Information Sheet #150 


January 27, 2006
 

Regression, Blunders, Errors, Mistakes, and Solecisms

[Whenever faculty converse about student writing, it’s not unusual for someone to express frustration concerning the frequency of editing errors.  It would require more than one Word Shop to explain why even the simplest editing errors (such as confusing its and it’s) are so resistant to correction, but one productive theory involves the role of “regression.”  In response to a query about the “regression” as the appropriate term for the often unexpected recurrence of errors in upper-class student writing, the University of Chicago’s Joe Williams, probably the field’s leading authority on this issue, offered the following response on a listserv for writing program administrators.   ~Bob Marrs]

I think "regression" captures a lot of it. The behavior you see in students writing in new contexts on new subjects or new audiences has one or more of these characteristics:

1. A heightened sense of "concreteness." Give them a schema to follow and they'll follow it relentlessly without seeing the nuances and variables. This includes seizing on the most obvious features of texts in their field and imitating them. . . .

2. A lot of "out-loud-cognition." They tend to represent their thinking overtly in their writing:  "First I did this, then I thought that, but . . . "

3. "Discovery" organization.  This is related to (2). They produce a narrative of their work.

4. A real breakdown in style and sometimes even grammar and punctuation.

Most theories of development postulate a beginning stage that is concrete.  The out loud cognition and discovery organization is just a manifestation of that concreteness.  The breakdown in style and grammar is close to real regression.


An excellent discussion of the regression phenomenon is in Richard Haswell’s Gaining Ground in College Writing: Tales of Development and Interpretation.  Here are some relevant passages from his chapter “Solecisms: Mistakes and Errors.”  Haswell uses the term “solecism” to identify an ungrammatical or substandard usage as a breach of etiquette.


Of all the signs of error in transformative learning, the most telling is regression.  It is the phase that is graphed in the classic “U-shaped growth curve,” where for a while later skill appears to perform worse than earlier.  The paradox, much studied, has produced some of the more fascinating phenomena of normal human learning.  Ten-year-olds can recognize upside-down photographs of acquaintances 90 percent of the time; twenty-year-old college students only 60 percent. . . . Language acquisition is attended by regression nearly from the beginning.  “I made that fall down” occurs earlier than “I fell that down.”  Three-year-olds make more mistakes than do two-year-olds in acting out the sentence “It’s the horse that the cow kisses.”  In all these cases, worse results are momentarily produced in a self-contradictory way by better but newer knowledge frames. . . . [The] three-year-olds misinterpret the direction of the kiss because they have acquired a syntactic forestructure, the actor-action-object sequence canonical in English.

Regression continues as long as language competence continues to grow, and transformative types can be found in college-age writing everywhere.  Rules are extended beyond their customary range (“overregularization”), as when students put quotation marks around block quotes.  Old frames are applied to new situations (“transference”), as when they spell, and probably read, “beatify” as “beautify.”  Previously learned coding systems interact with new ones, as when they write incorrectly what they then correctly read, or scribe mistakes that they later cannot find by proofreading but can correct once found.  Isolated rules are learned in a new system but not their internal relationship, as when they mix registers.  In a quid-pro-quo or interactive way, some subskills lag so that others can advance (“embeddedness”), as when they mispunctuate a novel syntactic form such as appositives.  During points of stress or indecision, retreat is beaten to previously acquired strategies (“fallback”), as when they write, “the future of work, marriage, and to be a good parent.”  Relics of old systems hang on, getting in the way of new (“fossilization”), as when they place an oral-language way to form possessives on top of a more formal way: “the appearance of a friend’s.”

 

Haswell’s own research demonstrates that error counts in student writing remain stubbornly stable through the college years.  His study looked at eight common solecisms: 

      • misformation of possessives

      • faulty predication

      • faulty pronoun reference

      • misformed syntactic parallelism

      • mispunctuation of final free modification

      • sentence fragments

      • comma splices

      • misspellings 

In compositions submitted through four years of college, students averaged 36-38 solecisms per thousand words.  While the total error counts remained level, students consistently improved in some ways while also revealing various regressions.  Here are a few insights from Haswell’s investigation:

  Misformation of possessives. The rate of possessives in student writing remains consistent, but misformation of possessives substantially increases as students progress through college.  In the Haswell study, from one-third to one-half of possessives were misformed by sophomores and juniors.  One reason for this increase is because older students are more likely to form possessives involving abstract, pluralized, or otherwise generalized nouns.

  Faulty predication.  The rate of predication errors remains level through the college years, which may be interpreted as a positive sign since upper-class students attempt to construct more complex sentences.  For example, in Haswell’s study, sophomores and juniors are twice as likely to attempt appositives in their essays as do freshmen. 

  Faulty pronoun reference. In Haswell’s study, pronoun-antecedent agreement deteriorated in sophomore writing but began to improve for juniors.  The regression is probably cause by an increased use of nominal modifiers, appositives, and final free modifiers–all of which double in frequency from freshman to junior year.

  Mispunctuation of final free modification.  The frequency of problems punctuating final free modifiers doubled between the freshman and sophomore year; however, sophomores attempted using final free modifiers twice as often. The rate of usage then levels off and the punctuation gradually improves into post-graduate years.  This is a particularly clear example of student writers’ developmental priorities: experimentation with a form precedes correctness of form.

  Sentence fragments and comma splices.  With undergraduate writers, fragment counts decline while comma splices become more common. Both trends are probably tied to upper-class students writing longer sentences. Haswell reports a study by Donna Kagan analyzing run-on sentences in college writing.  She found that the dominant examples were instances when the student writer would join together a long, complex clause with a short and simple clause at the end.  For example: “Because he had lost the money he did not buy a gift he arrived empty-handed.”  Such long-short combinations reflect the student writer’s development of a healthy, stylistically effective flow to the language.  Learning to punctuate the sentence comes with further practice.  Haswell warns that a simplistic red-inking of such examples may actually interfere with the natural evolution of these late-developing forms.

  Misspellings. The error counts for misspelled words remain level through four years of college writing; however, upper-class students double the freshman rate of “unusual words”–and thus the number of different words spelled correctly does increase.  Paradoxically, the misspelling of simple words increases in college; in fact, such errors continue to climb in the workplace.  This mis-scribing of common words (such as writing “all” for “are”) is probably the result of a greater concentration on content, a more rapid rate of production, and a decreased attention on minor issues that the writer assumes can be automatically handled without thought.

With this evidence in mind, Haswell suggests that college faculty consider the following implications as they reflect on their own teaching and how they deal with errors in student writing:

• When responding to student writers, instructors should distinguish between mistakes and errors.  Mistakes are due to performance factors or chance circumstances, caused by haste or distractions, memory faults, slips of the pen, lack of attention, good faith applications of bad teacher advice.  Errors are due to what can be described as “transitional competence.”  They are generated by developing writers in the process of restructuring their writing skills.  Errors are not the product of incompetence but come from transitional knowledge not yet perfected.  While mistakes are unsystematic, errors are systematic.  Teachers should not squelch errors (after all, they signify growth) but help students understand the nature and pattern of their errors.  To clarify this mistake/error distinction, it might help to consider this sequence of sentences from one paragraph in a recent first-year student’s FYS portfolio:

"Plato believed that the only real things, were things. . . ."

"He believed that the objects that make up the physical world, are imperfect imitations. . . ."

"Plato believed that every dog, was not really read. . . ."

What's intriguing about the punctuation errors in these three sentences is their recurrent pattern: in each sentence, the unneeded comma is preceded by a "that" construction, which is probably causing the student to feel the need for a comma–like inserting, a yellow warning light to announce the imminent arrival of a verb.  As we weigh our options for responding to these comma errors, we might keep in mind that we are likely witnessing a student writer’s unconscious experimentation with trying to communicate more complex ideas through more complex sentence structures.

• [In the following passage, Haswell warns that an obsession with solecisms can overwhelm a reading of a student paper.]  In the very act of marking, teachers usually see misformation first, thought second. . . . Some of the priority assigned to surface errors may owe to the fact that they are just that, on the surface, hence seen first and easily.  But the effect on a teacher’s appreciation of other rhetorical matters can be powerful.  Measures of error–usually misspellings–usually end up among the top three predictors of quality judgments by teachers of student writing, along with vocabulary and essay length.  Bennett A. Raforth and Donald L. Rubin compared the way teachers reacted to pairs of essays, one version mechanically perfect and the other identical except for fourteen surface mistakes.  On a general impression scale, the flawed essays dropped 40 percent from the unflawed ones.  Worse, they dropped 23 percent on an assessment of ideas and 17 percent on an assessment of organization, although neither ideas nor organization suffered any change.

Behind this large unintentional primacy assigned to error probably lies a pedagogical tale of almost gossip simplicity. Removal of blunders will better writing: SOLECISM REDUCTION IMPROVEMENT.  Solecisms take on a causal priority.  The writer errs, the errors are corrected, the result is a better essay and maybe a better writer. . . . Students too assume the causation, perhaps absorbing the notion from their teachers.  Just as teachers mark “weaknesses” first, students then often revise them first, and sometimes only them.  Sometimes students simply equate instruction in writing with emendation.  In Robert Stiff’s experiment manipulating teacher commentary on essays, the students who received fewer comments complained that they had not been given “full correction,” although they advanced during the course as well as did the other students. 

















This website created and maintained by the Coe Writing Center. Copyright 2001.
Email Dr. Bob Marrs with any questions, comments or suggestions.