Sherry Burgus Little
Anyone who uses a computer for composing knows what a marvelous tool it can be. My students write, in the computer logs that they keep, the common reactions of learning to compose on the computer: "It's fun, challenging, like a puzzle. It's better than a typewriter, faster, easier to edit." They report that they write more, revise and edit more. They also note that they type faster because they are more relaxed--no permanent errors--and they produce more accurate material. The miracles of word processing allow whole sentences, paragraphs, or even pages to be moved with the stroke of one key. Single-keystroke entry frees students from the drudgery of retyping the whole paper because of revisions and editing changes. Generally, students quickly become converts; as one student admitted, "I'm hooked on computers!" Using the computer for word processing, as a sophisticated electronic typewriter that performs marvelous feats, is not the only profitable use of a computer, however.
p. 107
Programs now exist that talk back to the user, acting as an audience, providing feedback on the writer's prose. This audience, though an imperfect or, perhaps better, an incomplete one, provides valuable feedback to students and also provides a vehicle for emphasizing the dangers of uncritically accepting readability information from computerized text programs as measurements of comprehensibility. Several programs provide feedback to writers. Some programs like GRAMMATIK and PUNCTUATION + STYLE for the IBM PC analyze prose for surface problems, mostly punctuation and wordy or pompous phrases. Others check only spelling, like RANDOM HOUSE PROOFREADER andCORRECTSTAR.
The programs I am mostly interested in analyze text for certain stylistic features, those incorporating some type of readability formula as a basis for analyzing and providing feedback. I am interested in this type of analysis for two reasons. First, such programs do provide valuable, though limited, feedback. And second, because such programs are increasingly being used in industry (and misused, I might add), students, especially technical writing students, profit greatly in having experience working with the feedback such programs provide. I would like to elaborate on these two points through example.
First, some background information. Besides gaining word-processing experience, my students also use HOMER, a text-analysis program written by Richard Lanham and Michael Cohen, UCLA, available from Charles Scribner's Sons. Although not as elaborate as other text analysis programs, such as WRITER'S WORKBENCH or CAMELOT, HOMER does provide feedback to the students on
p. 108
their prose, and it can be used on Apples instead of on more complicated mainframe computers as the more sophisticated, but more intimidating WRITER'S WORKBENCH (although an adapted version more suitable for students is now available in RIGHTWRITER). HOMER operates on Apple USCD Pascal, which requires two disk drives, one for HOMER (drive 2) and one for APPLEl:PASCAL l.l. Students enter their text with APPLE:l in order for HOMER to analyze their text. Although HOMER is a very friendly program, as I will discuss, APPLE:l is not. To simplify its use, I have written short instructions, covering only those instructions needed to complete individual assignments.
HOMER analyzes the use of passive voice, nominalizations, abstract words, and prepositional phrases. It also measures sentence length and gives statistics and comments about these statistics as well as short tutorials on these stylistic elements. Students can use this feedback in revising their prose. The program allows the instructor to change many features of HOMER's analysis, perhaps emphasizing passive voice if that is desired, or changing HOMER's comments or even eliminating them.
The analysis of the text has two formats. The verbal map, shown in Figure 1, identifies all prepositions, to-be verbs, nominalizations or shun words, and abstract or woolly words in a column next to the text. The other format, shown in Figure 2, isolates these words at the left margin, creating the laundry list look of Lanham's successful videotape Revising Prose. After the analysis, come the statistics as shown in Figures 3 and 4. First the total number of
p. 109
Prepositions = "P" "To be" verbs = "T" "Shun" words = "S" "Woolly" words = "W" Here's your map--GOOD LUCK, Sherry!!! Anyone who uses a computer for composing -----P- knows what a marvelous tool it can be. -------T My students write in the computer logs they ---P---- keep the common reactions of learning to ---SP-P compose on the computer: "It's fun, -P--T- challenging, like a puzzle. ---- It's better than a typewriter, faster, easier T------ to edit." P- They report that they write more, revise and -------- edit more. -- They also note that they type faster because -------- they are more relaxed--no permanent errors--and -T------ they produce more accurate material. ----- The miracles of word processing allow whole --P---- sentences, paragraphs, or even pages to be -----PT moved with the stroke of one key. -P--P-- Single keystroke entry frees them from the -----P- drudgery of retyping the whole paper because of -P----P revisions and editing changes. S--- Generally, students quickly become converts as ------ one student admitted, "I'm hooked on computers!" ---T-P- Figure 1. Verbal Mapp. 110 Here comes your text, Sherry: Anyone who uses a computer for composing knows what a marvelous tool it can be. My students write in the computer logs they keep the common reactions of learning to compose on the computer: "It's fun, challenging, like a puzzle. "It's better than a typewriter, faster, easier to edit." They report that they write more, revise and edit more. They also note that they type faster because they are more relaxed--no permanent errors--and they produce more accurate material. The miracles of word processing allow whole sentences, paragraphs, or even pages to be moved with the stroke of one key. Single keystroke entry frees them from the drudgery retyping the whole paper because of revisions and editing changes. Generally, students quickly become converts as one student admitted, "I'm hooked on computers!" Figure 2. Words Isolated at Left Margin
p. 111
>>>>>>>>>>>>>>>STATISTICS<<<<<<<<<<<<<<< Number of words: 133 Number of sentences: 8 Average words per sentence: 16.63 Number of prepositions: 14 Average prepositions per sentence: 1.75 Word to preposition ratio: 9.50 6 sentences contained prepositions Number of "to-be" verbs: 6 Average "to-be" verbs per sentence: 0.75 Word to "to-be" verb ratio: 22.17 6 sentences contained "to-be" verbs Figure 3. Statistics
p. 112
Number of "shun" words: 2 Average "shun" words per sentence: 0.25 Word to "shun" word ratio: 66.50 2 sentences contained "shun" words I'm glad we agree--I don't like too many "SHUN" words either!! Number of "woolly" words: 0 Average "woolly" words per sentence: 0.00 Word to "woolly" word ratio: 133.00 O sentences contained "woolly" words I'm glad you haven't pulled the "WOOLLY" words over my eyes. Figure 4. Statistics: Shun Words and Woolly Words
p. 113
words, then the number of sentences, and then the average words per sentence are recorded. Similar statistics are provided for each of the four stylistic elements identified in the analysis. The total number of prepositions are given first, then the average prepositions per sentence, then the word-to-preposition ratio, and finally the total sentences using prepositions. Usually, based on a pre-set ratio that can be controlled by the instructors, HOMER will comment on these statistics and in some cases, suggest possible revisions (Figure 5). Such comments as "Check your `SHUN' words--you might make some into verbs!!" not only directs the student's attention to a potential problem, but also suggests a solution to that problem. HOMER praises as well, "saying" things like, "I see you've shorn your text of its `woolly' words--I appreciate that" or "Good work--few `shun' words often mean clearer writing!" or "You keep your PREPOSITIONS under control--How admirable!" or "You don't overuse PREPOSITIONS-- I find that quite pleasing!" HOMER will also print out a bar graph, showing sentence length (Figure 6). Students have little problem seeing whether their sentences are mostly short, choppy sentences or "lengthy lingering sentences" (HOMER's words).
HOMER's comments, as you can see, are basically positive and human. Students like this feature. While enjoying the entertaining comments, they are gaining a personal and private evaluation of their prose. Though HOMER provides a limited evaluation, writing teachers of the future may, as Charles Smith and Kathleen Kiefer state, "rely on computers to help in the mind-numbing task of editing and evaluating
p. 114
>>>>>>>>>>>>>>>>>>>>>>STATISTICS<<<<<<<<<<<<<<<<<<<<<<
Number of words:
129
Number of sentences:
5
Average words per sentence:
25.80
TAKE CARE--long, tangled sentences can confuse your audience!!
>>>>>>>>>>>>>>>>>>>>>>STATISTICS<<<<<<<<<<<<<<<<<<<<<<
Number of words:
129
Number of sentences:
5
Average words per sentence:
25.80
Long, loose sentences slither like wounded snakes--OH MY!!
Number of "shun" words:
1
Average "shun" words per sentence:
0.20
Word to "shun" word ratio:
129.00
1 sentence contained "shun" words
Good work--few "SHUN" words often means clearer writing!
Figure 5. Samples of Homer's Comments
p.115
S. R. Little
=*=*=*=*=*=*=*=*=*=Sentence Length Graph*=*=*=*=*=*=*=*=*=*
+ 10 + 20 + 30 + 40 + 50 + 60
###############
#########################
#########
##########
#####################
#####################
###################
#############
Figure 6. Sentence-Length Graph
p. 116
student essays" freeing them "from editing so that they can focus on those matters of greater moment" (Smith and Kiefer, forthcoming). What's more, preliminary research studies by both the technical staff of Bell Laboratories, who developed WRITER'S WORKBENCH (Gingrich, 1983), and teachers of writing at Colorado State University, who used, tested, and adapted the program to their writing programs, suggest that students who use textual-analysis programs learn editing skills more thoroughly than those students who do not receive computer assistance, at least on those stylistic elements that the program identifies (Smith and Kiefer, 1983).
The majority of my students who used HOMER responded on questionnaires that they felt HOMER was valuable in helping them revise their prose, and most reported that they felt their writing and their writing processes had improved as a direct result of using computers in their writing and HOMER in their revising and editing stages. And from my viewpoint, such results were accomplished more effectively and with more enjoyment than if students were not using computers.
I base this observation on comments made by students in their computer logs and in classroom discussions. First was the obvious fun students had with HOMER's entertaining and human persona. As one student wrote in her log, "It's hard to argue with HOMER when he's so nice." It took some prodding to make some students realize that it's okay to argue with a machine, even when it's as nice as HOMER. A good number of students ignored HOMER's final message:
p. 117
You mustn't take my comments as gospel. I'm merely a machine. YOU must decide if you've written logically, organized effectively, and, above all, said something worth saying. I cannot. Only you or your reader can really judge your writin g's worth.
At first, some students had the impression that every flagged item was being criticized and needed revision, despite class discussion about the dangers of uncritically accepting all comments in the analysis. And here I want to emphasize that HOMER paves the way to stimulating discussion, centered on the student's own writing, about the dangers of computerized text analysis and uncritical application of readability formulas in general. HOMER's comments, besides being basically positive and human, are also deliberately nondogmatic. HOMER suggests that students check something because they might be able to improve it. Words like few, often, and might imply the need of human intervention and judgment. Students spend a good amount of time, usually in group workshops and readarounds, analyzing the analysis, an activity that sparked what may be the greatest benefit from using this program, or any text-analysis program.
The discussion is lively, specific, personal, and sometimes heated in these sessions. Instead of a ho-hum, noncommittal attitude, students, computer printouts in hand, ask indignantly, "What's wrong with using prepositions?!! or "I like the passive voice here! Why shouldn't I use it?" Interestingly enough, the writing teacher is not cast in the
p. 118
role of bad guy in these situations; instead, the teacher is now the resource person to be consulted as a result of The Other, Machine-generated evaluations and comments. Students question openly and look critically at these comments, discussing them freely with each other and the instructor. Most students reported that as a result of using HOMER and having these class sessions, they were more aware of these stylistic elements and their weaknesses. In addition to giving students actual experience in using text analysis, an activity used increasingly by industry to check style and readability, students more fully understood the importance of audience analysis in making the choices they eventually made in revising their prose. It was HOMER's inability to adapt to changing audiences that created some of the mistakes that students discussed so critically.
And this point is the crux of the matter: Though providing valuable feedback on certain aspects of writing, computerized text-analysis programs cannot think. Though human-like, they are not human. A company I know of in Southern California uses the reading grade-level of the document as determined by a computerized program of their own devising as a guide for effectiveness (a condition undoubtedly imposed by the documents they are writing and the to whom audiences they are writing). This company, fortunately, is aware of the fallacy of an uncritical application of this guide. Other industries are not so aware. Students, especially technical writing students, benefit from learning how best to use this electronic tool.
p. 119
Given this caveat, however, please do not misunderstand. I am not advocating the immediate destruction of text-analysis programs. Using the computer as audience provides students experience creating a fictive audience, that Other in their composing process. The immediate and private feedback from an Other helps them learn and apply the concept of writing reading-centered prose vs. writer-centered prose. It has added a valuable dimension to my classes and has brought the real world into my technical writing classes, providing students with experience that will aid them as they continue using computers in their careers.
p. 120
Gingrich, P.S. (July-August 1983). The UNIX Writer's Workbench Software: Results of a Field Study, The Bell System Technical Journal, 62(6), 1909-1921.
Smith, C.R. and Keifer, K.E. (forthcoming). Writer's Workbench:
Computers and Writing Instruction. Proceedings of the Future of Literacy Conference, University of Maryland, Baltimore County, April 1-3, 1982.
Smith, C.R. and Keifer, K.E. (1983). Using the Writer's Workbench programs at Colorado State University. In S.K. Burton and D.D. Short (Eds.), Sixth International Conference on Computers and the Humanities (pp. 672-684). Rockville, Maryland: Computer Science Press.