These computers are great! This new technology will revolutionize not only the production of writing, but the teaching of good style as well.
--The Enthusiast

Why should I learn about the computer? It may help my students produce a prettily-printed page, but substantively their writing will be the same: the same grammatical errors, the same wordy, flaccid style.
--The Skeptic

The Enthusiast and the Skeptic are both right: The computer is revolutionary; the computer is not a cure-all. Among the computer programs that the Enthusiast might tout is WRITER'S WORKBENCH, a set of computer programs, developed at Bell Labs to run under the UNIX operating system, that can analyze prose documents and suggest improvements. Its set of programs can proofread a text, analyze the style of the text, and provide online information about English usage and the WRITER'S WORKBENCH programs themselves (Bell Labs, 1982, p. 1-1).

WRITER'S WORKBENCH has already received considerable attention in various computer and writing journals. Notably, Smith and Kiefer (1982) heralded the package after a year's use at Colorado State University; their enthusiasm elicited a number of skeptical responses--see, for example, Oliver (1984). Use of WRITER'S WORKBENCH has spread to many campuses, and teachers at CSU continue to report good results with thousands of students using WRITER'S WORKBENCH (Kiefer & Smith, 1983, 1984; Reid & Findlay, 1986).

Based on a more limited and selective use at St. Olaf College, it appears that WRITER'S WORKBENCH is less than what a public relations specialist would have us believe, but considerably more than what the unbelievers profess. WRITER'S WORKBENCH is an effective aid to good writing. The program's ability to identify such things as potentially misspelled words, split infinitives, use of passive voice, and other common problems in diction provides an antidote to some of the most frequent, most persistent, and least interesting problems of writing. Moreover, because WRITER'S WORKBENCH is such a powerful collector of data, it allows the user to compare a text to standards based on similar works written by other college writers, rather than on some arbitrarily set criteria. Finally, the program acts as an effective writing coach or editor, drawing various elements to the attention of the user and giving the opportunity to edit and revise the text, depending upon the user's purpose.

The use we have made of WRITER'S WORKBENCH at St. Olaf is a function of the kind of school we are and of the kind of background and facilities for computing we have. Located in Northfield, Minnesota, 40 miles south of Minneapolis and St. Paul, St. Olaf is a four-year, liberal arts college with 3000 students and 225 faculty members. St. Olaf has a well-developed, unified computer system which is used by faculty and students in all areas of the curriculum for computation, analysis, information management, some computer-assisted instruction, special projects in the arts, and most notably text processing, which accounts for more logged-in time than any other use.

St. Olaf sees the Computer Center as a resource, much like the Library, which should be available to all students and faculty free of charge (Introduction to Academic Computing at St. Olaf College, pp. 1-2, 14-15). There is no college-wide or department-wide policy requiring use of text editors or of WRITER'S WORKBENCH. But the College encourages the use of these programs through the instruction sessions offered by the Computer Center that introduce students to text editing and to the use of computers for other purposes as well. In addition, the College provides extensive documentation, tailored to its computer environment, to support text editing and to lead users (students, faculty, administrators, staff, and their families) to more advanced uses of the system. [1]

In 1983, St. Olaf purchased the 2.0 version of WRITER'S WORKBENCH because of its general attraction as a style-analysis package, because it seemed the best such resource then available, and because it would be generally compatible with our UNIX operating system. With access to the source code, we fully integrated WRITER'S WORKBENCH into our system, and we also made a number of changes to the package itself. [2] Six months after we acquired the package, several members of the English Department began introducing WRITER'S WORKBENCH to selected sections of first-year English and advanced composition courses.

WRITER'S WORKBENCH consists of 28 basic programs. We have concentrated on WWB, a master program which controls eight other programs. The following diagram illustrates the relationship among these programs, all of which, except Split, can be run separately: [3]

The Enthusiast quoted at the outset would be correct about much of WRITER'S WORKBENCH--it provides useful information that should help users correct or avoid errors in their writing and develop their style. These essential virtues are easily illustrated by Punct, a primitive and mechanical program that identifies apparent errors involving quotation marks, apostrophes, and parentheses. first, it can do some things as well as or better than the human editor; for example, Punct will always highlight the illogical typographical convention governing the placement of commas inside quotation marks. Second, it frees the human reader to make judgments about comma splices and other more sophisticated aspects of punctuation. (See similar conclusions reached by Harris and Cheek, 1984).

Other features of WRITER'S WORKBENCH attractive to the Enthusiast include Style (which measures word and sentence length, discriminates sentence types, and identifies parts of speech), the subprograms of Style (which help the user focus on particular issues highlighted by Style), and Prose (which comments on sentence variety, use of verbs, and readability based on data gathered by Style). finally, WRITER'S WORKBENCH does not force a user to compare an essay to some arbitrary standard; instead, WRITER'S WORKBENCH enables groups of users to establish particular standards for different kinds of essays, based on 20 or more excellent student essays having the same rhetorical purpose. These features are major advantages of WRITER'S WORKBENCH over similar but less powerful style-analysis programs. [4]

But the Skeptic can also find things to point to in WRITER'S WORKBENCH. The package is less than a cure-all because the output must be used with care and proper guidance. This is true not simply because the content of some of the responses are inappropriate for college writers but, more fundamentally, because WRITER'S WORKBENCH cannot discern the appropriateness of such things as word choice, sentence types, and level of style for every situation. This lack of discrimination is most clearly illustrated by problems in Diction, which are analogous to the difficulties with the quantitative aspects of Prose.

Diction locates and displays certain incorrect, wordy, commonly misused, and possibly sexist expressions and displays each sentence in which they occur; Diction also displays a "Table of Substitutions" suggesting alternative expressions. Besides a number of minor imperfections (for example, the format of the "Table" is often unclear), Diction has several shortcomings, especially from the perspective of the student user. For example, the 2.0 version flagged "girl" every time it appeared and suggested it be replaced with "secretary" or "assistant." This suggestion reflects the package's origin as an aid for computer documentation writers at Bell Labs. Diction does not, however, effectively help college writers deal with sexist language. No doubt, the 3.0 version and the Collegiate edition now available from Bell Labs will have edited the contents of the dictionaries with the college user in mind. But, even careful tailoring of the contents of the dictionaries and revising the comments will not eliminate the basic problem with Diction: It cannot distinguish the use of various expressions that are wrong, or at least infelicitous, in some situations but which are perfectly fine, often exactly on the mark, in other situations. Not all beginning college writers are able to tell the difference.

During the summer of 1984, several English Department faculty wrote Using WRITER'S WORKBENCH at St. Olaf to help our students use and interpret WRITER'S WORKBENCH. [5] Here is an extended excerpt from it concerning Diction that illustrates the way we try to guide our students use of WRITER'S WORKBENCH:

Diction matches expressions in your text with a dictionary of problem words and phrases. It flags these words or phrases regardless of context; not everything it flags needs changing. Moreover, Diction cannot identify all errors in your text. Nonetheless, it is useful in identifying and suggesting alternatives.

[At this point, the document provides several examples from a student paper and the WRITER'S WORKBENCH analysis of it, both of which are reproduced in full in the document. Here is one example and the conclusion of the section on Diction.]

Not every suggestion should be acted upon. Diction routinely flags "which," as it did in this essay, and displays a list of possible misuses. In fact, in this case, "which" is correct.

In cases like this, think of Diction as a thesaurus. For each word in its glossary, a thesaurus lists numerous "synonyms," yet no two words are exactly synonymous. Therefore, when choosing a word to substitute, you must first decide if the new word fits. (pp. 16-17)

In our view, students need similar guidance interpreting the quantitative responses in Prose. lust as the responses in Diction may not apply to a given sentence in a student essay, the quantitative output in Prose will not always apply to a given student essay as a whole. For example, being told that an essay can easily be read by someone with six or more years of schooling may not help the author think constructively about how to revise a short story full of colloquial sentences. In an argumentative piece on a complicated subject, short simple sentences may be perfectly in keeping with the author's purpose. Variety for variety's sake does not lead to good writing. Students need some guidance (such as we provide in our document) in taking advantage of the results of WRITER'S WORKBENCH concerning these and other areas.

Despite the package's inherent shortcomings, which a thorough Skeptic might focus on, we have continued to introduce our students to WRITER'S WORKBENCH. We encourage them to use it, not with the slavish abandon of an Enthusiast, but with realistic care.

Those of us at St. Olaf using WRITER'S WORKBENCH have employed it as part of our overall approach to teaching writing. We have integrated it into our syllabi, usually introducing it as an aid to revision when we begin the revision-intensive stage of our courses. Using an introductory pamphlet, we introduce WRITER'S WORKBENCH to students, require them to use the package a specified number of times, and then leave it to them whether or not to continue using WRITER'S WORKBENCH.

After introducing WRITER'S WORKBENCH teachers report receiving more polished products. As was noted in previous comments on WRITER'S WORKBENCH, students themselves have information available to them to improve papers before teachers see the papers (Kiefer & Smith, 1983).Similarly, our students report that they have assimilated a variety of points about grammar and style through using WRITER'S WORKBENCH. So, because students proofread more carefully, their papers arrive on the teacher's desk with many common errors having been avoided or edited out.

The faculty agree with most of the assumptions about good writing underlying WRITER'S WORKBENCH. As my colleague Mary Steen put it:

Variety in sentence structure and length is good. Active voice is better than passive. Nominalizations thwart lucid prose. "It is" and "there are" make flabby sentence openers. Plain words are better than fancy ones: 'use" is better than "utilize." One word is stronger than a phrase: prefer "many" to "a large number of." A verb should replace a phrase: "decide" substitutes for "arrive at a decision." In other words, clear and unpretentious writing is the goal. (quoted in Day, 1985, p. 18)

Not surprisingly, the developers of the package specifically cite Strunk and White's Elements of Style (1972) and Richard Lanham's Revising Prose (1979) as having determined the principles of good writing built into WRITER'S WORKBENCH (Bell Labs, 1982, 5.5-7).

Use of WRITER'S WORKBENCH need not be confined to dealing with final drafts. Information provided by the package can be integrated easily into conferences by providing another point in the composing process where the instructor can engage the student writer. For example, in addition to discussing brainstormings, outlines, and drafts of papers, the instructor can also discuss the WRITER'S WORKBENCH output with a student. "What does it mean," the instructor might ask, "that your paper uses passive verbs 25% of the time, whereas most good essays of this kind written by first-year students at St. Olaf use the passive voice less than 10% of the time?"

We have tried to ascertain student response to WRITER'S WORKBENCH systematically in three ways. First, we asked two questions on our general course evaluation forms: "Do you recommend that other students be informed about WRITER'S WORKBENCH?" and "Should WRITER'S WORKBENCH have a place in English 11?" Some students were very enthusiastic in their responses to these two questions ("Yes." "Yes."); others were skeptical ("No." "No."). But most were more cautious in their responses than either the Enthusiast or the Skeptic. Here is a typical middle-of-the-road answer: "Yes, students should be aware of WWB and should use it. It is helpful to some extent."

This cautious appreciation characterized the tenor of student responses to our second effort at student evaluation. Beginning in the fall of 1984 and in all subsequent semesters, we asked students to respond to the following open-ended question after their first use of WRITER'S WORKBENCH at or before the middle of the term: "Comment on the usefulness of WRITER'S WORKBENCH in revising your essay." Although some students were skeptical of the efficacy of parts of WRITER'S WORKBENCH, almost all students saw value in one or more parts of the package. To paraphrase one student, "I'll take whatever help I can get from wherever it comes." Almost everyone appreciated the value of the spelling checker (although one student commented that it fostered a servile dependence everyone should learn to spell, in her opinion). Most found the minimal punctuation section useful. Few commented on either Double (which identifies where a word may be inadvertently repeated) or Split (which identifies split infinitives).

Students were about evenly divided on Diction, which highlights potential problems of word usage. Some found it very helpful in locating wordy expressions: "It helped me take out a lot of unnecessary words and focus on what I was saying." But others discovered shortcomings: "The word choice section . . . offended me at first, because . . . the phrases it printed out were exactly what I wanted to say and had no connection whatsoever to the meaning in the table of substitutions."

Concerning the Prose output, some students didn't like to think of their writing in quantitative terms. Others didn't want to deal with issues of their writing in ways the on-screen messages suggested. More typically, students made observations like the following: "It did make me think a little about types of sentences, lengths, and word choices. It provided a starting point for my revision."

Our third and most comprehensive student evaluation involved a detailed end-of-the-semester WRITER'S WORKBENCH questionnaire first circulated in the spring of 1985 and then in all subsequent semesters. The results of this questionnaire helped sharpen our understanding of how students were using the package. All the respondents, about 175 students, came from seven sections of first-year English in which they were required to use WRITER'S WORKBENCH. Not all respondents replied to all four central questions on our survey, and some made more than one response to a given question. (See Appendix for a tabulation of the results.)

We first asked students whether using WRITER'S WORKBENCH led to making changes in their writing processes. Five students commented in general about the package's usefulness in revision; three reported that they changed their processes, using what they learned from WRITER'S WORKBENCH to include an extra draft. Except for these eight students, it appears that WRITER'S WORKBENCH did not, in the students' estimation, lead them to change their writing processes significantly. As indicated, over a third (59) said "No" or "Not really" directly. Most of the other responses did not refer to issues of process; instead, they specified particular aspects of writing that WRITER'S WORKBENCH commented on.

When asked about using WRITER'S WORKBENCH for revision (Question 2), some students responded with a global "Yes" (28) or "No" (13), but most students volunteered a great number and variety of specific comments. These included not just spelling (15), but word choice and related issues (a total of 26), sentence-related issues (a total of 20), and other specific grammatical issues (5). Although only nine commented directly on how they used WRITER'S WORKBENCH in revision, the specific responses indicate that students did use WRITER'S WORKBENCH to advantage in revising. In light of the shortcomings of the package discussed above, it is notable that at least 26 students paid attention to issues of word choice highlighted by Diction (despite the problem that Diction cannot discriminate context) and that an almost equal number (20) paid attention to comments on variety in Prose.

When asked if they had learned specific things about their writing by running WRITER'S WORKBENCH and applying this new knowledge (Question 3), students responded with a total of 84 items learned. These ranged from very specific instances of word choice or punctuation to broader aspects of sentence-related issues, to a general comment about incorrect phrases or conjunctions. A significant minority (34) indicated little or nothing learned.

Taken together, the responses to these three questions suggest several positive effects of using WRITER'S WORKBENCH. Students used WRITER'S WORKBENCH for more than a spelling checker; they learned specific things about grammar and style, and they assimilated these insights during the process of composition. As suggested previously, fewer errors and fewer stylistic blemishes reached the teacher's desk.

The responses to these three questions also highlight some areas of concern. Remarkably, students did not see using WRITER'S WORKBENCH as affecting their writing process significantly, nor did most students generalize from the specifics learned to a more abstract sense of the package's place in revision. This lack of abstraction, combined with the wide range of specifics in the student responses, supports to some extent critics of style-analysis programs who contend that these packages focus student attention on surface characteristics. In addition, the absence of responses to some questions and the negative responses to others are worth noting: Some students did not learn much at all from WRITER'S WORKBENCH.

Finally, in Question 4, we asked specifically about earlier student comments on the statistics that appear in the Prose output (see p.8). The instructors were worried, too, that some users would slavishly make revisions so as to get their statistics within the suggested ranges. Although some said simply "Yes" (12) and more simply "No" (46), most responded with a qualified "Yes" that echoed their instructors, who were at some pains to emphasize that the "acceptable ranges" were suggestive, things they ought to consider. (See Appendix for a partial listing of responses.)

Use of WRITER'S WORKBENCH is slowly spreading at St. Olaf, but it is unlikely to be imposed upon any large group of instructors or students, although there has been some talk of introducing the package as an option for advanced students in our writing-across-the-curriculum program. Many of my colleagues (especially those most comfortable with using the computer) use WRITER'S WORKBENCH in their own writing, and knowledge of the package is spreading by word of mouth among students. The students we surveyed agreed that more people should know about WRITER'S WORKBENCH as an available option.

The responses to our surveys confirm our initial assumption that WRITER'S WORKBENCH is neither a cure-all (as some Enthusiasts might claim) nor a technological gimmick (as some Skeptics might charge). Rather, it is a useful tool and an aid to good writing. The attitude we espouse in Using WRITER'S WORKBENCH at St. Olaf carries over into our comments to students about the package. We continually remind our students that WRITER'S WORKBENCH highlights parts of an essay so as to suggest alternate phrasing of possibly incorrect expressions or makes other observations about that essay. The writer is always in charge and must decide whether or not to make changes. When to make changes, and what changes to make, must be determined by criteria not specified by the package. We have found that using WRITER'S WORKBENCH as a part of our approach to teaching writing is effective. We have also found that our strategy of introducing the package as an aid to revision, after some larger issues of writing have been discussed, is a good one.

We also remind our students that doing everything WRITER'S WORKBENCH suggests will not guarantee the quality of a paper. Many of our students are dealing with issues of writing more sophisticated than those with which WRITER'S WORKBENCH is able to deal. As a result, some students may see running WRITER'S WORKBENCH as not worth the effort; they lack the patience to read the output carefully to determine which suggestions apply and which do not. This may continue to be true even when the dictionaries and comments are more directly tailored to student users--comments will continue to appear that do not apply to the particular case in a given student essay.

Nonetheless, our experience and the student responses to our surveys make clear that WRITER'S WORKBENCH does help re-enforce notions of good writing and does help students deal conceptually with issues of writing. WRITER'S WORKBENCH helps students focus on problems in their writing and makes solving those problems easier. It draws users' attention, helping those users become healthily more self-conscious and, in the end, better writers.

When used with appropriate guidance and caution, with an awareness of what it can do and can't do, WRITER'S WORKBENCH is a useful tool. Moreover, it helps teachers hold students to higher standards, at least in the surface characteristics of their styles. And, if they choose, it allows students and teachers to focus their human attention on more interesting and complicated issues of writing--on the higher reaches of questions of grammar and style, on persona, tone, purpose, and audience.

John T. Day teaches in the Department of English at St. Olaf College, Northfield MN.

1. For additional information about the St. Olaf computing facilities and the use of text editors, see Day(1985) and Holdstein (1988).
   

2. We edited the various on-screen messages, adjusted the format of the output, changed the numbers from decimals to integers, reset the range of acceptable standards in several areas, changed various routing commands, simplified certain flags, and made other alterations to make the package more "user-friendly."
   

3. For a complete description of WRITER'S WORKBENCH, see Bell Labs (1982). For a full treatment of the package as used at St. Olaf, see Using WRITER'S WORKBENCH at St. Olaf (1984). Other schools have chosen to emphasize different parts of the package. The fullest and most readily available discussion of the package as implemented elsewhere is Kiefer and Smith (1984), which discusses its use at Colorado State University. For current information on WRITER'S WORKBENCH see WRITER'S WORKBENCH User's Group Newsletter (WUGN), by C. R. Smith and B. Stewart and published by the Department of English at Colorado State University, Fort Collins, CO 80523.
   

4. Several other programs related to those in WWB illustrate other aspects of the power and versatility of WRITER'S WORKBENCH. Findbe, which displays a copy of an essay with all the to be verbs in capital letters and underlined, helps students get at problems connected with passives, expletives, and nominalizations. Spelltell, Punctrules, Worduse, and Splitrule are interactive programs enabling students to get behind the on-screen messages to learn more about grammatical issues that the package comments on. Dictadd, which allows the user to add words or phrases to the dictionaries that form the basis of Diction, Spell, and Sexist, makes the package more adaptable to individual users.
   

5. This pamphlet has two parts: a 24-page introduction (including an overview, instructions for using the package, a sample student paper with accompanying output, a discussion of how to interpret the results, and a listing of further uses) and a 12-page appendix that summarizes all 28 commands in the package. It is available from the Academic Computer Center, St. Olaf College, Northfield, MN 55057.
   

Bell Labs. (1982). UNIX WRITER'S WORKBENCH Software User's Manual. Piscataway, NJ: Bell Labs, System Training Department.

Day, J. T., Frey, O., Hunter, L., Schwandt, P., & Steen, M. (1984). Using WRITER'S WORKBENCH at St. Olaf. (Available from the Academic Computer Center, St. Olaf College, Northfield, MN 55057)

Day, J. T. (1985). Writing and computers: The St. Olaf experience. Selected papers from the conference on computers in writing: New directions in teaching and research, University of Minnesota, April 1984 [Special issue]. Computers and Composition.

Harris, M., & Cheek, M. (1984, February). Computers across the curriculum: Using WRITER'S WORKBENCH for supplementary instruction. Computers and Composition, 1(1), 3.

Holdstein, D. (1987). On computers and composition. New York: Modern Language Association.

Introduction to academic computing at St. Olaf College. (1982). (Available from the Academic Computer Center, St. Olaf College, Northfield, MN 55057)

Kiefer, K. E., & Smith, C. R. (1983). Textual analysis with computers: Tests of Bell Laboratories' computer software. Research in the Teaching of English, 17, 201-214.

Kiefer, K. E., & Smith, C. R. (1984). Improving students' revising and editing: The WRITER'S WORKBENCH system. In W. Wresch (Ed .), The computer in composition instruction: A writer's tool (pp. 65-82). Urbana, IL: National Council of Teachers of English.

Lanham, R. A. (1979). Revising prose (1st ed.). New York: Macmillan.

Oliver L. J. (1984). Pitfalls in electronic writing land. English Education, 16, 94-100.

Reid, S., & Findlay, G. (1986). WRITER'S WORKBENCH analysis of holistically scored essays. Computers and Composition, 3(2), 632.

Smith, C. R., & Kiefer, K. E. (1982, April 1-3). WRITER'S WORKBENCH: Computers and writing instruction. Proceedings of the Future of Literacy Conference. Baltimore, MD: University of Maryland.

Strunk, W., Jr., & White, E. B. (197 '). The elements of style (2nd ed.). New York: Macmillan.

1. Has using WRITER'S WORKBENCH led you to make any changes in your writing process? If so, what kind of changes?
   

   No or Not really   59
   

   Specified changes:
   Word choice, especially avoiding
     unnecessary words   23     passives   9
     nominalizations   4     to be verbs   4
   

   Sentence structure and variation   17
     sentence openers   4
   

   Other specific, single-response comments   6
   

   Non-Specific comments:
     General comments about its importance in revision   5
     Changed writing process to include an extra draft to use data from WRITER'S WORKBENCH   3
     Made writer more aware of writing abilitie   1
   

2. Do you usually revise your essays? If so, how have you made use of WRITER'S WORKBENCH for revision?

   Yes   28     No   13
   

   Most specified one or more aspects:
     spelling   15
     spelling and word choice   6
     wording   23
     wording and sentence types and length   2
     wording and punctuation   1
   

     sentences   4     sentence variety   2
     beginnings   1     phrases   4
     sentence length   2     sentence structure   7
   

   Single grammatical responses:
     to be verbs     punctuation
     awkward expressions     readability
     grammatical errors
   

   Single general responses:
     touch-up
     for reference
     for first copy before revision
     for revision
     haven't used it enough for revision
     by paying attention to what they say
     to recheck minor mistakes for final copy
   

3. If WRITER'S WORKBENCH has flagged something several times, do you find that you then avoid that problem on your own before running WRITER'S WORKBENCH? In other words, have you learned specific things by running WRITER'S WORKBENCH? If so, what have you learned?
   

     word choice or unnecessary words (often specified)   40
     to be verbs   7     passives   5
     spelling   4     nominalizations   2
     adjectives and adverbs   2
   

     sentence types and variety   6     sentence structure   5
     sentence openers   4 sentence     length   4
   

   Single grammatical responses:
     dangling modifiers     quotes and commas together
     expletives     space after comma
     incorrect phrases or conjunctions
   

   Negative responses:
     No, just to be more careful   1     No   10
     Nothing I can think of now   1     Not really   2
     Not yet   9     Not much   10
     No, I repeat the same mistakes   1
   
   

4. How do you use the statistics (for compound sentences, passive voice, etc.) you get when you run WRITER'S WORKBENCH? Do you, for example, work to make corrections until all the percentages are within acceptable ranges?

   Simple responses:   yes   12   no   46
   

   Qualified responses (partial list):

   I read through my paper again to determine if I need to make changes in sentence structure.

   Yes, sometimes. But at times it is appropriate to do things differently.

   I take them with a grain of salt.

   No I write in second-grade sentences for a reason sometimes.

   If a large difference, I make changes, if not--No.

   Yes, except if I don't think the corrections are correct.

   No. Only when I'm satisfied with the paper as a whole. I use WWB only as a guide.

   I pay attention to the stats given, but . . . I try to make changes that work well within my paper.

   I don't really pay attention to those percentages. They are interesting, but the only one I would really think about would be the percentage of passive voice.

   Yes, usually. Only once did I totally disregard the Kincaid readability. It said it was too high (long words) but the essay was as letter to a senator.

   The percentages make me examine the paper, but I rarely change things. In particular, the ratio of complex to compound sentences--I usually have a high percentage of complex sentences. Changing these however would change the meaning of the paper.