in instructors* responses to and grading of themes. In fact, it is precisely how impor tant a role it does play that I want to measure. Chapter Three will provide a detailed outline of the procedures used in this study, but here let me try to establish the basis for these procedures. There is much discussion in the professional journals of the need for agreement among teachers of English as to what constitutes a "good" essay, an "average" essay, or a "poor" essay. Paul Diderich's findings on this problem have already been cited; let me give additional weight to the importance of the issue with a statement made by E. D. Hirsch: One goal at the top of everyone’s list will be to develop reliable methods of evaluating the quality of writing. This is the sine qua non of all future composition research, since without it we have no way of testing anything else.1® 7 I have made the assumption that if syntactic density plays a significant role in literary response, it should play a significant role in instructors' responses to and grading of themes* To test this assumption, fifty Freshman essays were holistically evaluated on a one to four scale. Each essay was given three readings by three different evaluators. Rater reliability and scale criteria will be discussed in Chapter Three. Educational Testing Service has used holistic ratings to evaluate essays written for College Boards for years. In National Assessment and the Teaching of English, John Mellon gives a detailed description of how holistic ratings were applied to the 1969 and 1973 National Assessment data.A good description of the holistic rating process is given by Charles Cooper in Evaluating Writing: Describing. Measuring and Judging: Holistic evaluation of writing is a guided procedure for sorting or ranking written pieces. The rater takes a piece of writing and either (1) matches it with another piece in a graded series of pieces or (2) scores it for the prominence of certain features important to that kind of writing or (3) assigns it a letter or number grade. The placing, scoring or grading occurs quickly, impressionistically, after the rater has practiced the procedure with other raters. The rater does not make corrections or revisions in the paper. Holistic evaluation is usually guided by a holistic scoring guide which describes each feature and identifies high, middle, and low quality for each feature.12 8 In our scoring, we assigned each paper a number grade* the guide we used will be presented in the third chapter. Cooper goes on to say that a scoring reliability of as high as .90 can be obtained for individual writers. There are multiple uses for holistic rating; let me just say here that for pur poses of this dissertation, it was the most reliable way I knew to achieve a score for the over-all quality of the writing in each of the fifty essays. Having obtained these scores, what I then wanted to do was apply the syntactic density theory to them. Did those essays which were characterized by a high level of syntactic density receive a higher rating on the scale than those which were not? Or is the situ ation reversed, with low level density receiving higher ratings? In either case, the kinds of syntactic structures that account for the density will be very important, which is why a later chapter in the disser tation will provide a breakdown of the syntactic structures studied. The importance of individual structures in contri buting to syntactic density was discussed ten years ago by Christensen in his article, "The Problem of Defining a Mature Style." Christensen’s article was based in part on disagreement with the conclusions in Kellogg Hunt’s 1964 study. Specifically, he felt that following 9 Hunt's suggestions (i.e., teaching students to increase clause length) would decrease the readability of the resulting prose: "A mature style must say much in little, agreed, but a mature style must be easy to 13 decode." Christensen found the long noun phrase to be the worst offender in terms of decreasing readability. Using long noun phrases, however much it would increase the syntactic density of a text, would have a detri mental effect on the clarity of the sentences, increasing the difficulty for the reader. The problem now, as then, was summed up by Christensen: Without a clear understanding of the syntactic features of a mature style and some semblance of a consensus among ourselves, we can do nothing to.advance our students' progress toward it.14 Christensen*s statement defines the central problem addressed by this dissertation. A group of teachers abstractly discussing which syntactic structures are important in the teaching of Freshman Composition, or which are indicative of a mature style, would probably sit until doomsday attempting to reach a consensus. On the other hand, with a minimum amount of time spent per session, teachers can be trained to read papers and rate them to achieve a high level of agreement among themselves. Why not use the former situation to enlighten the latter one? 10 How that I have on hand a random sampling of student papers, each read and rated by three trained readers, it should be interesting to see which kinds of syntactic structures consistently appear in those papers receiving high ratings. The method is a bit indirect; however, if we again called upon our hypo thetical group of teachers to read a sample of essays and rate them for maturity of style, my feeling is that the old problem of four or five different ratings for every paper would crop up. But by asking them to rate the papers holistically after a training session where everyone "practices*' rating, they are allowed the delight of giving a quick intuitive rating to the whole paper. It is ray belief that sophistication of style is one of those things like love or happiness: it is never found when directly sought. However, as Pirsig argued in Zen and the Art of Motorcycle Main tenance . quality in writing, as in most other things, is recognized and appreciated when encountered. It is simply very difficult to define. The purpose of statistically identifying the kinds of structures characteristic of high and low rated writing is not simply descriptive. The final concern of all composition research must be pedagogical; in other words, how do we apply what we have learned from our reader/rater responses to individual 11 students wrestling with the most effective way to get their thoughts on paper? I'm not naive enough to think that knowledge of which syntactic structures receive high or low ratings and an ability to guide students in use of these structures are going to resolve all the uncertainty connected with the teach ing of style. But it will be a start. In his book Style and Stylistics. Graham Hough speaks to those who doubt the wisdom of applying statistical quantification to the study of style. Although he is here speaking of literary texts, the same can be said of this method applied to student texts. Something should be said at this point about statistical methods of investigation, now becoming much more prominent with the use of computers. They are probably repugnant to most students of literature. It is felt on one hand that insensitive and inappropriate methods are introduced into literary scholar ship by such means; on the other that liter ary students are unwilling to submit their observations to positive verification. If it is asked whether statistical information is ever relative to style-study, the answer must be a qualified 'yes.'^ Hough, like many others who have written about style, defines style as choice; "choice between the varied lexical and syntactic resources of a particular language." In his article "Rhetorical Choice and Stylistic Option; The Conscious and Unconscious Poles," 12 Milic presents the question “How much conscious control does a writer have over his style?“ Milic unravels the argument traditionally given by students (though not so clearly) when they are told they must "develop style,w that "how" they say it deserves con sideration as well as "what" they say. the writer or speaker intent on the succession of semantic items (meanings) he is producing cannot spare any mind to a conscious examination of the syntactic structures in which those meanings are presented. If he allows himself to be distracted from the semantic stream he at once loses his control, being forced to stop in order to avoid producing substan tive nonsense or ungrammatical structures. It does not seem possible, for example, to decide in media res that a given bit of information will be put into the form of a relative clause or an infinitive phrase without losing track of meaning. How often I have shifted my weight from one foot to the other while attempting to answer that precise objection in the classroom. Yet we do have such things as the Gettysburg Address, with its balance and parallelism; we do have William Faulkner’s tedious but rhythmic page-long sentences, and Hemingway’s frequently parodied short ones. At some point all of these writers had to think about what they wanted to say, as well as the form they would choose to say it in. Milic contends: From the professional to the novice is a very large step but no different in kind. 13 The student of composition like the profes sional writer, does not know what he is doing when he composes. And the teacher of composition seems helpless to tell him.18 I disagree here. What the professional writers might not know is the nomenclature; they perhaps couldn't identify the absolutes, the noun phrases, the verbals. But what professional writers do know intuitively— and students frequently don't— is the sound of their words on paper. How one will not substitute for the other; because a student can recognize an absolute does not mean that he or she will begin using it as skillfully and naturally as a professional writer. But there is a level at which recognition, coupled with frequent manipulation of various syntactic structures, can result in improvement for the novice. It is here that Milic's distinction becomes important: I distinguish between decisions made uncon sciously while the language-generating mechanism is proceeding as stylistic options and decisions made consciously while the mechanism is at rest as rhetorical choices . . . Stylistic options do indeed generally take place below the sentence level, usually in the design of constructions, but rhetori cal choices can be made anywhere because any aspect of the text may be consciously.^ scrutinized at will. (Emphasis mine) To paraphrase Milic, only at the revision level does the writer consciously choose one structure over another. Perhaps this is generally true; it is at the conscious level that we catch an awkward phrase or a 14 loose transition. However, it is not enough to tell students they must revise. The tools of revision, the rhetorical choices Milic speaks of, need sharpening for the student. Simply telling a novice writer to "find a better way to say that" is not effective if he or she is unaware of what better ways axe available. The student's range of rhetorical choices cannot be increased unless he or she is trained to read carefully and observantly for choices used by professional authors. And even then, nothing will be accomplished unless these choices are regularly practiced as part of the writing assignments. The degree of metalanguage involved here is really up to the teacher. There are now several "fluency" textbooks available; some rely 20 heavily on grammatical terms; some use none at all. However, increasing skills at the conscious level is only a start. Many might claim that teaching our students to be good editors and careful revisionists is more than enough of a task in the average Freshman Composition course. I would agree that achieving success in these areas is no small thing, but there is a larger consideration. We should hope to teach our students to become good writers. not just editors. In other words, to be comfortable with the process of getting thoughts and ideas out of the nebulous area of the mind and onto a sheet of paper without interference 15 caused by worry over a lot of rules. In Milic's terms, then, we must somehow stimulate stylistic options, those decisions he claims are made uncon sciously and usually below the sentence level. In a discussion of pedagogical stylistics, Winterowd sums up the situation I am referring to this way: If modern linguistic theory is correct and there is plenty of evidence that it is, on the following point at least— then a high school or college student who is a native speaker enters his composition class with virtually a total knowledge of his language. That is to say, his competence is total even though his per formance may be "substandard.” It would seem, then, that the job of the composition teacher is not to "teach" structure, but to activate competence so that it spills over into the area of performance.21 Much of the composing process is still a mystery, and we cannot say with certainty how good writers got to be good writers, or as Winterowd puts it, how their competence became activated to achieve high performance levels. One thing I have said already is that writers intuitively know the sound of their words on paper. Thanks to biographical criticism, we also know one other thing writers have in common: they read exten sively. What they seem to absorb by osmosis from all this reading is what we must try to give our students: a feeling for the variety of expression that exists in language. They must come to know what Oscar Wilde meant _ - 16 when he said: "A book is neither good nor bad; it is either well or poorly written." Winterowd offers some suggestions on how to accomplish this: It seems to me that the skilled teacher can use modern theories and techniques within the rhetorical situation by working with the sentences that students produce in their own writings—-writings that attempt (we should hope) to convey ideas to an audience for a purpose— and by analyzing the structures that other writers use and relating these structures to the effects that those writers achieve. Thus the business of "teaching" sentence structure will have all of the glorious and productive lack of system that characterizes most real language learning tasks. ^ It is important that teachers not confine themselves strictly to literary examples when analyzing writing for sentence structure and adjustment to audience. Stu dents too easily assume "Oh well, the teacher expects me to be another Hemingway— and he was so good he got a prize." Hemingway, Faulkner, Fitzgerald, Steinbeck, Melville all appear regularly in my examples. But so do Loren Eiseley, David Halberstam, Irving Kolodin, Pauline Kael, and others whose articles and reviews appear in the pages of current magazines. Students must realize that there is a practical application of these techniques they are supposed to absorb, that they do not belong only to the world of fiction writers, that fluency in syntax is essential to all types of writing. More and more I find that students respond to * 17 well or poorly written magazine articles with interest they do not show for the traditional literary giants whom they feel are too far removed from the language problems they themselves are struggling with. They can see readily that magazine writers must work under constraints similar to theirs, i.e., (1) the article must be geared to a specific audience (readers of that particular magazine— let's assume magazines like Harper'siAtlantic Monthly. Saturday Review) (2) the writer must work to keep his readers interested, or the editor won't buy in the first place. He or she cannot use monotonous sentence structure or begin every para graph with "the point X want to make . . ." (3) fre quently the writer's task is to present a new finding in a specialized field (medicine, archaeology, outer space) to lay readers. The student who wants to argue the merits of soccer over football in his essay soon learns that the excuse "Well, you don't know anything about my subject" won't wash; and more importantly, he learns why it won't. This last item works particularly well with magazine articles in a way that it can't always with literature. A student can easily claim that Faulkner or Joyce have not tried to write in the most understandable way, and arguments by the teacher for artistic merit can lose 18 ground in a beginning writing class, where the emphasis is frequently on clear communication. Several articles in both College English and College Composition and Communication have urged introduction of popular writing into the Freshman Composition courses. All of the articles have common goals: to increase the fluency of college students, to activate their compe tence, to help them use what they know about the oral language when transferring it to paper. One article puts the problem this way: They adopt a "college voice" . . . in their writing, stringing together long words and empty standardized phrases. . . Instructors rarely recognize or know how to make use of -3 the knowledge their students already possess. In an issue of College English devoted to the teaching of composition, David Pichaske sums up what has always been the sorry conundrum of the Freshman English class, and provides a solution: nobody— from students to faculty to the New York publishers who support real books with revenues from composition handbooks and readers— nobody considers freshman composition to have much of anything to do with good writing. It is an exercise to be performed with a minimum of pain and a maximum of A's, salary, and profits. In some places, however, writing is enjoyable: Harper1s. Atlantic. Rolling Stone, the New Yorker. Ramparts. . . . Here is a possibility to be considered: would we not be truer to our discipline (writing, not literary criticism) and our students if we aimed more consciously at 19 the general, non-academic, "popular* 1 expository style, and left the other disciplines (including literary criticism) to teach their own styles?24 To sum up then, one a priori assumption of this dissertation is that superior writing is characterized by a high level of syntactic fluency. This seems a fair assumption when one considers the conclusions presented in the vast amount of research already done in this area. What I hope to do is to determine which syntactic structures receive a favorable response from a group of trained readers who are composition teachers. The rationale for this kind of study is one that has often been expressed: we need to provide our students with something more concrete than the "it's-not-elves- exactly" method of teaching style. However, as composition research gradually iso lates the kind of syntactic structures that achieve high ratings among readers, we are left with a pedagogi cal problem: how can we teach these particular struc tures in the overall context of the writing process? Given that pattern practice exercises, though useful at times, may grow stale without a larger context, I have spent the last few pages outlining an alternative method for "teaching" syntactic fluency. I say "teach ing" tentatively, because as Winterowd pointed out, our prime responsibility is to activate the competence stu dents already possess as native speakers of the language. 20 Examining magazine article writers for technique in addition to traditional fiction writers allows the student to realize that fluency is something any writer must achieve, that it applies to all varieties of writing. Given our pecuniary-minded society, stu dents begin to see fluent writing as a marketable and therefore worthwhile skill, since article writers are actually paid for their accomplishments. If analysis of these articles provides syntactic fluency with a living context, then I say use it. Winterowd states We have the means whereby we can help students increase their syntactic fluency. And anyone who thinks about it will realize that a high degree of syntactic fluency is an accomplishment hardly to be overestimated . . . ideas cannot be expressed— adequately at least— unless the writer has the ability to embody those ideas in appropriate structures. Syntactic fluency is the ability to use the syntactic resources of the language in order to express ideas.25 Few people are more syntactically fluent than professional writers. If we can't bring the writers to the classroom, the classroom must go to them— and learn from them. 21 FOOTNOTES Kenneth L. Donelson, “Why English Teachers Must Write." Educational Forum. 42, No. 4 (May 1978), p. 423 . 2 John Lindberg, "Composition: Cult or Competence," North American Review. Winter 1977, pp. 11-21. 3 Paul Diderich, Measuring Growth in English. (Urbana, 111.: NCTE, 1974), p. 5. 4 W. Nelson Francis, "Syntax and Literary Interpre tation" in Essays on the Language of Literature. Chatman and Levin, eds. (Boston: Houghton-Mifflin, 1967), p. 57. ^Louis T. Milic, Stylists on Style. (New York: Charles Scribner's Sons, 1969), p. 13. 6 Louis T. Milic, "Theories of Style and Their Implications for the Teaching of Composition" in Contemporary Essays on Style. Love and Payne, eds. (Oregon: Scott Foresman and Company, 1969), p. 20. 7 E. D. Hirsch, The Philosophy of Composition (Chicago: University of Chicago Press” 1977), p. 9. 8 Joseph Williams, "Defining Complexity." Paper presented at College Composition and Communication Conference, April 1978. g Richard Ohmann, "Literature as Sentences" in Contemporary Rhetoric: A Conceptual Background with Readings. W. Ross Winterowd, ed. (New York: Harcourt, Brace and Jovanovich, 1975), p. 296. 10E. D. Hirsch, p. 172. 11John C. Mellon, National Assessment and the Teaching of English. (Urbana, Illinois: NCTE, 1975). 12 Charles R. Cooper and Lee Odell, Evaluating Writing: Describing. Measuring and Judging (Urbana, Illinois: NCTE, 1977), p. 3. 13 Francis Christensen, "The Problem of Defining a Mature Style," English Journal. 57 (1968), p. 576. 14 Christensen, p. 572. 22 15 Graham Hough, Style and Stylistics. (London: Routledge and Regan Paul, 1969), pp. 53-54. 1 Hough, p. 3. 17 Louis T. Milic, “Rhetorical Choice and Stylistic Option: The Conscious and Unconscious Poles” in Literary Style: A Symposium. Seymour Chatman, ed. (New York: Oxford University Press, 1971), p. 81. 18Milic, p. 84. 19Milic, pp. 85-86. 20 It is possible that both rhetorical choices and stylistic options are acquired in good writers. See Stephen D. Krashen, "On the Acquisition of Planned Discourse: Written English as a Second Dialect" in Handbook 42nd Claremont Reading Conference 1978, M. Douglas, ed. 21 W. Ross Winterowd, ed. Contemporary Rhetoric: A Conceptual Background with Readings. (New York: Harcourt, Brace, Jovanovich, 1975), p. 253. 22 Winterowd, p. 254. 2^ Lee Haring and Ellen Foreman, "Folklore in the Freshman Writing Course" College English. 37, No. 1, (September 1975), pp. 13-21. 24 David R. Pichaske, "Freshman Composition: What is this Shit?" College English. 38, No. 2 (October 1976), pp. 117-24. 25 Winterowd, p. 338. 23 CHAPTER TWO REVIEW OF THE LITERATURE The recent Sentence-Combining Conference in Oxford, Ohio represented the first of its kind. It was a scholarly gathering: Kellogg Hunt, John Mellon, James Kinneavy and William Strong were among the out standing speakers. The conference provided an excellent overview of both the descriptive work in language development assessment and the experimental studies measuring syntactic fluency over the past fifteen years. The spectrum of ideas presented for scrutiny and experi ment was vast indeed. This chapter is an attempt to present the research discussed at the conference, along with other pertinent studies, particularly those dealing with college-level student writing. The sections include descriptive studies, experimental studies, density scores and a review of two practical programs. Descriptive Studies Kellogg Hunt gave us the T-unit, defined as "the minimal terminal unit" or as a group of words that cannot be further divided without loss of their essen tial meaning.1 In his 1964 study, Differences in Grammatical Structures Written at Three Grade Levels, 24 The Structures to be Analyzed by Transformational Methods, Hunt established that if sentence length is assumed to be an index of language maturity, then the child who under-punctuates the most or uses and the most will, regrettably, be credited with the greatest language maturity.2 And so was born his more reliable instrument of measure ment, the T-unit. Hunt's twelfth-graders achieved a T-unit length of 14.4 words, compared with 8.6 in grade four, and 11.5 in grade eight. In a later study by Hunt, clause length proved to be the second most valuable measure of maturity, and ratio of subordinate clauses to main clauses or to the 3 total number of clauses the third. Recent literature indicates that clause length is the most effective 4 measure beyond the twelfth grade level. Nearly all research in sentence-combining since Hunt has applied his measures. Hunt posited that "the ability to use noun clauses in positions other than as direct object appears to be :*5 a mark of maturity." It was this emphasis on the long noun phrase, a measure reinforced in Mellon's 1969 study, that Christensen took issue with in his article "The Problem of Defining a Mature Style." Hunt did caution against the acceptance of sentence-combining as a cure-all, however, and Christensen may have over looked this in his criticism: 25 Whether the sentence-combining process should be taught at all is not self-evident. If skill in that process is the most impor tant factor in sentence maturity, then a teacher is certainly tempted to try to force the growth. But forced growth is not always firm growth.6 Hunt's statement still holds; even now there is much debate as to the most effective way to teach sentence-combining, so that it will indeed be "firm growth" and not just a temporary spurt. The Loban Study Walter Loban's 1976 language development study was the first longitudinal study to span the thirteen years from kindergarten through Grade Twelve. As such, it provided more precise data on the average age for development of various syntactic skills than had ever been available. The researchers used three groups, each comprising 35 subjects, on whom they had data spanning thirteen years: High Language Proficiency Group, Low Language Proficiency Group, and a Random Group, selected from all 211 subjects involved. Ratings of at least seven elementary teachers plus six or more secondary English teachers were averaged to determine the High and Low 7 Groups. The attributes examined were: average number of words per communication unit (same as T-unit), syntactical elaboration of subject and predicate, number of grammatical transformations (especially 26 multiple-base deletion transformations), reading ability, writing ability, range of vocabulary, scores on listening tests, use of connectors (unless, although, etc.), use of tentativeness (supposition, hypotheses, conditional statements, number of dependent clauses) use of adjectival clauses. The results of such an extensive inventory do not all pertain here; I will not be concerned with oral language development or reading ability. My summary will only include those factors pertaining to written language. Loban notes that his random group is always slightly below the scores presented in the research of Hunt and O'Donnell; their subjects' scores more 8 closely paralleled his High Group. One significant finding was that all three groups showed rapid growth in writing from grades nine to ten, but only the High and Random groups showed another velocity surge from eleven to twelve. This finding clearly relates to the problems of the college student in the Basic Writing Course. Loban hypothesizes that the Random and High groups are anticipating a college education, and so are motivated to beep improving their writing. But with so many colleges developing Special Admissions Programs to give students a chance who have low SAT scores and low cumulative averages, students are entering college who may be as much as 27 five years below their peers in the High Group in 9 writing development. The Year-to—Year velocity for the Low Group in Grade 12 is +1, for the Random Group +19, and for the High Group, +17. Thus colleges or universities offering Special Admissions Programs, but not yet equipped with writing labs, or with teachers experienced in teaching Basic Writing, are putting the students enrolled in these programs in a Catch-22 situ ation. The students do poorly in many of their classes because of their inability to write an essay exam demonstrating their knowledge of the subject. It is too much for them; they know they "can't write,” always had trouble with "grammar"; they want to drop out. But they also know they may never have another chance. Here are students whose writing ability has had no active spurts since 10th grade, and even then when it did increase, it increased only to the level many of their peers had achieved in eighth grade. Loban*s study is descriptive, not experimental, but though it doesn't offer solutions, it certainly presents challenges for the writing teacher. Loban found that the Low Group used subordinate clauses excessively in high school, to the exclusion of the more sophisticated strategies (cumulative sentences) used by their peers. 28 Loban noted also that Hunt's research showed the adjective—dependent clause increasing steadily from the earliest grades to the latest, with it still more frequent among shilled adult writers than among students finishing high school. In his own research, he found it interesting that the High ©roup excelled in written adjectival dependent clauses until grade ten, when the Low Group "catches up” with the High Group in terms of structures used, while the High Group begins using adjectival participial phrases and other more sophisti cated structures."^ Loban sums up this phenomenon later when he describes the elaboration index— the use of all strate gies by which communication units are expanded from simple one-word subjects and predicates. The elaboration index indicates that the High Group shows an obvious superiority when ALL elaborated structures are considered whereas such a superiority does not exist (on written language during the high school years) if one ignores other elaborated structures and cony, centrates exclusively on dependent clauses. Loban*s conclusion here goes far toward solving "the problem of defining a mature style." It suggests that researchers must be very careful about the kinds of structures they count, before claiming that what they have counted demonstrates syntactic maturity. As Christensen pointed out, what is reliable as a measure of maturity on the elementary school level is not 29 necessarily reliable in measuring sophistication in writing of skilled adults. In a descriptive study on the college level, Hurray Stewart reported the results of his experiment with high school and college students, including a group of graduate students. The study used Hunt's "aluminum passage" as the instrument for evaluating syntactic maturity; Stewart simply applied a measure and compared the results obtained at the various levels. The measures used were Hunt's words per clause, clauses per T-unit and words per T-unit. It is significant because it is one of the first studies to evaluate the writing of university students in terras of syntactic maturity. Stewart found that in the measures employed, students in the first three years of uni versity do not display significant gains over the averages for the last two years of high school. Fifth year and graduate students, however, do display some gain over the regular high school and undergraduate levels.12 Stewart's findings would indicate, then, that there is much we can try to activate in those first three years during which he found no significant improvement. Recent experimental studies have tried to do that. Experimental Studies In 1962-63, Bateman and Zidonis hypothesized that the study of transformational grammar would affect the structure of the sentences students wrote, and performed 30 13 a pioneer study to test that hypothesis. Hindsight being what it is, Mellon and subsequent researchers leveled many criticisms at this initial experiment. Nevertheless, the results published in 1964, the same year as Hunt's study appeared, were still significant. To quote Frank O'Hare: The significance of this study lies in the discovery that students who study transfor mational grammar end up writing sentences that have fewer errors and are more complex syntactically than students who do not. That is significant indeed. Some of Mellon* s.criticisms were valid, some a bit strong, considering that nearly ten years later, he still rankles under OVHare's criticisms and improve- 15 ments of his own study. However, Mellon's comments will serve both to sum up the Bateman-Zidonis study, and to introduce discussion of his own. Briefly, they are; 1) Bateman and Zidonis reported that the increase in average structural complexity scores for well-formed sentences was 3.79 for the control class and 9.32 for the experimental, an increase of over five transforma tions per sentence. But the greatest changes in the experimental group were made by only four students, which led Mellon to question whether analysis of vari ance was the appropriate statistic here. Four students equalled approximately one-fifth of the experimental 31 population; Mellon's question here was valid, although it cannot be overlooked that the experimental students i 6 as a whole did write sentences of greater complexity. 2) Mellon criticized the investigators for not using the findings of Kellogg Hunt, overlooking the fact that the results of Hunt's study did not appear until 1964, the same year that Bateman and Zidonis published the results of their two-year study. 3) Mellon suggested that the researchers did not adequately describe what their control group studied; he was perhaps correct that the phrase "regular curric ulum" was insufficient, but one can assume that Bateman and Zidonis would have noted any outstanding deviation from the usual ninth grade English course. 4) Mellon's final criticism carries the most weight. He argued that the rationale of the experiment was a difficult one to accept, and disputed the Bateman-Zidonis claim that "pupils must be taught a system that accounts for well-formed sentences before they can be expected to produce more of such sentences themselves." Here Mellon may be right; this claim does run counter to some of the evidence acquired through research of the competence/performance theory. Theo retically, by the time a child reaches kindergarten, he should have already internalized most of the rules for producing well-spoken sentences. But there may be 32 three types of students: those who have acquired the code, those who have not, and those who have acquired a "passive*’ knowledge. The trick is to get the stu dent to apply this tacit knowledge to writing, and to acquire or learn the communicative strategies that 18 differentiate writing from speaking. Mellon's 1967 study de-emphasized the possible corrective function of sentence-combining, the error- reduction effect noted in several previous studies, and concentrated instead on an increase in syntactic complexity. His main point of departure from the Bateman-Zidonis study was his suggestion that it was the sentence-combining practice itself, and not the study of grammar that affected the students' writing positively. At the sentence-combining conference referred to earlier in the chapter, that claim was carried even one step further. Sentence-combining skeptics suggested that it is the increased writing practice, rather than sentence- combining, that is responsible for students' improvement. This claim was raised in connection with the recent Daiker, Kerek, and Morenberg study. Perhaps there can never be any real certainty as to what exerts the most influence in the mysterious composing process, but certainly directions and trends can be definitely estab lished. The researchers cited in this chapter have all 33 helped to do that, and Mellon was certainly an innovator in the field. Mellon*s study provided for three separate treat ments, one more than Bateman-Zidonis. The five control classes studied traditional grammar; two placebo classes studied no grammar at all but had extra lessons in literature and composition, with no additional writing assignments. Finally, the five experimental classes studied a year-long course in transformational grammar, including a large amount of sentence-combining practice. All of the classes studied the regular English program for their particular schools, all located in a white, middle-class Boston area. (For precise details of counting, see Mellon’s study). After pre-and-post-testing, results indicated that the experimental group had increased in some factors of syntactic fluency (T-unit length, subordination-co-ordin- ation ratio, number of nominal and "relative** clauses and phrases, including adverbial clauses of time, place, and manner, clustered modification and depth of embed ding). The gains were significant beyond the .01 level of confidence. Comparing his increases with those established by Hunt, Mellon found that his experimental group exceeded Hunt’s norms of per year growth by an increase of 2.1 to 3.5 years. His control group, on 34 the other hand, failed to show even one year's growth. Certainly, then, Mellon's hypothesis was substantiated. However, 0*Hare's 1971 study poses some interesting questions about Mellon's work, and even suggests that by the time Mellon*s work appeared in an NCTE report in 1969, the Epilogue to that publication indicated Mellon may have changed his mind about the implications of his . , 19 study. O'Hare's first question is "what constitutes trans formational sentence-combining practices?" Specifically, O'Hare's concern "is whether the word 'transformational' refers to the Researcher * s knowledge of transformational grammar, or to the student's." O'Hare points out that Mellon taught generative grammatical concepts all the way through, and that he demanded three things of his experimental students: 1) learn transformational rules like T:rel, Tsgerund, T-del-NP, T:infin, which they applied in combining practice. 2) learn concepts like passive infinitive phrase, appositive noun phrase, participial compound etc., which they were never asked to apply consciously in the combining practice and 20 3) learn a quite difficult set of grammatical rules. O'Hare concluded that many of Mellon's rules (to use the examples he cites: T:Der-NP=NP + AXJX + 35 VERB + NP NP + VERB + URE + OF + HP) ANCE ME NT TION AL were too difficult for seventh graders; he was therefore left with the problem that formed the hypothesis for his own research. Was it the study of this particular transfor mational grammar that led to the syntactic gains made by Mellon's experimental group? Or was it the combining practice only that led to these increases? Or was it the inter action of the grammar study and the combining practice?^ O'Hare did away with grammatical terminology in his experiment in an attempt to answer the question raised in his mind by Mellon's work: what effect did the grammar study have on Mellon's experimental group? For his purposes, O’Hare defined maturity of sentence structure “in a statistical sense as the range of the sentence types found in samples of the students' 22 writing." Two of Mellonls claims were particularly questioned by O'Hare. The first is that the combining practice was an integral part of the grammar study; the second that the combining practice must be a-rhetorical. O'Hare's curriculum also involved an a-rhetorical setting, but he was convinced that this was not the best way to present it. He considered that it held a real attractiveness when considered as an integral part 36 of composition instruction because of its direct bearing on style and the revision process. Current researchers are further examining its potential in these directions. The basic substance of O'Hare's study consisted in eliminating Mellon*s grammar and substituting a series of practical "little helps." The simplicity of the instruction would allow the student to deal, unhampered, with concepts which language development research has demon strated he has already mastered . . . With grammar done, the full potential of Mellon's kernel—embedding system might be realized.2 O'Hare's sentences, which were in fact revisions of Mellon's sentence-combining problems, looked like this: A. SOMETHING is not easy. Mrs. Adams condoned SOMETHING (IT—FOR—TO) Her son was sent to Vietnam. ('S + ING) B. It is not easy for Mrs. Adams to condone her son's being sent to Vietnam. If one refers to the Mellon example presented earlier in the chapter (pg. 7), one can see that O'Hare has eliminated the necessity for a student's knowing rules like T:rel, T:del, etc. The problems gradually build up, each section of the text "teaching" the embed ding of a larger number of propositions, using more and more language variety. I will not attempt to summarize O'Hare's methods here, but rather move on to briefly describe his results 37 and conclusions. The experimental group experienced growth significant at the .001 level on the same six factors of syntactic maturity Mellon measured, out distancing the control group again at the .001 level of significance. On five of the six factors of syn tactic maturity (all but noun clauses per 100 T-^units), O'Hare's eighth graders scored as well as Hunt's twelfth graders. Most important, the experimental group's composi tions were judged to be significantly better at the .001 level in overall quality than those written by the control group, and there was substantial agreement among the eight teachers who judged the overall quality of the compositions. O'Hare's conclusions led him to reject Mellon's assertion that "the sentence-combining practice had 24 nothing to do with the teaching of writing." The basis of this rejection was his realization that sentence-combining involved semantic considerations as well as syntactic ones, considerations such as "How does it sound? Does it make sense? Does it include all the input information?" This is precisely the point that current research ers like William Strong, and Daiker, Kerek and Morenberg made at the Sentence-Combining Conference. Their views supported O'Hare's statement that "Rhetoric 38 and sentence-combining should be viewed not as mutually 26 exclusive or even discrete but rather as complementary." O'Hare also substantiates Christensen’s claim that form can in some sense generate content, even if the fact that the experimental students had more to say resulted from their increased confidence in manipulation of syntax, rather than any cognitive change in their development. James Hey dampened the hopes of composition teachers when he suggested that "freshman students may simply be past the stage when sentence—combining exercises are of 27 benefit." But in an article appearing in College Com- 28 position and Communication. Daiker, Kerek, and Morenberg strongly disagreed with Hey’s conclusions, and more recently they have provided a good reason why. In October of this year, they presented their very signifi cant results. These researchers used 290 students from Miami Uni versity’s Freshman English course— six sections were control, six experimental. Eight compositions were written during the semester, the first and last serving as pre- and post-tests respectively. The four variables of teacher, topic, time and mode of discourse were carefully controlled by the researchers in accordance with the guidelines laid down by Braddock, Lloyd-Jones, O Q and Schoer (1963). 39 The six control sections followed traditional (The researchers are very specific about the term tradi tional, even listing the texts used.) methods of instruc tion, while the experimental sections dispensed with essay anthologies and rhetoric handbooks in order to make SC activities the exclusive content of the course. The single text used by the experimental sections was Strong's Sentence-Combining: A. Composing Book. To supplement Strong's book, additional model exercises on appositives, infinitives, complex prepositional phrases and parallel construc tions were created. Some of Strong's exercises were adopted in order to emphasize such writing elements as organization and tone.-*0 The students' essays were measured for the three factors of words per clause, words per T-unit, and clauses per T-unit. In addition, the papers were rated for writing quality by three different methods: holistic, analytic, and forced choice. Because holistic rating is the closest to actual paper grading, it indicates how well the writer is able to produce a mature skill fully written composition. This measure was considered, therefore, the most critical of the rating methods, the principal basis for accepting or rejecting the second hypothesis (that a group trained in SC would score sig nificantly higher than a control group on overall writing quality as judged by a panel of experienced college teachers of college composition). Like the syntactic pre-test, the holistic pre-test showed no significant difference between the two groups. On the other hand the post-test difference of -36g^ between the two groups proved significant. 40 Significant indeed. The results achieved were significant beyond the .001 level of confidence, some thing every statistician dreams of. Results were also significant for the analytic and forced choice ratings. Not only does the Daiker, Kerek, and Horenberg experiment disprove Ney’s theory, it goes far toward removing the stigma attached to sentence-combining after the Mellon study in 1969, where gains in syntactic maturity occurred, but without a corresponding improve ment in overall writing quality when the papers were rated holistically. In a Canadian experiment, Murray Stewart also used the Strong book, coupled with exercises based on the Christensen model (1967, 1976). He too achieved results significant at the .001 and the .002 levels of confi dence. Stewart sums up the interpretation of his results this way: The examination of writing quality differences indicates that these freshmen could concentrate on sentence-combining and related writing over a six week period and raise their technical level of syntactic maturity while still main taining, and even raising, the average level of the quality of their writing as judged by trained and experienced raters. In sum, the results support the proposition that freshmen can take a six-week module of grammar-free sentence-combining on the Strong-Christensen model with a definite expectation of positive effects in the area of syntactic maturity and a good likelihood of some improvement in writing quality as w e l l . 2 41 Finally, Maimon and Nodine have done an interesting study of the kinds of errors students may make in sentence-combining when their reach exceeds their grasp. The researchers hypothesized that errors of embedding would be related to the improvement measured in syntactic maturity by words per T-unit. Their desig nated errors were the following: vague pronoun refer ence, faulty parallelism, comma splice/run—on sentence, faulty subject-verb agreement, dangling verbal, rais- 33 placed modifier, and fragment, as defined in Corbett. Though Joseph Williams quarreled with the kinds of errors chosen as being insignificant, it is my opinion, and that of several of my colleagues at Illinois State University, that these kinds of errors, compounded, produce problem prose. There were only 14 subjects involved in the Beaver College study; however, work with these fourteen extended over two semesters. Its importance consists in the fact that the researchers examined differences in skill possibly attributable to differences in modes of discourse, an area that needs much more probing. Five writing assignments were examined for all subjects. Of these, two were Hunt's aluminum passage, which served as both pre- and post-test. The other pre-post test measured was the "lunch essay" which required stu dents to use their imagination. They were given a list 42 of people and told "You have a chance to lunch with one of the following people. Whom do you choose and why?" The other assignment was an in-class writing requiring the students to answer a question of their choice on Hamlet. which they had read in class. The students worked on sentence-combining exercises designed by O’Hare (1975) and Strong (1973) for two weeks at the end of the first semester, and for seven weeks at the beginning of the second. Later in the second semester, they were introduced to the cumulative sentence as explained by Christensen, and for two weeks were given pattern practice in completing a base clause with modifiers. On both the aluminum and the lunch post-tests, students increased significantly in their number of words per T-unit. However, the absolute values of the scores differed on the two assignments. The students wrote longer T-units— about 40% longer— on the lunch essay than they did on the controlled aluminum exercise. The lunch essay was written on a topic familiar to the writers, freeing them, as part of the composing pro cess to add modifiers to the sentence elements, an opportunity not available when manipulating the sentences about processing aluminum. But Maimon and Nodine judged words per T-unit inaccurate as the sole measure of syntactic maturity, and so examined the kinds of errors mentioned above. Interestingly, the mean number of errors decreased sig 43 nificantly in the lunch essay between pre- and post test, but increased in the aluminum passage. The Hamlet essay, viewed as a college-level assignment with a higher level of conceptual complexity than the other two, yielded 20.01 words per T-unit, 2.97 words shorter than the lunch essay. Also there were more embedding errors (1.02 for Hamlet compared to .42 in the lunch essay). Both of these measures indicate that the greater difficulty of the assign ment was not reflected in longer T—units.35 Certainly the study is an interesting one, for it begins to ask the kinds of questions for which we must provide answers. As hinted at earlier in the chapter, increasing words per T-unit must not become the sine qua non of sentence-combining. Rather, attention must be paid to developing the student's eye and ear for the kind of detail appropriate to the mode he or she is writing in, be it narrative, descriptive, analytic, or persuasive. Much remains to be done before we can activate our own competence, and that of our students, to its highest level of performance. Calculating Syntactic Density Scores Loban's elaboration index, and the conclusions arising from it, perhaps the most important contribution of his study, have created a need for calculating a syntactic density score. Many studies have attempted to do this. The concern here is with moving beyond the 44 T-unit, so I will examine only one study still focusing on T-unit length as a primary measure, then move on to others which suggest that the T-unit may not be the most effective measure at the college level and beyond. Among the most interesting of these studies, and one which has received a great deal of attention because of its potential for use with the computer, is the Golub Syntactic Density Score. I will summarize it in some detail. In 1974, Golub and Kidder published their Syntactic Density Score in Elementary English. Their aim was to develop a syntactic measure which points to speci fic teachable linguistic structures which are empirically likely to make a difference in the syntactic density of what a growing child or student reads or writes.36 To accomplish this, Golub first examined which syntactic structures would predict most accurately whether the students* written discourse would be rated as high, medium or low by their teachers. The researchers found twelve variables correlating significantly with teacher judgments of writing samples. If occurrence of these variables was high, then generally the papers would receive a high score. The variables were analyzed and weighted according to their contribution to "syntactic density," the name given to the combination of these variables. 45 Golub's finding closely paralleled Hunts; the loading of their SDS reflects the measures of T-unit length and subordinate clause length as the main indi cators of maturity, with words per T-unit equalling a .95 loading, subordinate clauses per T-unit a .90 load ing. However, it must be remembered that once again, Golub was working with elementary school children, so it would not be accurate to apply his measures to college students with the same loadings. By college, as Loban points out, subordination is no longer an indi cator of maturity; those students who are rated "high" in written discourse are moving on to more sophisticated structures. Golub's work was interesting, though, as an attempt to see how much of a role development of the various structures played in an overall rating of the students' work. In addition to words per T-unit and subordinate clauses per T-unit, Golub's SDS included main clause length, number of raodals, number of be and have forms in auxiliary, number of prepositional phrases, number of possessive nouns and pronouns, number of adverbs of time, and number of gerunds, participles, and absolute phrases (unbound modifiers). The loadings for number of modals and number of be an<* have forms in the auxiliary were given values of 46 .65 and .40 respectively. This contrasts with the higher loading, .85, given the free modifiers. In their 1977 article, Hold and Freedman found final free modifiers to have a high positive correlation to the quality of a paper; (they measured 17 variables) while a high percentage of be*s and have1s in the auxiliary significantly weakened the essay. Also, short essays contained a greater percentage of finite verbs tempered by modals (should, could, would). when sentences containing modals are not properly supported by others containing examples and justifications of the judgments and hypotheses expressed by the modals, the essay is weakened. 7 So Golub's lower weighting of modals and be and have forms is borne out significantly on the college level. In his article "A Critique of Some Indices of Syntactic Maturity," Roy O'Donnell suggests some weak nesses in the Golub model and others previously 38 developed. His first criticism is that Golub's items have a high degree of redundancy in what they measure; secondly, he questions the decisions made in assigning loadings to items. O'Donnell's real concern in this area is the provision Golub makes for converting the SDS to a grade level without taking into account "the diversity of individual writing styles, the diverse 47 demands imposed on language by different modes and 39 other factors yet unknown.” Belanger points out a mathematical problem presented by Golub's formula: However, a problem more immediate than these is a mathematical scoring anomaly in the scale whereby the SDS received by a writing sample depends more on the number of T-units analyzed than on any factor of syntactic density. Thus if a researcher analyzes five T-units of a given passage he will arrive at a far higher score than if he had analyzed ten T-units, which in turn is higher than if he had chosen fifteen T-units. Belanger suggests some modifications (see his tables), but remains concerned that even with the modifications, 60% of the SDS is still determined by only two factors: T-unit length and the number of prepositional phrases. For this reason, he questions whether the SDS results in a valid measure of syntactic maturity. Golub's score is useful in that a self-instructional packet for learning to tabulate his SDS is available, and the tabulation has been programmed for the computer by Carole Kidder. Certainly for elementary students, and even for students of basic writing at the high school level, Golub's formula may be applicable, with some modifications. The concern with "what to count" still remains. Present research questions T-unit length as the primary indicator of syntactic maturity, particularly on the 48 college level. The Freedraan-Nold article previously referred to concluded: Though we have found some significant pre dictors of quality, we do not claim to have identified all the important variables which will distinguish between higher quality and lower quality essays. We do know that words per T-unit and other standard developmental measures are not useful in predicting per ceptions of quality on the college level.41 The researchers go on to suggest that there are many areas for further research and that what constitutes a good essay varies from mode to mode. Differences from mode to mode are, of course, difficult to measure. But Joseph Williams may have provided researchers with some quantifiable measures that are "cross-modal," because they are based not on the writer's conception of what is complex, but on the reader's. The two may be inversely proportionate. Williams puts the problems this way: So if we try to measure a complex style by the number of dependent clauses in a sentence, then we have the paradoxical situation where the more complex the style is, the more easily we process it; and the simpler the style, the less easily we process it. (There is a point where this is not quite true, but in most cases it is.)4^ Williams uses two sentences to illustrate his point; the first is a series of phrases constituting a compound sentence: 49 The government's investigation into the ship ment of the wheat by the exporter was met by his refusal in regard to an examination of his method of payments for its domestic transportation. The government investigated how the wheat was shipped by the exporter, but he refused to let the government examine how he paid to have the wheat transported domestically. Part of the problem in the first sentence is our old friend, the long noun phrase. But as Williams points out, writers frequently have difficulty expressing "every idea in its full clausal form rather than its more abstract phrasal form." Thus Williams finds T-unit length not very useful for measuring simple and complex styles because the numbers that are accumulated "do not tell us whether a writer has written clearly something that his readers can understand easily." It is clear that as writers mature, they write longer T-units, particularly longer clauses. But however these numbers change, we have no direct way to relate any of them to any thresholds we might have in the way we respond to sentences composed of more or less complex T-units . . . T-units provide no straight-forward way to understand how we respond to these numbers (longer T-units), much less how difficult it is for a writer to achieve them. Williams briefly reviews the Flesch formula for readable writing, but concludes that it provides no way to teach a clear style. Before coming to his own hypothesis, Williams covers several measures of complex 50 style used in the past: number of transforms used from deep to surface structure, left and right branch ing, passive sentences according to whether or not they may be reversed, grammatical-semantic expectancy (whether or not the agent immediately precedes the verb), and of course heavy use of nominalization. He and a colleague, Rosemary Hake of Chicago State Univer sity, experimented with subject-verb displacement. They created papers in which we systematically displaced agents from the subject position and what those agents do from the verb position . . . We gave these passages on different days to about 75 typists of different levels of skill and background. They typed the ver sions in which the semantic agents and grammatical subjects coincided about 15% faster than the versions in which they did not and made about 25% fewer errors.4^ Williams gives an example of the kinds of sentences used: A, The scientists analyzed what caused the genes to mutate. B. The causes of the mutation of the genes received analysis from the scientists. Of course, it is in the "B" form of the above that the higher number of errors occurred. This is not sur prising, considering the examples given earlier in the chapter by Christensen and Winterowd, demonstrating the effects of wide separation of subject from verb. Draw- 51 ing on what we know about the role of redundancy in reading theory, Williams states: In short the clearest style is one in which the grammatical structures of a sentence most redundantly support the perceived semantic structure. The more consistently the grammatical structure re-enforces— or reflects— the semantic structure, the more easily a reader takes up that semantic structure.45 Williams acknowledges that it is not always possi ble to write in this style, but suggests that writers can come far closer to this kind of maximally redundant pattern than most of them do. The coup of Williams' paper is that he writes each section in the style he is illustrating; but as he notes throughout the paper, it is extremely complex to write in a style that is "easy" to read. Williams claims that the way semantic and grammatical structures do or do not coincide would appear to be a very salient metric of a complex style, if we agree to define complex in a special way— as the style in which grammatical structures systematically fail to support semantic structures and thereby make sentences rela tively more difficult to read. The features that create this kind of style are very frequent in modern prose; they are crucial to the way we process sentences; they are structures that we can teach our students to avoid, and we can demonstrate what happens when they do not. So as a metric of complex and simple styles, it meets every one of our criteria.45 52 Williams practices what he preaches, not only in his paper, but in many areas of technical writing, in all cases working with adults in "the real world.'* He has worked for the Department of Agriculture, for a large corporation in the Chicago area, for a major railroad system, and has taught physicians from a large medical school how to write for medical journals. In all cases he met with success because his audience was able to perceive immediately that he was speaking to problems adult writers must deal with. Unfortunately, I was not able to apply that parti cular metric in my study of Freshman college students, although I hope to do so in a future study. Certainly it makes sense, and certainly it bears more directly on the problem students will face when they attempt to use their writing skills in the real world. They will seldom be asked to write in the narrative and descrip tive modes, or "to pay attention to a particular point of view," but will be asked to communicate ideas— a business report, a grant proposal, an itinerary summary— in a way easily understood by their readers, who may have upteen other reports to get through in the same day. I have devoted so much space to Williams because I feel his coordination of readability with "write- ability" is moving in the direction we must go, not only 53 in teaching, but as he suggests, in measuring what we have taught. But now I want to move on to those criteria of Williams' that I have chosen to examine in this study. Williams has tentatively coined the term "signifi cant style," consisting in a measure of how clauses end. Williams claims: Clauses written in an emphatic and confident style end about two times out of every three with one of the following three structures, or with some combination of them. (And that sentence illustrates all three). 1) A coordinate structure either within the last phrase or as the last phrase 2) A nominalization (a noun derived from a verb or an adjective, including gerunds) 3) A prepositional phrase introduced by .•of . «47 And then he adds what he calls a very big generalization about one kind of style: "Sentences that are both clear and strong are predominantly agent-action sentences, sentences that move briskly through relatively short subjects and verbs into a comment that climaxes with one of those three structures." (Narrative sentences i 48 do not reflect these distributions). Williams included an appendix giving the clause endings for an article on lasers appearing in Time Magazine. To illustrate his point more concretely, however, he contrasts pairs of sentences ending in different ways, listing a few from Somerset Maugham. 54 He does not feel the need to indicate the original; neither will I. A. I have never had much patience with the writers who claim from the reader an effort to understand their meaning. B. I have never had much patience with the writers who claim from the reader an effort to understand what they mean. A. You would have thought that men who passed their lives in the study of the great masters of literature would be sufficiently sensitive to the beauty of language to write if not beautifully at least with perspicuity. B. You would have thought that men who passed their lives in the study of the great masters of literature would be sufficiently sensitive to the beauty of language to write if not beautifully at least perspicuously. A. It is natural enough that he should not find a precise expression for a confused idea. B. It is natural enough that he should not be able to express a confused idea precisely.^9 Again Williams suggests that although this kind of style is a mature one, requiring practice and experience, it is not difficult to process. Since the kinds of endings Williams suggests are easily identifiable and readily counted, they lent themselves well to my study. It will be interesting to see, in the following chap ters, how use of these "significant" clause endings correlated with the holistic evaluations given the papers, as well as with clause and T-unit length. One final aspect of Williams' paper is that while he rails against number of words per T-unit as tlie sine gua non of good writing, he does not give sentence-combining the kiss of death. He acknowledges that some students do write one ten word sentence after another, and would profit from some sort of com bining activity, providing they did not lose sight of the goals of clarity and directness. Toward those goals, Williams suggests two special kinds of modifiers characteristic of good expository prose, rather than the narrative writing Christensen's patterns seem directed to. These are what he calls resumptive and summative modifiers, and they are easily illustrated: (His examples) Resumptive: I would call the kind of writing that I've tried to develop during the last pages textured, writing that goes beyond the simplest communication of the simplest ideas, beyond the plainest of the plain styles. It is a structure that allows the writer to break the flow of a clause, a structure that lets the reader take a mental breath and then sustain the syntactic line for several more words. Like that. ^ The summative modifier is similar, providing a mental breath for the reader that an unbroken string of phrases and clauses prevents: 56 Summative: The plain style is by no means difficult to process, a , claim which I hope this paper has demonstrated. It is something that students would surely profit from, and might see more clearly than other structures because both resumptive and summative modifiers are merely a form of conscious repetition. Williams' work exhibits an awareness of the "con sequences" of writing, an awareness Winterowd stresses in his freshman text, The Contemporarv Writer. After a semester of teaching four freshman writing courses, it is easy for me to remember many instances of tangled prose that Williams' dictums could have helped untangle. His Ttfork, I think, goes far toward answering the ques tion of "what to count." Finally, a recent study in analyzing maturity on the college level showed that the three developmental factors of syntactic maturity (words per T-unit, clauses per T—unit, and words per clause) explained less than two percent of the variance in the holistic rating scores. The percentage of T-units with final free modifiers had the strongest influence on writing effectiveness (16%) but all five syntactic variables considered accounted for only twenty percent of the 52 variance in the ratings. The implications of this and of my own study will be considered in the final chapter. However, it is 57 important to note here that the literature on college- level students is finally moving away from the T-unit, viewing the writing process as a whole which is not necessarily the sum of its parts, but not as a mysterious "unteachable" process. The search for quantifiable characteristics is not, 1 think, solipsis- tic; rather it is an affirmation of our responsibility for providing substance to that still undefinable standard we hold up to our students, that standard we call Mgood writing." Practical Programs The Christensen Rhetoric Program appeared in 1968. It was a program too quickly overlooked in its vast possibilities; and for one main reason: the nomencla ture was difficult. The issues Christensen tried to deal with, and in fact the rationale underlying his program is outlined in his 1968 article "The Problem of 53 Defining a Mature Style." In that article, Christensen argues with the implications of Hunt's findings, and with some aspects of the Mellon study, which had appeared in its original form in 1967. As support for his arguments, Christensen cites the fact that teachers who read (College Board Fashion) Mellon*s pre-post essays rated the control group higher, despite the fact that the experimental essays doubled the syn- 58 tactic growth described by Hunt. Christensen stated the problem like this: Producing the kind of growth stipulated by these studies (Hunt and Mellon) may be possible, but is the kind of growth stipu lated the kind of growth we want? . . . Maybe the kids are headed in the wrong direction. Maybe the lines of their growth projected upward, would never meet the lines projected downward from the writing of skilled adults. Maybe, unless the direction is changed, unless the twig is bent, they will never write like skilled adults, but write, like most adults, the lumpy, soggy pedestrian prose that we justly deride as jargon or gobbledygo ok. Christensen's particular argument was with the long noun phrase, a phrase he described as "the very hallmark of jargon," easy to write and difficult to read. In his Freshman Composition text, The Contemporary Writer. W. Ross Winterowd devotes a section in his chap- 55 ter on the sentence to nomxnals and ease in reading. His concerns echo Christensen's with clear examples of "what happens when the predicate is deferred by an over packed nominal slot." The next sentence talks easily and lucidly about the process of manufacturing cars: Automatic machines, linked by transfer equipment, move engine blocks through a complete manufacturing process, performing 530 precision cutting and drilling opera tions in 14 1/2 minutes as compared to 9 hours in a conventional plant. Ben B. Seligman 59 But rewritten with an overpacked nominal slot: Automatic machines performing 530 pre cision cutting and drilling operations in 14 1/2 minutes as compared to 9 hours in a conventional plant, linked by transfer equip ment, move engines blocks through a complete manufacturing process. In the second example, the reader must hold 25 words in suspension before he reaches the predicate, thus creating the comprehension difficulty. Both Christensen and Winterowd are driving at the same point: it is not enough merely to teach students a wide range of syntactic devices; rather we must also teach them the effect these devices have on the reader. Students must become rhetorically as well as syntactically sophis ticated, writing always with an audience in mind. Recent studies, like the one by Joseph Williams, are beginning to show more awareness of the relationship between a work's syntactic fluency and its readability level. Christensen's answer was to introduce the con cept of free modifiers: relative clauses, noun clus ters, prepositional phrases, verb phrases, absolutes, subordinate clauses and adjective clusters. Below are some examples: Relative clause: The concert, which was success ful, ended an hour late. Noun cluster: The concert, a smashing success, ended an hour late. 60 Verb phrases The concert, ending an hour late, was very successful. Absolute: The concert having been successful, Ms. Hyler was pleased. Adjective cluster: The concert, successful, ended an hour late. His program was based on four principles: addition, direction of modification, levels of generality and texture. His examples were drawn from the great names in fiction— Hemingway, Steinbeck, Fitzgerald, Welty, etc. Here is an example of one of his exercises: 1 now draw much of (noun phrase) their support and encouragement from (noun phrase) 2 __________ (absolute phrase) 2 __________ (absolute phrase) 3 ____________________________ __________ (relative clause) Christensen's Program, then, not only combined sen tences, but forced the students to think of some original material. As with any innovative approach, the Christensen Rhetoric Program had its problems, and plenty of critics to point them out. One of these critics was Sabina Johnson, whose article "Some Tentative Strictures on Generative Rhetoric" initiated the logo machy between Bonniejean Christensen and herself. Mrs. Johnson has some legitimate concerns; however, I do 61 disagree with her on at least two counts. I included her article in this survey because it raises several points still uncertain in the teaching of sentence- combining. Mrs. Johnson disagrees with Christensen’s dictum not to teach initial verb clusters because students usually dangle them. She writes "I find that when students do dangle their initial verb cluster it is 56 easy to correct the fault.” Students may see this point at the time it is made to them, but they cheerfully continue to dangle modifiers anyway. Johnson also dis agrees with Christensen that form can generate content; however, William Strong, author of the very successful book Sentence-Combining: A Composing Process would agree with Christensen, and expressed his confidence in the generative powers of form at the Sentence-Combining Conference in Oxford. There are others in this camp as well, but the form/content argument is still very much alive. It was this contention of Johnson's that offended Bonniejean Christensen most. She replied: "Mrs. Johnson, as a matter of taste, rejects the idea that structure may be an aid to invention, that learning 57 how to say helps learn what to say." It is my own contention that as students gain con fidence in their "how-to" abilities, they also have more to say in writing, and more frequently. 62 However some other of Johnson's concerns are ques tions that underly this dissertation. Johnson contends that students seldom write sound, precise base clauses, and for them the writing of the cumu lative sentence is not unlike building a house on sand. Furthermore . . . they depend heavily on modifiers to make up for what they rightly sense to be the inadequacy of their base clauses. But I_ think I . can show that their choice of modi fiers is almost always relative and subordinate clauses— not the structures Christensen empha sizes— and that these long clauses complicate the problem in a way shorter free modifiers would h o t . Johnson's last point here is exactly the reason I chose not to count relative and subordinate clauses in my study, but restricted ray counting of free modifiers to verbals, absolutes, noun phrases and adjective phrases, those devices seldom employed in student writ ing. Students use the former structures almost to the exclusion of any other kind of modification. This is perhaps the great benefit of the Strong methods after working on combining several propositions in a small group, students vote on the best combined sentence. This way they arrive at three or four versions of a sen tence, learning from their classmates other ways to 59 present embedded materials than in a relative clause. Finally, Mrs. Johnson raises a question that has always interested me, a question that to my knowledge has not been tested: whether and to what degree 63 Christensen's method of building sentences can be made to carry over from the narrative and descriptive to the expository essay. To conclude the discussion of Christensen, Mrs. Johnson acknowledges that his work "is more promising than what we teach now." Also, A. M. Tibbetts applauds Christensen's attempt to integrate literature, rhetoric and grammar, something that most English departments still fail to do. But he warns: Unless the writer uses it carefully and with discretion, the cumulative sentence can tempt him to deviate from his logical course, caus ing him to pitch and yaw about in his para graph.60 Perhaps Christensen erred on the side of over ambition for his students; however, X have dwelled on his work at length because his was the first attempt at putting Sentence-Combining into a text, a program avail able to all students, rather than to those cooperating in an isolated experiment. Actually, while there was much experimentation performed between his time and the present, there was nothing in the way of a text for teaching sentence-combining to college students until the appearance of William Strong's book in 1973, and now Daiker, Kerek and Morenberg's book, The Writer1s Option: College Sentence-Combining. The recent studies cited here (Stewart 1978, Daiker, Kerek and Morenberg, 1978, Maimon and Hodine, 64 1978) have all used William Strong's Sentence-Combining: A Composing Book as the main text for their experiments. It is important to examine the kind of contribution Strong made to the field with the appearance of his book in 1973. No one describes a book's purpose better than its author. To quote Strong: Sentence-Combining is a ski11-building text. It won't help you find something to say when you're asked to write a research paper nor will it offer rules about organization style, usage, diction— the matters that composition books often take up. Its main purpose is to help you "hear" the stylistic options available to you and to help you "see" patterns of development, both in sentences and paragraphs. You'll probably find the skills more useful than any number of “rules." Over and over again, Strong emphasizes to the stu dent that he or she must read the sentences aloud. This goes back to the point I made in ray first chapter: what students frequently don't know (and professional writers do) is the sound of their words on paper. In a sense, Strong revolutionized sentence-combining. He removed grammatical terms from his exercises and relied instead on the writer's competence, on his innate ability to manipulate syntactic options and discover for himself the most effective way to say something. Here is a typical beginning exercise: Coffee 1. He sips at his coffee cup. 2. The cup is chipped along the rim. 3. The taste is bitter. 65 4. The taste is acidic. 5. The taste is faintly soapy. 6. There is a film. 7. The film is brown. 8. The film is on the inside of his cup. 9. He takes extra care. 10. The care is so that he doesn*t spill any on his clothes. 11. He is afraid. 12. The fear is that it might eat holes in the material.62 Strong suggests that students work together in groups on these sentences. Each group then presents its sentence, and all sentences are displayed on a transpar ency, using an overhead projector. In this way, stu dents get to see and hear several versions of the sen tence, and are able to discuss the rhetorical effects of the various options chosen. As a student progresses through the texts, the exercises and the suggestions following them become increasingly complex. Of course, one of the great merits of the book is the high interest level of its exercises, ranging widely in topics that included stereotypes, motorcycles, politics, holidays, etc. Even as I page through the book myself, I am challenged by the material, and would almost like to give it a go. Students are eventually encouraged to generate their own material, listing details to describe a particular situation, and then combining them into an effective paragraph. Results with the Strong book have been very good. 66 FOOTNOTES CHAPTER TWO This is what A. F. Watts calls "the natural linguistic unit." See A. F. Watts, The Language and Mental development of Children, pp. 65-66. See also Kellogg W. Hunt, "A Synopsis of Clause—to Sentence Length Factors,** English Journal 54 (1965): 300-309. 2 Kellogg Hunt, differences in Grammatical Struc tures Written at Three Grade Levels. The Structures to be Analyzed by Transformational Methods. Co-operative Research Project No. 1998, Florida: Florida State University, 1964, p. 12. 3 Kellogg Hunt, Sentence Structures Used by Superior Students in Grades 4 and 12 and by Superior Adults. Co-operative Research Project No. 5—0313, p. 10. ^See Stewart (1977) Daiker, Kerek and Morenberg (1978) and Stewart (1978). 5 Kellogg Hunt, Differences at Three Grade Levels. p. 59. 6 Hunt, p. 147. 7 Walter Loban, Language Development: Kindergarten through Grade Twelve. (Illinois: NCTE, 1975), p. 24. Q Loban, p. 31. 9 Year-to-Year Velocity is the percentage change in any given group from one year to the following year. 10_ . Loban, p. 48. 11 Loban, p. 60. 12 Murray F. Stewart, “Syntactic Maturity from High School to University: A First Look," Research in the Teaching of English. 12, No. 1 (February 1978), p. 45. 1 3 D. R. Bateman and F. J. Zidonis 1964, The Effect of a Knowledge of Generative Grammar Upon the Growth of Language Complexity. " CoTumbus: The Ohio State University. U. S. Office of Education Cooperative Research Project No. 1746. 67 14 Prank O'Hare, Sentence-Combining: Improving Student Writing Without Formal Crammar Instruction. (WCTE: Urbana, 111., 1973), p. 9. 15 John C. Mellon, "Issues in the Theory and Practice of Sentence-Combining: A Twenty-Year Perspec tive" Paper delivered at Conference on Sentence- Combining and the Teaching of Writing. Miami University, Oxford, Ohio. October, 1978. 16 / - * John G. Mellon, Transformational Sentence- Combining; a Method for Enhancing the Development of Syntactic Fluency in English Composition. Research Report fto. 10, (UrEiha, ill.: NCTE, 1969). A11 further references to Mellon in this chapter pertain to this study. 17 Bateman and Zidonis, p. 3. 18 See, for example, Elinor O. Keenan and Tina L. Bennett, eds., Discourse Across Time and Space. "Southern California Occasional Papers in Linguistics, Mo. 5" (Los Angeles: University of Southern California, 1977). 19 O'Hare, p. 11. All further references to O'Hare pertain to this study. 20 It is possible that both rhetorical choices and stylistic options are acquired in good writers. See Stephen D. Krashen, "On the Acquisition of Planned Discourse: Written English as a Second Dialect" in Handbook 42nd Claremont Reading Conference 1978, M. Douglas, ed. 21 O'Hare, p. 12. 22o*Hare, p. 19. 23 O'Hare, p. 27. 24 Mellon. Transformational Sentence-Combining.., p. 79. 250'Hare, p. 68. 9 fi O'Hare, p. 69. 27 James W. Ney, "The Hazards of the Course: Sentence-Combining in Freshman English," The English Record. 27 (1976), pp. 70-77. 68 28 Donald Daiker, Andrew Kerek, and Max Morenberg, "Sentence-Combining and Syntactic Maturity in Freshman English," College Composition and Communication. 29, Ho, 1 (February 1978), pp. 36-41. 29 Donald Daiker, Andrew Kerek and Max Morenberg, "Sentence-Combining at the College Level: An Experi mental Study," Research in the Teaching of English. 12, Ho. 3 (October 1978), pp. 245-256. 30 Daiker, Kerek and Morenberg, p. 247. 31 Daiker, Kerek and Morenberg, p. 252. 32 Murray F. Stewart, "Freshman Sentence Combinings A Canadian Project" Research in the Teaching of English. 12, No. 3 (October 1978), pp. 257-267. 33 Elaine Mairaon and Barbara Nodine, "Measuring Syntactic Growths Errors and Expectations in Sentence—Combining Practice with College Freshmen" Research in the Teaching of English. 12, Ho. 3 (October 1978), pp. 233-243. 34 Maimon and Hodine, p. 235. 35 Maimon and Hodine, p. 241. 36 Lester Golub and Carole Kidder, "Syntactic Density and The Computer" Elementary English. 51, Ho. 8, November-December 1974, pp. 1128-31. 37 Ellen Hold and Sarah Freedman, "An Analysis of Readers' Responses to Essays" Research in the Teaching of English. 11, Ho. 2 (Fall 1977), pp. 165-174. 38 Roy O’Donnell, "A Critique of Some Indices of Syntactic Maturity" Research in the Teaching of English. 10, Ho. 1 (Spring, 1976),ppl 31-38. 390'Donnell, p. 37. 40 J. F. Belanger, "Calculating the Syntactic Density Score: A Mathematical Problem" Research in the Teaching of English. 12, Ho. 2 (May 1978), p. 149. 41 Hold and Freedman, p. 174. 42 Joseph Williams, "Defining Complexity" Paper presented at 1978 College Composition and Communication Conference in Denver. 43 Williams, p. 5. 44Williams, p. 12. 45 Williams, p. 13. 46 Williams, p. 21. 47 Williams, p. 25. 4®Williams, p. 26. 4 9 Williams, p. 23. 50 Williams, p. 30. ^Williams, p. 30. 52 Lester Faigley, "Problems in Analyzing Maturity in College and Adult Writing" University of North Dakota. Paper presented at Sentence-Combining Confer ence, Ohio, October 1978. 53 Francis Christensen, "The Problem of Defining a Mature Style" English Journal, 57 (1968), pp. 573-579. 54 Christensen, p. 575. 55 W. Ross Winterowd, The Contemporary Wrxter. (New York: Harcourt, Brace, Jovanovich, 1975)', pp. 356-357. C Sabina Thorne Johnson, "Some Tentative Stric tures on Generative Rhetoric" in Contemporary Rhetoric: A Conceptual Background with Readings, ed. W. Winterowd TNew York: Harcourt, Brace, Jovanovich, 1975), p. 353. 57 Bonniejean Christensen, "Strictures on Mrs. Johnson's Strictures" College English. 31. No. 8 (May 1970), pp. 878-881. 58 Johnson, p. 358. 59 William Strong, "Doing Sentence-Combining: Some Practical Hints" Paper presented at Conference on S-C and the Teaching of Writing. Miami University, Oxford, Ohio. October 1978. 70 60 A, M. Tibbetts, "On the Practical Uses of a Grammatical System: A Note on Christensen and Johnson" Colleqre English. 31, No. 8 (May 1970), p. 874. 61William Strong, Sentence-Combining: A Composing Book (New York: Random House, 1973) p. xiiiT ^Strong, p. 11. 71 CHAPTER THREE METHODS AND RESULTS Hypothesis The comparison of holistically evaluated writing scores to various features of freshman texts centered on two hypotheses. First, that those students receiving a score in the upper third would be above the mean in a majority of the factors analyzed. Second, that upper- third papers would be characterized by a higher number of both free modifiers and Williams' "significant" clause endings. The Desicrn Sixty-six Freshmen from three Freshman English classes at the University of Southern California were the subjects of this study. The classes were chosen randomly, representing various time periods during the day. The Teaching Assistants who administered the essays were told that the essays were to be used experi mentally for the new Freshman Writing Program as a dry run for evaluating Freshman writing ability in future placement testing. Only this researcher and the administrators of the Freshman Writing Program knew that these essays would also be the data for this dissertation. 72 Students participating were asked to write for twenty minutes on the assigned topic, and told that they would receive their scores within the week. The same directions were read to all students. Test Questions In Moby Dick. Ishmael says: Whenever I find myself growing grim about the mouth; whenever it is a damp drizzily November in my soul; whenever I find myself involun tarily pausing before coffin warehouses, and bringing up the rear of every funeral I meet; and especially whenever my hypos get such an upper hand of me, that it requires a strong moral principle to prevent me from deliberately stepping into the street, and methodically knocking people's hats off— then I account it high time to get to sea as soon as I can. When you find yourself in that same condition, what do you do— eat a giant sundae, make a list of everyone you know who is worse off than you are, beat up your little brother, go for a long drive, call a friend? WRITING ASSIGNMENT: How do you handle depression or frustration? Consider the following points in your response: a description of what you do, an indication of its effectiveness, and an explanation of why it does or does not work. You have £0 minutes to write this essay; spend a few minutes planning. This topic was chosen because it elicits both descriptive and expository writing. Also, the adminis trators of the Freshman Writing Program felt if they were going to give students only 20 minutes (a 45 minute argumentative sample was administered to a 73 different set of classes), it would be best to provide a ‘ ‘personal experience" topic that would not give one student field-specific knowledge over another. Reading The readers for the holistic evaluation were a team of ten teaching assistants, all of whom had at least one year's experience in teaching composition. Also, all ten were students in the University of Southern California Rhetoric Program, a program designed to train people in the teaching of composition. The raters were trained during the first hour by Allan Casson, an associate professor in the department who had participated in holistic ratings for Educational Testing Service for a number of years. Professor Casson guided the raters through two sets of sample essays previously rated by the Freshman Writing Commit tee. The following rubric was used: Rubric. Four Point Scale. Reward the students for what they do well. A very well-written essay may be scored a point higher than it would be scored on the basis of content alone. A poorly written essay should be scored a point lower. A very poorly written essay must be scored in the lower half of the scale. Ignore occasional errors of spelling, grammar, punctuation, or omitted words that are likely to have been caused by the pressures of time. Do not judge by the length of the essay. 74 Do not judge by the readability of the handwriting. Do not judge an essay by a single foolish or aberrant sentence. Read steadily through the entire answer and make your choice on your judgment of the whole, deciding first whether the essay belongs in the upper or lower half of the scale. Use the whole scale. There will be l's and 4's, though fewer of them than 2's and 3*s. Bear in mind that if all the essays get 2's and 3's the essay section (of the Freshman Placement Test) loses weight and scores on the objective sec tion become more important. Do not think in terms of letters. A 4 answer is a four, not an A. Think only in numbers. Scores 4— A 4 answer will be a well-organized, well- written essay that deals with all of the parts of the writing assignment. It may deal with more than one response to depression, but it will deal with at least one in some detail. A 4 answer is likely to be concrete and confident, with a discernible formal shape. It may contain minor mechanical errors. 3— A 3 essay will cover all of the assigned questions, though it may be less thorough, less specific, less shapely than a 4 answer. 2— A 2 essay will respond only partially to the assignment. It may: — give adequate attention to one of the speci fied tasks but little to the others. — treat the t asks superficially — be lacking in supporting detail; — drift away from the topic or display con siderable irrelevancy; — have serious faults in writing. 1— This score should be given to any response that is on the topic but suggests incompetence. Papers that are blank or completely off the topic should be given immediately to the ques tion leaders. Papers that are not essays should be given immediately to the question leaders. 75 During the first hour, each reader received a packet containing the above rubric and essay questions, along with eight training essays judged by Professor Casson and the Freshman Writing Committee to be repre sentative of the quality of the group. After readers rated the eight training essays, Professor Casson led a discussion focusing on disagreements among the raters, and attempting to guide them to achieve similar responses. Then readers were told the criteria the Freshman Writing Committee had used to rate the training essays, and since these essays had already received a three-reader rating, they were removed, leaving the raters with 58 essays. However, these training essays were included in the sample used for this dissertation. The readers had a brief coffee break after the training session, then read and rated the remaining 58 essays, coded to conceal the identity of the writer. The essays were distributed to the raters in random order by the Freshman Writing Committee, who collected each essay after one rating, then passed it to two other readers, so that each essay was rated three times by three different raters. Dependent Variable A quality rating for each essay was determined by summing the scores given by the three readers. The quality score could range from a possible three (for 76 the worst essay) to 12 (for the best). The researcher eliminated from the sample those essays receiving more than a one point difference among raters. For example, if an essay received a 4, a 2, and another 2, it was discarded. Six essays in all were eliminated by this factor, so the total sample for this dissertation represents 59 essays. Independent Variables There were several considerations in selecting the independent variable to be used in this study. I thought it important to include the three variables contributing heavily* to composition research since the days of Kellogg Hunt: words per T—unit, words per clause, and subordinate clauses per T-unit. My next consideration, since I was essentially doing an analysis of reader's responses to essays, was to include some of the variables previously used in that kind of study. For this reason, I chose to adopt some of the variables from the Nold-Freedman study (1977), particularly those the researchers found most promising. Hold and Freedman found a high correlation between number of free modifiers used, and the quality rating the paper received. Also, Lester Faigley found that while Hunt's three developmental measures together accounted for only 2% of the variance in the quality ratings among his raters, number of free modifiers 77 accounted for 16%. Like Faigley, I did not distinguish among initial, medial and final free modifiers, but counted only the general category of free modifiers. Students use so few initial and medial modifiers that it is not worth the distinction. In addition, I used four other of the Nold-Freedraan variables associated negatively with quality: number 1 of common verbs, number of be*s and have's in auxiliary, number of modals in auxiliary, and their dummy variable, length. I collapsed the number of be's and have’s, and the number of modals into one general category: number of be’s, have's and modals per auxiliary. I did this because judging from the Hold study, both categories weakened the essays equally. Finally, I used the Williams' clause endings because I was hoping to find a measure of quality that would, in the tradition of sentence-combining, prove teachable. Williams' clause endings (a coordinate structure, a prepositional phrase beginning with "of," or a nominalization) met all of my criteria: they are readily identifiable, easily hand-counted, and I believe they may be simply taught either without sen tence-combining or more effectively, in addition to it. The list of variables, then, looks like this: (1) Words per T-unit (2) Subordinate clauses per T-unit 78 (3) Words per clause (4) % of be's, have's and raodals per finite verb (5) % of common verbs per finite verb (6) Number of free modifiers (7) Number of Williams* clause endings per T-unit (8) Mean number of words per essay (See Appendix A) Results and Interpretations Before presenting individual trends and correla tions, I would like to present the descriptive results of the data. Table 1 gives the means and standard deviations for the eight variables in the study (one dependent variable, quality, to be predicted and seven independent variables to predict). To determine which variables seemed the most reli able predictors of quality, and to ensure that the dis tinctions were clear-cut, I divided the essays into thirds. 10*s, 11*s and 12*s were upper third, 9's and 8's middle third, and 7 and below were lower third. No essay received a three, the lowest possible score. The distributions in these categories were, respectively, 23 essays, 14 essays, and 22 essays. Table 2 represents the number of essays in each third which were above the mean for the following variables: Words per clause, words per T-unit, free modifiers, Williams' clause endings per T-unit, and total words. According to Nold and Freedman better essays are below the mean in common TABLE 1 MEANS AND STANDARD DEVIATIONS FOR THE VARIABLES Mean S.D. Quality (Qual) 8.3 2.3 Words per T-unit (Wdprt) 15.04 3.16 Subordinate Clauses per T-unit (Subprt) .91 .33 Words per Clause (Wdprcl) 9.1 1.9 % of Finite Verbs with be, have, or modal auxiliaries (be-have-mod) 22.3 % 11.5 % % of common finite verbs (common) 63.0 % 12.4 % Number of free modifiers (Fmod) 1.7 2.2 Number of Williams* clause endings per T-unit (Wild) .37 .22 Number of words per essay 251.2 68.8 -4 V O eo verbs, and in be's, have's and modals, so those figures are recorded in the table. Table 2 shows that neither words per clause nor words per T-unit are significant as indicators of quality. Exactly 16.9% of the population in both the upper and lower third of the scores were above the mean of 15.0 in number of words per T-unit, while the discrepancy in words per clause for these two sections was only 1.7 percentage points. Only 18.1% of the lower third essays, and 7.1% of the middle third were above (or below, in the case of be-have-mod and common) the mean in five or more of the variables. Only one essay (upper-third) out of the 59 manipulated all seven variables successfully. The remaining four variables appeared promising as indicators of quality: free modifiers, Williams' clause endings, percent of common verbs per finite verb, (lowered to below .50 in table) and mean word length. Table 3 presents chi square values for these variables, using a 3 x 2 contingency table with 2 two degrees of freedom. The table reveals Williams' clause endings as significant beyond the .001 level. Wext are free modi fiers, falling somewhere between the .05 and .01 level of significance. The percentage of common verbs per 81 r-i o 0 4 0 4 i n H w r*4f 04 Pi ft? o o * i-O o o IX in c\ #-i r o O O O r-i ^ C Q* i 04“ »S* I TABLE 3 CHI SQUARE VALUES FOR VARIABLES CHARACTERISTIC OF QUALITY VARIABLE CHI SQUARE VALUE P Williams* ClprT 16.6 .001 Free Mods. 5.84 .05 Common Verbs 4.18 .06 Mean Length 3.56 .08 C O N) 83 essay is a less determining factor, but nevertheless, one that approaches the .05 significance level. Finally, length as a determiner is slightly above the .10 level, just approaching significance. The upper-third group did have a higher percentage of essays above the mean of 251.2 words than either of the other two groups. 84 FOOTNOTES CHAPTER THREE ■^The following were counted as common verbs: be, break, do, feel, follow, find, get, give, go, have, hold, keep, know, look, live, make, mean, need, put, say, see, seem, stop, take, talk, tell, think, turn, try, want, work, write, use. 2 Chi Square shows the relationship between the individual variable and the essay's placement in upper, middle or lower third of the scoring range. 85 CHAPTER FOUR DISCUSSION, COMPARISONS AND CONCLUSIONS This is the first time the Williams' clause endings have been tested as a determining variable in holistic evaluation; something should foe said concerning where their significance lies. The structures demonstrate a kind of syntactic control not mastered by poor writers. Although their usage does not guarantee an essay an upper third score, the structures were used almost exclusively by upper and middle third writers. Lower third essays were characterized by a statistically significant lack of the clause endings, a lack of what Williams tentatively termed "significant style." According to Williams, the clause endings are markers of clear expository prose, rather than of descriptive or narrative. Since the expository style is the one most needed at the college level and beyond, control of these structures would be a useful tool for the Freshman writer. The findings here reaffirm Christensen's notion that free modifiers, particularly the tighter structures such as absolutes and verbals, are indicative of a more sophisticated style. The Freedman-Nold study also supported the Christensen hypothesis, as did the 86 Paigley study. The Paigley study found that incidence of free modifiers influenced raters in making a quality judgment more than any other syntactic factor. Comparisons between the Freedman-Nold study and the present one are found in Table 4. Thirty-six of the fifty-nine essays used free modifiers. Where they appeared, there was an average of 2.8 modifiers per essay (overall average = 1.7). In the Freedman-Nold study, 41 of the 88 essays used free modifiers with an average incidence of 1.5 per essay (where they appeared). However, mean word length of modifiers was higher in that particular study: ten words as compared with the present 6.4. In this study, free modifiers were found in 52.1% of the upper third essays, and 57.1% of the middle group, but only in 22.7% of the lower third essays. Nold and Freedman also found that a high number of be's, have's and modals as auxiliaries weakened the essay. While this was my own intuitive feeling as I read through the papers, chi squares on this factor yielded no significance. Nold and Freedman also found that essays using less than their mean of 54% common verbs were rated high. I found that twenty-one of the twenty-two essays in the lower third used over 50% (well over in most cases) common verbs, so I think it TABLE 4 FREE MODIFIERS: A COMPARISON WITH THE NOLD STUDY Avg. Length of Fmods Fmods (% of Wds. In Free Modifier Position) Incidence of Fmods In Essays Where They Appear Martin 6.4 3.8% (total essays) 6.2% (essays where they appear) 2.8 Nold 10.0 4% (Final) 1.5 2% (Medial) 00 -4 88 reasonable to assume that superior essays are character ized by more vivid verb use. It is equally important to discuss some of the variables which were not significant, i.e., words per T-unit, words per clause, and subordinate clauses per T-unit. Neither words per T-unit, nor words per clause proved distinguishing factors in the essay's rank; 43.4% of upper third essays were above the 15 words per T-unit mean as were 45.4% of the lower third. This finding correlates with Faigley's study where words per T—unit, clauses per T-unit, and words per clause together accounted for less than 2% of the vari ance in qualitative scores. Nold and Freedman also concluded that words per T—unit and other standard developmental measures are not useful in predicting perceptions of quality on the college level. Table 5 provides comparisons of the standard developmental measures here with other studies. When I undertook this study, I thought Nold's findings were perhaps due to Standford students' unusually high mean of 17.2 words per T-unit. But among USC students with a mean of 2.2 words less, there was still absolutely no correspondence. Perhaps, then, on the college level and above, it is no longer how many words, but rather how well they are used and in what structures. Which is not to deny the efficacy TABLE 5 COMPARISON OF PRESENT FINDINGS TO PREVIOUS STUDIES Wds pr T Wds pr cl Sub cl pr T % be-have-mod Martin 15.0 9.1 .91 22.3 Hunt (Grade 12) 14.4 8.6 1.68 Hunt (Superior Adults) 20.3 11.5 1.74 Loban (12) (Random) 13.2 31.0* .60 Loban (12) (High) 14.0 33.8* .66 Loban (12) (Low) 11.2 27.5* .52 Stewart (1st. yr. Univ.) 12.25 8.2 1.48 Stewart (5th. & 6th. Univ.) 13.85 8.81 1.60 03 TABLE 5 (continued) Hold 17.27 10.6 (main 8.4 (sub. cl.) cl.) .77 19. 10. (modal) (be-have) Daiker (pre-test) 15.31 8.75 1.76 Daiker (post-test) 16.05 9.64 1.68 * words in dependent clauses as a % of words in T-units V O o 91 of sentence-combining; Daiker, Kerek and Morenberg's experimental group were judged superior in all six rating criteria: ideas, supporting details, organi zation and coherence, voice, sentence structure, and diction and usage. Sentence structure was defined for the raters as "the extent to which variety, maturity and effectiveness of sentence structure is achieved." Based on my findings, X would hypothesize that if Daiker et al. had examined free modifiers and Williams' clause endings among their variables, the experimental group would have used a higher number than the control group. The point here is that perhaps we should teach a kind of "selective" sentence-combining, emphasizing some structures and de-emphasizing others. Consider the implications of the longitudinal research. Use of dependent clauses seemed to precede the more sophisti cated verbal structures. Perhaps we can't "skip over" this stage if students haven't reached it. Students writing at different levels may need different types of sentence combining practice. For example, many Basic Writing students cannot work with Strong's Sentence-Combining because subordination patterns are unfamiliar. Here is how one student manipulated the "Coffee" exercise (see p. 60): 92 He sips at his coffee cup. It is chipped along the rim. The taste of which is bitter, acidic and faintly soapy. The brown film on the inside of his cup. The extra care is so that he doesn't spill any on his clothes which he is afraid that it might eat holes in the material. The purpose of the Strong book is to prod the stu dent into using various syntactic options, and making rhetorical choices he or she was previously unaware of. But the assumption underlying this purpose— that the student already possesses an internal competence which only needs to be activated— is not valid for college level Basic Writing students. For this reason, then, these students would perhaps profit from working with cued sentence-combining until they have mastered some basic sentence patterns. In other cases, students use too much subordination. Subordinate clauses may add more words per T-unit than do verbals or absolutes; yet as Loban has shown, they are a less sophisticated structure, over-used by his low group, and decreasing in use among more mature writers in grades eleven and twelve. In the present study, 65% of the upper-third writers were below the mean of .91 subordinate clauses per T-unit, a greater percentage than in either of the other two groups. Since the upper third also used more free modifiers, it could be that word per T-unit and clauses per T-units as non—significant variables reflect a difference 93 between the kinds of structures used by upper and lower third writers. Lower third writers clutter their essays with a variety of "which*sn and "thats," adding length to their T—units, but not skill to their writing. Some additional support for this hypothesis is gained from the Daiker et al. study. While their experimental group gained in words per T-unit and words per clause after exposure to sentence-combining, they decreased in number of clauses per T-unit, indi cating that their longer T-units were using structures other than the over-worked subordinate clause. A third problem for sentence-combining is the writer who is almost there, but one who needs a bit of guidance. One who uses absolutes and verbals easily, but whose lack of parallelism is noticeable^and jarring enough to keep it out of the upper third range of essays. Below are the first and last sentences of that kind of writer's essay: Whenever I become totally angry or displaced with my immediate environment, I feel the instant physical need to reach out and destroy something, whether it be merely a bicycle standing upright waiting to be kicked down, a door swung open so I can slam it back again, or inflicting some of my pain upon an individual that is enjoying a pleasanter moment than myself. Nevertheless some of the fury of my frustrations do exude past my most adamant expression, and I find myself helplessly crushing an innocent piece of fresh paper, and then mercilessly 94 plummeting it to its demise on the bottom of my wastepaper basket. Obviously, a teacher cannot hand this student a copy of Strong's Sentence-Combinincr and tell him or her to work in it. And this is the dilemmanof the Freshman Composition class: the levels of competency are very mixed, and one prescription will not remedy all the ills. The writer of this essay needs to be warned against overwriting, against using vocabulary merely to impress, and against sentences that are long at the expense of grade and ease of expression. However, this kind of writer will learn quickly to recognize and to revise for these finer distinctions, once they are pointed out to him, because unlike the student of Basic Writing, he already possesses a fair amount of language fluency. It is a different case entirely from the student who writes one ten word sentence after another, and cannot see the point of "correcting" and revising sentences which to him are error-free. Some Exceptions Statistics, of course, do not tell all. There were a number of essays using all the right variables, yet still receiving low scores, as well as the reverse phenomenon. Some discussion of these anomalies may help to determine other factors influencing raters involved in holistic scoring, factors more difficult 95 to measure empirically, such as content and organi zation. First I would like to present two contrasting essays for analysis. One essay used twelve free modifiers, more than any other essay in the study, and was also characterized by a low percentage of common verbs. Yet it was a lower third essay, receiving a total score of 7. It was an imaginative account of a lone basketball player acting out his frustrations on the court. I quote a few sentences to illustrate the writer's variety of expression: For me the best way to relieve the pressures of school or work is to "pop the pill," no, not that type of pill. I am referring to that wondrous 28 inch leather ball which spins through the air like a satellite. . . . Alone on the court opportunitys [sic] abound [sic] that ever so elusive hoop sus pended ten feet above the ground chalanges [sic] you to beat it as you know you must. . . . One by one the shots fall going straight down the hoop and shaping the beautiful white cord net with a swishing sound. As the ball hits the ground it has just that perfect amount of rotation on it that sends it scurring [sic] back to you, like a dog to his master. . . . And the eloquent conclusion: "Just look for a lone boy soaring to the basket with the ball cocked behind his head, like Dr. J., ready to jam it home in an effort to take on the world." This essay followed the rubric, describing the means the writer used to alleviate frustration and depression, and stating subtly but succintly why that 96 means was effective: "Deep down you know however it is you against the world and the world is that basket." In many cases, (but not all) the writer did ignore conventional punctuation and his spelling was less than accurate, though not appalling. I trans cribed the quoted passages exactly as they appeared. However, there were no sentence fragments, only one run-on sentence, and when the writer did remember his punctuation, he used it correctly, suggesting that its omission in other cases may have been due to the pres sure of the 20 minute time-limit. Raters were warned to take this into account, yet two out of three rele gated this to the lower half of the scale, giving it two out of a possible four points. I think most readers would agree, however, that the essay demon strates a variety and richness of expression, using metaphor in an apt and original fashion, something that rarely happens in freshman essays, and something, I think, that should outweigh its mechanical weakness. Let me compare it to passages from the following essay which, to my amazement, received a perfect score of twelve. When I feel depressed, I usually go out with a bunch of my friends or take a long drive. . . . A long drive is probably my single most useful tool to aleviate [sicj depression. A long ride permits me to think over my problems, it relaxes me. It allows me the time to search for answers 97 and solutions. A drive is especially useful to me, if I take it in the country. The beautiful scenery allows me to take my mind off my problems. And the conclusion, a typical rehash of the previous paragraphs: "A long drive helps me to overcome my depression because it gives me a time to think, to reflect on my problems. Ususally when I remember them, they seem much less important." Trite, banal, few specific details. Then why a 12? Well, the organization was beyond reproach, for one thing. The first paragraph introduced the topic of frustration and depression, and stated 2 possible solutions: (1) going out with other people or (2) a long drive. The second paragraph (not quoted) dealt with solution number one, explaining its effectiveness with impeccable logic: "When I am surrounded by happy people, I usually become happy." Paragraph number three revolved around solution number two— the long drive and its effectiveness: "Sometimes after I have relaxed or forgotten my prob lems for a while, I realize that they are not worth worrying about or they seem less important." The fourth paragraph summarized once again why being around people is an effective alleviator of depression, and the last paragraph re-summed the bene fits of a long drive. The paper followed the directions specifically, had only one grammatical error (a run-on), 98 and few spelling errors. In fact, it is a perfect example of the kind of paper that presents the greatest difficulty for the Freshman Composition teacher. The difficulty arises when attempting to explain to the student who has just received a "C" what, exactly, is wrong with the paper. Well, what is wrong with it? Looking at the tabulations, I can say that it has a low 13 words per T-unit, and 7.9 words per clause, uses only one free modifier, and has a high percentage of common verbs. Unfortunately, that won’t satisfy my student to whom these terras mean nothing. And they don't entirely satisfy me. But ray alterna tive is to tell him or her that the essay is dull, repetitive, lacks specific details, vivid vocabulary, achieves no variety of expression, and uses no sentence structure beyond an eighth grade level. But after a blank stare, "it's organized," will undoubtedly be the reply; "It's organized the way I was taught in high school." As yes, so it is. And proof that organiza tion, coupled with minimal mechanical errors, is some times rewarded above all else lies in the essay's perfect score of 12, a score given it by raters who should have college-level expectations. Because these two things are the only virtues of this lack-lustre essay. 99 I could refer my student to the basketball essay, pointing out to him or her the vivid verb usage, and the richness of the figurative language. But chances are the student, like the essay raters, would pounce immediately on the essay’s loosely structured organi zation and number of mechanical and spelling errors. What have we wrought by preaching the doctrine of correctness? I realize that in choosing these two particular essays to quote from, I am in some sense guilty of setting up a straw man. But not entirely. I've had too many students with essays like the 12 telling me they always got A's in high school, and too many with the imagination and competence level of the basketball writer telling me of their C's and apologizing for their "awful spelling." Other things being equal, the problems of the 7 score writer are remedied easily enough by a few weeks work in a Writing Lab. Certainly more easily than the lack of imagination and poor language manipulation of his peer with the "perfect" score. And most teachers of Freshman Composition would recognize that, I'm sure. But for those occasional lapses when lack of punctuation causes the profuse spilling of red ink, let us remember which values in writing are more important before we go around the 100 bend entirely by assigning a fat ”C" and commenting on the bottom of the page "Your mechanics are atro cious I M And let us not be so mollified when an essay exhibits rote organization and grammatical correctness that we write "Bh good grasp of mechanics'1 1 and move hurriedly on to the next paper. Sometimes reach exceeds grasp, and sometimes that reach should be rewarded. However, cumulative use of the variables generally produced essays ranked high for their stylistic sophistication. Essays above the mean in five out of the seven variables appeared twice as frequently in the upper third category as in the lower third, and seven times more frequently than in the middle third. Style, then is an important factor in holistic evaluation, an important factor in reader-response. If it can be measured quantitatively, to some degree at least, it possesses identifiable features. In that case, it is up to composition teachers to take style from the nebulous regions of their mind and give it to their students in the concrete terms repre sented by the variables studied: Williams' clause endings, free modifiers and vivid verb usage, to name the three most significant. It is time to translate the intuitive response we feel to a well or 101 poorly written paper into tools the student can work with to build his stylistic repertoire. Finally, though, style as measured by syntactic control is not everything. Style is also represented by a successful adjustment to the audience. When this adjustment, this creation of empathy, is highly successful, it may outweigh weaknesses in manipulation of syntax. Essays low in the variables measured but receiving high scores were marked by this quality. They handled the escape from frustration and depression in a unique way, with a sense of self that made the actual writing of the essay appear to be therapeutic. The following are sample passages illustrating this skillful achievement: Essay 25— Score = 12 There are other times, however, when I don't know what to do. I take long walks at night, but often to no avail; I only depress myself more. Sometimes the only thing which can bring me back to ray senses is a cold shower. It wakes me up from something which is like a nightmare; it stimulates my nerves and makes me feel alive again. When I'm depressed, I'm so remote from reality that it seems as if an alien force has taken over, and is now controlling my every move. Who can ignore the aptness of those descriptions? Few among us can deny having similar feelings at one time or another. Even though this essay was one of the shortest in the group, it said enough for readers to empathize, placing it in a higher bracket. 1 0 2 One more brief example from an essay highfin words per T-unit, but low in everything else. In this essay, I think the writer’s simple but accurate analysis of her means of coping strikes a responsive chord. She recognizes her devices for what they are and says so— something that many of us are not always able to do. She was rewarded for her perception. Whenever I find myself frustrated or « depressed, I like to surround myself with friends and keep myself busy so I can forget about what’s bothering me. My friends and I will usually go to a movie, play tennis or maybe go somewhere to eat. Some may say I am running away from my problems but this is not always the case. Activities such as those I’ve listed are sometimes just excuses to be with:;people I love and trust. I feel that my way of handling depression and frustration mayanot always solve what’s really bothering me, but it helps me get through thosepperiods in my life. It is not easy to devise a way to measure the 1 extent to which a writer draws in his or her audience. But I think all four sample essays illustrate that this kind of style is also a factor in holistic rating, where the reader is encouraged to read quickly and impressionistically. Audience involvement is one of the most elusive skills to teach; the closet we seem to come is to talk about focus and narrative detail, and to provide as many examples of writing successfully in4 :this capacity as possible. "Tone** is a hard term to define for a______ 103 beginning writer. One student can write about the time she didn't make cheerleading with a pathos that will make you weep; another writes on the same topic and reader's only reaction is "so what?" If we could translate the difference between those two extremes into a pedagogical device, we could come far toward teaching successful writing, and farther still toward a more complete concept of style. Limitations of the Study The variables examined here were somewhat limited in scope. Looking toward directions for future re search, an analysis of the transitions described by 2 Winterowd in "A Grammar of Coherence" would be an excellent measure to correlate with a holistic rating. And perhaps number of embeddings per T-unit might be 3 a more accurate measure of syntactic maturity. However, as this was ray first venture into research in the field of composition, I felt there was some merit in my covering the traditional ground as well as applying one new measure— the Williams' clause endings. The last limitation has to do with the rating system itself. While a holistic rating is an appropriate tool, research has shown that it is most valuable when combined with analytic or forced- 4 choice ratings. 104 FOOTNOTES CHAPTER FOUR ■*"See James Britton, The Development of Writing Abilities (11-18)• (New York: Macmillan, 1975). 2 W. Ross Winterowd, Contemporary Rhetoric: A Conceptual Background with Readings. (New York: Harcourt, Brace, Jovanovich, 1975) p. 225. 3 See Walter Kintsch, "Notes on the Structure of Semantic Memory" in Organization of Memory. Tulving and Donaldson, eds. (New York: Academic Press, 1972) pp. 249-305. 4 See Cooper and Odell, Evaluating Writing. (Urbana, 111.: NCTE, 1977) and Daiker et al., 1978. 105 CHAPTER FIVE IMPLICATIONS: PEDAGOGICAL AND THEORETICAL Implications for Teaching: After Sentence-Combining. What Next? Sentence-Combining, for all ©f its proven effectiveness, cannot care all the ills of student writing. For one thing, it has been my experience that it can frequently lead to the hinds of errors tabulated by Maimon and Nodine, errors in parallelism, dangling modifiers, vague pronoun reference, etc. Perhaps we need to work at developing a kind of "selective” sentence-combining, one that would empha size some structures over others, particularly for college level programs. For example, as I have pointed out in earlier chapters, some students already use an over-abundance ©f relative clauses. And for some, sentence-combining only encourages this over-use. These students, then need to learn a reduction process: clause to phrase, and sometimes phrase to word. Reduction exercises can certainly be incorporated into sentence-combining; students might be given a variation of the Martin Joos exercise of writing a well-known fairy-tale in 200 words, then 100, then 50. Joos claimed that "gerunds 106 and absolutes began popping out all over the place."1 My own advice to students is simply to go through their writing, getting rid of as many "which*s" and "that's" as possible. Also, this study and the Hold study both pointed to verb-use as a determining factor in writing quality. I think sentence-combining can be used to encourage more vivid verb choice, as well as to teach students the principle of rearrangement. I have noticed that when I tell students to go through their papers and rid them of all the "is," "was" and "have" verbs, they will frequently return to ray desk with the following problem: they cannot replace their "empty" verbs with another verb because they have boxed them selves into a linguistic corner. They are right; another verb in that particular position would not make sense. But it doesn't occur to students to rearrange their whole sentence in a way that would allow their verbs more flexibility. For example, one student wrote: "When we were driving to the concert it was raining, which was a shame because the scenery was beauti ful." Except for changing "was raining" to "rained," the student was in a real quandary concerning the other verbs. It simply did not occur to him to rearrange the sentence to eliminate his problems— 107 something like this: "It rained while we were driving to the concert, spoiling the beautiful scenery." He was locked into a lengthy clause structure, and a strict subject-verb arrangement. My point is that we need to realize, and in turn emphasize to our students, that there is more to sentence-combining than longer sentences. It can be used to demonstrate any number of problems that arise at the revision level, and perhaps to help the stu dent make what are now conscious choices part of his internal competence, enhancing his repertoire of sty listic options. I would also like to develop the teaching of the clause endings researched here, endings Williams claims mark an "emphatic and confident style." These too could be included as part of sentence-combining. Once again, they involve an element of reduction; they are short and summary in purpose. They are not relative clauses, nor even complete clauses since they have no subject and verb. They are merely types of endings, easily encouraged by example and practice. Perhaps they could be taught by contrasting pairs of sentences as Williams did in his paper, then later providing one half of the pair, and asking students to write the other by ending the sentence a different way, using a 108 coordinate structure, a prepositional phrase beginning with "of," a norainalization, or some combination of them. I think more and more we need to draw upon the work of professional non-fiction writers to learn what constitutes clear expository prose. For that is really the mode of writing most students will use, no matter what their future occupation. From the auto mechanic writing a list of car repairs to the engineer drafting a project proposal, all need to write clearly and concisely with varying degrees of style in between. The non-fiction articles in magazines like Harper * s. Atlantic and Saturday Review frequently must explain concepts like solar energy or changes in the curriculum at Harvard to a somewhat unknown audience who may be largely unfamiliar with the subject. Students writing persuasive or expository essays have much the same task. They are writing on a subject of concern to them (at least in good composition classes) and must convey their knowledge to an audience equally unknown. For example, many of my students are from farming areas, and choose to explain a particular farming method for their expository essay. That's fine, but many have great difficulty with audience, accusing me of prejudice against farmers when I tell them they must define terms like "la amaze method" and "combine." Their answer: "Everybody I know knows what these 109 terras mean.” They are astounded that their classmates from Chicago may not know these terms, and even more astounded at the idea that one day their audience may not be so homogenous, and that they may still wish to get their ideas across. Suppose they are applying for government aid for new implements. They will need to explain what they need and why to the bureaucrats at the other end of their request. So I think Williams is quite right to draw his markers of clear expository prose from an article on lasers in the New York Times Magazine. That is pre cisely the kind of source we need to research for our students to bring them the kind of help they need, and to help them write sentences that are not only longer than 10 words, but clear and forceful as well. Toward this end, Daiker, Kerek and Morenberg have developed a new program: The Writer1s Options: College Sentence-Combining. (New York: Harper and Row, 1979). Their program is as carefully designed as the research that led to its development. In their chapter on appositives, they deal precisely with the patterns suggested by Williams as appropriate to expository prose (suramative and resumptive modi fiers, Chapter Two). It also helps resolve my problem of ”too many which’s and thats” in a simple yet effective way. Here is a sample passage from the Chapter: 110 An excellent. all-around student with a con genial personality, she is a promising candi date for WOW, an international scholarship program for outstanding women around the world. In this last sentence, an excellent. all- around student with a congenial personality is an appositive to she. and an international scholarship program for outstanding women around the world is an appositive to WOW. The second revised sentence is five words shorter than the first; it states more con cisely the very same idea that in the first sentence requires the connectives since and which. as well as the repetition of she and is. By replacing whole clauses, the appositives seem to make the second sentence move more briskly. But it jus important to be aware of your options; sound them out, listen to them, weigh them. Increasingly, your final choice will be the "right" one. Test your ears and eyes on sentences (a) and (b) below to decide which is more effective: a. Even a brief visit to Greece, which is a modern gateway to the glory of the past, gives you a profound sense of the roots of our civilization. b. Even a brief visit to Greece, a modern gateway to the glory of the past. gives you a profound sense of the roots of our civilization. Clearly, which is in (a) in unnecessary, and it could even cause momentary confusion as you read the sentence, even a brief visit to Greece. which is . . . At this point, can you be sure which refers to Greece and not to visit? Using an appositive instead of which can be a smart way of avoiding possible confusion. The chapter overcomes the great weakness of the Christensen program: it does not deal with "pretty" examples from fiction writing, examples that intimidate the student because he is unable to see how he can Ill make them work in his own writing. Instead, it focuses on samples that might appear in modern expository prose. The book is divided into three sections: "Struc tures," "Strategies," and "Beyond," a section dealing with the selection and organization of ideas, moving the student into the whole essay. The "Strategies" section includes a unit on a problem mentioned earlier, rearrangement. It helps the student to realize the various rhetorical effects created by shifting a sentence around. In fact, I think far more than any other textbook I have seen on sentence-com bining, this one successfully integrates sentence-com bining with the whole process of composition. Those who extol the virtues of teaching literature over composition sometimes forget that they are in fact teaching the end product of a composing process. To some extent, I sympathize with their complaints about Freshman Composition. We all tire of explaining what sentences are, and beyond that, how to give them some variety. I know sometimes it feels as though we have explained the same concepts over and over until we are brought to the brink of lunacy. But research into all the aspects of composition, and especially research directed to the demands of the various modes of discourse, is essential if we are 112 to achieve professional growth. We cannot be satisfied with increases in only the syntactic maturity of our students. Logical and organizational maturity, as well as mastery of tone and audience are needs that should impel us to intensive theoretical and pedagogical activity. Our scholarship in turn should profit the whole profession: colleagues in every area of the discipline, and our students at all levels. Implications for Research Each time I read through my essay sample, and whenever I grade my freshman papers, X realize how much research remains to be done in the field. One area needing more exploration is the hind of error analysis done by Maimon and Nodine. Students who have worked with sentence-combining successfully do produce longer T-units; however, some are less success ful than others in the number of errors they are able to control. Of these errors, the dangling verbal and faulty parallelism are the most glaring. This goes back to my statement in the introductory chapter that many students are not aware of the sound of their words on paper. An even more important aspect of the Maimon- Nodine study was its concern with the various modes of 113 discourse, an aspect that researchers are increasingly turning to as we learn to explore levels of conceptual difficulty. It is important to know which kinds of x^riting tasks are most complex for the student, and will therefore cause him or her difficulty in con trolling syntactic maturity, diction, mechanics, organization and content simultaneously. The question we have to answer is this: At what conceptual level do the concerns of content interfere with the student's performance and negate the effects of previous class room practice with syntactic manipulation? Error analysis and classification may help us to forewarn the student; as we work with sentence-combining in the classroom, we can anticipate under what circum stances these errors will occur, and let the student practice sentence-combining simultaneously with the manipulation of complex ideas. The kinds of errors examined by Maimon and Nodine are easily identified, and therefore measurable. They also work well with teaching a selective sentence-combining. Students can be pre-tested for their level of syntactic control, then work with the appropriate level of sentence-com bining (as outlined in the previous chapter). The number and kinds of errors resulting from their attempts to use the skills they are acquiring should help gauge when they are ready for a higher level of sentence- 114 combining. In this fashion, the student will acquaint himself with the equal importance of style and content. In our zeal for empirical research, however, we must guard against a tendency to look for "new things to count,” solely because they will yield measurable results. We must not be afraid of less measurable, more exploratory studies dealing with content, and the sometimes elusive qualities of tone and voice. For in reading Freshman essays, it is plain that some students know how to take a personal experience and endow it with those qualities that draw a reader into the work, creating empathy. Others ramble on and on, i<rith no awareness of which details focus the incident, which only distract the reader, and which are utterly inconsequential. For example, a student of mine wrote an essay whose main focus (I learned on the last page) was a rock concert he attended in the mountains, and the kinds of effects created by the musicians involved. But he rambled for about five pages, providing details which included the following: who he went with, how long the drive was, where they got lost, how much beer was consumed, where the car was parked, etc. By the time he got to describing the concert itself, a description fairly well-written in comparison with the rest of the paper, he had ex hausted his writing time for the week, and my patience. 115 In talking with colleagues and in monitoring my own grading process, I have come to realize that a large part of the difference among grades given to Freshman papers reflects how well the student has handled his or her audience, a definite sign of maturity. Can this handling be measured by counting various syntactical devices? Perhaps in part— the student who uses "you" all the time is generally not sure how to involve his reader other than by a direct approach unacceptable in most kinds of adult writing. Britton characterizes the school situation as "the set task and the extent to which the writer makes it his own," claiming that this "makes the difference 3 between involved or perfunctory writing." However, Britton concluded that except for writing in the poetic mode, permitting expressive use of language, it was difficult to distinguish perfunctory from involved writing. The reason for this, Britton suggests, is that the ability to use a detached style of address appro priate to an unknown public audience is a relatively late development while creative writing is not. It is precisely in this area of "relatively late develop ment" that we need research. Some freshman have it; some don't. What if anything, can we measure as characteristic of this development, and most important 116 as the end result of any research, how can we teach it effectively? The answer will not be found in number of words per T, clauses per T, or even, X think, in number of free modifiers. Though as Williams suggests, it may appear somewhat in the kinds of modifiers chosen, specifically the summative or resumptive modifiers discussed in Chapter Two. Perhaps we have to examine also the various semantic categories of words and concepts used by the successful audience-adjuster to see how he or she abstracts from a personal experience, or point of view, those elements that will best lend themselves to a broader, though unknown, audience. More simply put, how does a writer turn “what happened to me" or ' ’what I think" into "at some time, in some situation, we've all had these feelings" and "what do you think?" There are probably some measurable markers, and some that are not so measurable. But it is certainly an area that needs more research. James Kinneavy raised some interesting questions in his paper "Sentence-Combining in a Comprehensive Language Framework," Kinneavy reported that sentence structure predicted 69.6% of the grades in the Daiker et al. study. He suggested a possibility that the holistic evaluator underwent "seduction by stylistic maturity," and questioned whether English teachers knew 117 enough about inductive validity to rate a paper in 4 the same way a science teacher might. I would like to think we are not so easily seduced, but there are ways we could test ourselves in this area, and either achieve vindication or suffer the indictment of such a claim. It would be easy enough for us to use a philosophi cal essay question, or perhaps even a scientific process essay, as the post-test for students who had worked with sentence-combining. The only difficulty would lie in ensuring that students were at the same level in the subject area. This can be done by having our colleagues in philosophy or science devise a fairly simple objective pre-test which is then administered to the students in a sentence-combining class, and graded by the philosophy or science teacher to ascer tain the students' knowledge in the field. With a sample of students who are at a fairly equal level, the writing teacher can proceed with his or her own post-test essay, then have it holistically graded by English teachers. At that point a group of teachers from the selected field can be trained in holistic rating, using the same rubric and the English teachers and the results compared. Only then could we evaluate how, much, if at all, we are seduced by stylistic maturity over content. 118 Kinneavy also raised two other points dealing with the effect of sentence-combining practice on students' writing. Kinneavy suggested that since the Daiker et al. method of teaching sentence-combining encouraged the evaluation of students' sentences by other students, perhaps the semantic discourse criteria was the most important emphasis in sentence-combining. Students then developed their own metalanguage, forcing them to use what may be their implicit knowledge of rhetorical principles. This new awareness of rhetori cal principles may be, more than anything else, what carries over into their writing, lending it more maturity. Finally, Kinneavy hypothesized we learn to write by writing, and that the act of writing itself, the large amount of writing done in the Sentence-Combining classes, was responsible for the improvement. As support for this hypothesis, he cited a study done at Oklahoma State where traditional teaching methods were used, but lots of actual writing took place. Students gained significantly in Hunt's measures of maturity, surpassing the levels he cited. Kinneavy concluded that too much theory in the composition class prevents learning to write. My own view is the less time spent lecturing in a writing class, the better. The teacher's job is simply to guide the process. 119 To quote Irwin Weiser, a colleague of mine at Illinois State, “some students don't know what writing looks like, or how it should feel coming out of their 5 pens.” Separating the effects of sentence-combining from the effects of writing practice itself is diffi cult, unless one has a traditional class as a control, and time devoted to writing in both classes is care fully measured and controlled. Maybe, in the end, technique is not so important as an evaluation of whether or not the students have measurably improved in their ability to write clear prose that is stylistically graceful and conceptually mature. In any case, we cannot be satisfied with any one method until our results are continuous and gratifying, until we can demonstrate to our students and to our colleagues in other disciplines that a course in Freshman Composition has not been a waste of time. 120 FOOTNOTES CHAPTER FIVE 1 Martin Joos, The Five Clocks. (New York: Harcourt, Brace, and World, 1962), pp. 68-80. 2 Daiker, Kerek and Morenberg, The Writer*s Options: College Sentence-Combining (New York: Harper and Row, 1979), pp. 43-44. 3 James Britton et al. The Development of Writing Abilities. (11-18) (New York: Macmillan, 1975), p. 78. 4 James Kinneavy, "Sentence-Combining in a Comprehensive Language Framework." Paper presented at Sentence-Combining Conference at Miami University, October, 1978. 5 Irwin Weiser, "Imitation in the Basic Writing Course." Unpublished manuscript. 121 BIBLIOGRAPHY Bateman, D. R. and Zidonis, F. J. The Effect of a Knowledge of Generative Grammar Upon the Growth of Language Complexity, Columbus: The Ohio State University, U, S. Office of Education Cooperative Research Project no, 1746, Belanger, J. F. "Calculating the Syntactic Density Score: A Mathematical Problem" in Research in the Teaching of English. 12, No. 2 (May 1978T7 pp. 149—151. Braddock, Richard, Lloyd-Jones, Richard, Schoer, Lowell, Research in Written Composition. Champaign, 111,: NCTE, 1963. Breland, Hunter M. “Can Multiple-choice Tests Measure Writing Skills?" in the College Board Review No. 103 (Spring 1977), pp. 26-30. Chatman, Seymour. Literary Style: A Symposium. New York: Oxford University Press, 1971. Chatman and Levin. Essays on the Language of Literature. Boston: Houghton Mifflin, 1967. Christensen, Bonniejean. "Strictures on Mrs. Johnson's Strictures" College English. 31, No. 8 (May 1970), pp. 878-881. Christensen, Francis. "The Problem of Defining a Mature Style," in English Journal. 57 (1968), pp. 572-579. Cicourel, Aaron V. et. al. Language Use and School Performance. New York: Academic Press Inc., 1974. Cooper, Charles and Odell, Lee. Evaluating Writing: Describing. Measuring. Judging. Illinois: NCTE, 1977. Daiker, Donald, Kerek, Andrew and Morenberg, Max. "Sentence-Combining & Syntactic Maturity" College Composition and Communication. 29 (February 1978), pp. 36-41. 122 Daiker, Kerek and Morenberg. "Sentence-Combining at the College Level: An Experimental Study: Research in the Teaching of English. 12, No. 3 (October 197877 PP- 245-256. B'Arcais, Giovanni B. Flores and Levelt, Wm. J., eds., Advances in Psycholinguistics. New York: American Elsevier Publishing Company Inc., 197G. Diderich, Paul B. Measuring Growth in English. Urbana, 111.: NCTE, 1974. Dolozel and Bailey, eds. Statistics and Style. New York: American Elsevier Publishing Comp., Inc., 1969. Faigley, Lester. "Problems in Analyzing Maturity in College and Adult Writing." University of North Dakota. Paper presented at Sentence-Combining Conference, Ohio (October 1978). Foreman, Ellen and Haring, Lee. "Folklore in the Freshman Writing Course" College English. 37, No. 1 (September 1975), pp. 13-21. Freeman, Donald C. Linguistics and Literary Style. New York: Holt, Rinehart and Winston, Inc., 1970. Gibson, Walker. Tough. Sweet and Stuffy. Indiana: Indiana University Press, 1966. Golub, Lester and Kidder, Carole. "Syntactic Density and the Computer" Elementary English. 51, No. 8 (November-Deeember 1974), pp. 1128-31. Grimes, Joseph E. The Thread of Discourse. Paris: Mouton, 1975. Hough, Graham. Style and Stylistics. London: Routledge and Regan Paul, 1969. Hunt, Kellogg W. Differences in Grammatical Structures Written at Three Grade Levels. The Structures to be Analyzed by Transformational Methods. Coopera tive Research Project 3 1998, Florida: Florida State University, 1964. Hunt, Kelloggr Sentence Structures Used by Superior Students in Grades 4 and 12 and by Superior Adults. Cooperative Research Project No. ¥-0313, p. 10. 123 Kachru, Braj B. and Stahlke, Herbert P. Current Trends in Stylistics. Illinois: Linguistic Research, Inc., 1972. Kingston, Albert J. Toward a Psychology of Reading and Language: Selected** *Writings of Wendell W. Weaver. Athens: The University of Georgia Press, 1977. Kintsch, Walter. The Representation of Meaning in Memory. New York: John Wiley and Sons, 1974. Kintsch, Walter and Keenan, Janice. "Reading Rate and Retention as a Function of the Number of Propositions in the Base Structure of Sentences" Cognitive Psychology. 5, No. 3 (November 1973), pp. 257—274. Lindberg, John. "Composition: Cult or Competence," in North American Review. Winter 1977, pp.' 11-21. Loban, Walter. Language Development: Kindergarten through Grade 12. Urbana, 111.: NCTE, 1976. Love, Glen and Payne, Michael. Contemporary Essays on Style. Oregon: Scott Foresraan and Company, 1969. Mairaon, Elaine and Nodine, Barbara. "Measuring Syntactic Growth: Errors and Expectations in Sentence-Combining Practice with College Freshmen" Research in the Teaching of English. 12, No. 3 (October, 1978), pp. 233-243. Mellon, John C. "Issues in the Theory and Practice of Sentence-Combining: A Twenty-Year Perspec tive" Paper Delivered at Conference on Sentence- Combining and the Teaching of Writing." Miami University, Oxford, Ohio. October, 1978. Mellon, John C. National Assessment and the Teaching of English. Urbana, 111.: NCTE, 1975. Mellon, John C. Transformational Sentence-Combining: a Method for Enhancing the Development of Syntactic Fluency in English Composition. Research Report No. 10, (Urbana, 111.: NCTE, 1969). 124 Messing, Gordon M. “The Impact of Transformational Grammar Upon Stylistics and Literary Analysis’ * Linguisties. 66 (February 1971), pp. 56-73. Milic, Louis T. Stylists on Style. New York: Charles Scribner’s Sons, 1969. Ney, James W. “The Hazards of the Course: Sentence- Combining in Freshman English,” The English Record. 27 (1976), pp. 70-77. Hold, Ellen and Freedman, Sarah. "An Analysis of Readers' Responses to Essays" Research in the Teaching of English. 11, No. 2 (Fall 1977), pp. 165-174. Norman, Donald A. and Ruraelhart, David E. Explorations in Cognition. San Francisco: W. H. Freeman and Company, 1975. O'Donnell, Roy. "A Critique of Some Indices of Syntactic Maturity" Research in the Teaching of English. 10, No. 1 (Spring 1976), pp. 31-38. O'Hare, Frank. Sentence-Combining: Improving Student Writing Without Formal Grammar Instruction. (Urbana, 111.: NCTE, 1973). Sebeek, Thomas A. Style in Language. Massachusetts: The M.I.T. Press, 1960. Steinberg, Erwin R. Needed Research in the Teaching of English. Proceedings of a Conference Held at Carnegie Institute of Technology, May 5-7, 1962. Steinmann, Martin. New Rhetorics. New York: Scribner's Sons, 1967. Stewart, Murray F. "Freshman Sentence Combining: A Canadian Project" Research in the Teaching of English. 12, No. 3 (October 19787', pp. 257-267. Stewart, Murray F. "Syntactic Maturity from High School to University: A First Look" Research in the Teaching of English. 12, No. 1 (February 197817 p. 45. 125 Strong, William. “Doing Sentence Combining: Some Practical Hints" Paper Presented at Conference on S-C and the Teaching of Writing. Miami University, Oxford, Ohio, October 1978. Strong, William. Sentence Combining: A Composing Book. New York: Random House, 19~3. Tulving and Donaldson, eds. Organization of Memory. New York: Academic Press, 1972. Winterowd, W. Ross. Contemporary Rhetoric: A Conceptual Background with Readings. New York: Harcourt, Brace, Jovanovieh, 1975. Winterowd, w. Ross. Rhetoric: A Synthesis. New York: Holt, Rinehart and wTnston, Inc., 1968. 126 APPENDIX COUNTING PROCEDURES Finite Verbs: I counted all finite verbs, including those in subordinate clauses within T-units. The number of Be's, Have's, and modals and the number of common verbs were counted only when they were part of a finite verb. Free modifiers: I did not count relative clauses used as free modifiers, since students use these excessively; nor did I count prepositional phrases. The following structures were counted as free modi fiers: noun clusters, adjective clusters, verb phrases (present and past participles; infinitives), and absolutes. Williams' clause endings: These are counted only in the last clause or phrase of the sentence. If there were two nominaiizations within the last clause, I counted this as only one kind of clause ending. (for example, "frustration and depression" = one nominalization).

Asset Metadata

Core Title 00001.tif

Tag OAI-PMH Harvest

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC11257588

Unique identifier UC11257588

Legacy Identifier DP23074

Linked assets

University of Southern California Dissertations and Theses