Writer-Based Prose: A Cognitive Basis for Problems in Writing

Author(s): Linda Flower

Source: College English, Vol. 41, No. 1, (Sep., 1979), pp. 19-37
Published by: National Council of Teachers of English
Stable URL: http://www.jstor.org/stable/376357
Accessed: 19/08/2008 21:11

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ncte.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org
LINDA FLOWER

Writer-Based Prose: A
Cognitive Basis for Problems in
Writing

IF WRITINGIS SIMPLYTHE ACT of "expressing what you think" or "saying what you
mean," why is writing often such a difficult thing to do? And why do papers that do
express what the writer meant (to his or her own satisfaction) often fail to communi-
cate the same meaning to a reader? Although we often equate writing with the
straightforward act of "saying what we mean," the mental struggles writers go
through and the misinterpretations readers still make suggest that we need a better
model of this process. Modern communication theory and practical experience
agree; writing prose that actually communicates what we mean to another person
demands more than a simple act of self-expression. What communication theory
does not tell us is how writers do it.
An alternative to the "think it/say it" model is to say that effective writers do not
simply expressthought but transformit in certain complex but describable ways for
the needs of a reader. Conversely, we may find that ineffective writers are indeed
merely "expressing"themselves by offering up an unretouched and underprocessed
version of their own thought. Writer-Based prose, the subject of this paper, is a
description of this undertransformedmode of verbal expression.
As both a style of writing and a style of thought, Writer-Based prose is natural
and adequate for a writer writing to himself or herself. However, it is the source of
some of the most common and pervasive problems in academic and professional
writing. The symptoms can range from a mere missing referent or an underde-
veloped idea to an unfocused and apparently pointless discussion. The symptoms
are diverse but the source can often be traced to the writer's underlying strategy for
composing and to his or her failure to transform private thought into a public,
reader-based expression.
In function, Writer-Based prose is a verbal expression written by a writer to him-
self and for himself. It is the record and the working of his own verbal thought. In
its structure,Writer-Based prose reflects the associative, narrativepath of the writer's

LindaFlower is directorof the Businessand ProfessionalCommunications Programat Carnegie-Mellon University.

She is workingundera grantfrom the NationalInstituteof Educationto studycognitiveprocesses
in writing. Shehas
publishedessaysin COLLEGE ENGLISH and Composition and Teaching, and she has just completeda book,
Problem-Solving Strategies in Writing.

COLLEGEENGLISH Vol. 41, No. 1 * September 1979

19
20 COLLEGEENGLISH

own confrontation with her subject. In its language, it reveals her use of privately
loaded terms and shifting but unexpressed contexts for her statements.
In contrast, Reader-Based prose is a deliberate attempt to communicate something
to a reader. To do that it creates a shared language and shared context between
writer and reader. It also offers the reader an issue-centered rhetorical structure
rather than a replay of the writer's discovery process. In its language and structure,
Reader-Based prose reflects the purposeof the writer's thought; Writer-Based prose
tends to reflect its process.Good writing, therefore, is often the cognitively demand-
ing transformation of the natural but private expressions of Writer-Based thought
into a structure and style adapted to a reader.
This analysis of Writer-Based prose style and the transformations that create
Reader-Based prose will explore two hypotheses:
1. Writer-Based prose represents a major and familiar mode of expression which
we all use from time to time. While no piece of writing is a pure example, Writer-
Based prose can be identified by features of structure, function, and style. Fur-
thermore, it shares many of these features with the modes of inner and egocentric
speech described by Vygotsky and Piaget. This paper will explore that relationship
and look at newer research in an effort to describe Writer-Based prose as a verbal
style which in turn reflects an underlying cognitive process.
2. Writer-Based prose is a workable concept which can help us teach writing. As
a way to intervene in the thinking process, it taps intuitive communication strategies
writers already have, but are not adequately using. As a teaching technique, the
notion of transforming one's own Writer-Based style has proved to be a powerful
idea with a built-in method. It helps writers attack this demanding cognitive task
with some of the thoroughness and confidence that comes from an increased and
self-conscious control of the process.
My plan for this paper is to explore Writer-Based prose from a number of
perspectives. Therefore, the next section, which considers the psychological theory
of egocentrism and inner speech, is followed by a case study of Writer-Based prose.
I will then pull these practical and theoretical issues together to define the critical
features of Writer-Based prose. The final section will look ahead to the implications
of this description of Writer-Based prose for writers and teachers.

INNER SPEECHAND EGOCENTRISM.In studying the developing thought of the

child, Jean Piaget and Lev Vygotsky both observed a mode of speech which seemed
to have little social or communicative function. Absorbed in play, children would
carry on spirited elliptical monologues which they seemed to assume others under-
stood, but which in fact made no concessions to the needs of the listener. According
to Piaget, in Vygotsky's synopsis, "In egocentric speech, the child talks only about
himself, takes no interest in his interlocutor, does not try to communicate, expects
no answers, and often does not even care whether anyone listens to him. It is similar
to a monologue in a play: The child is thinking aloud, keeping up a running accom-
paniment, as it were, to whatever he may be doing."1 In the seven-year olds Piaget

lLev Vygotsky, Thoughtand Language,ed. and trans. Eugenia Hanfmann and Gertrude Vakar (Cam-
bridge, Mass.: M.I.T. Press, 1962), p. 15.
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 21

studied, nearly fifty percent of their recorded talk was egocentric in nature.2 Ac-
cording to Piaget, the child's "non-communicative"or egocentric speech is a reflec-
tion, not of selfishness, but of the child's limited ability to "assume the point of view
of the listener: [the child] talks of himself, to himself, and by himself."3 In a sense,
the child's cognitive capacity has locked her in her own monologue.
When Vygotsky observed a similar phenomenon in children he called it "inner
speech" because he saw it as a forerunner of the private verbal thought adults carry
on. Furthermore, Vygotsky argued, this speech is not simply a by-product of play,
it is the tool children use to plan, organize, and control their activities. He put the
case quite strongly: "We have seen that egocentric speech is not suspended in a void
but is directly related to the child's practical dealings with the real world . . . it
enters as a constituent part into the process of rational activity" (Thoughtand Lan-
guage, p. 22).
The egocentric talk of the child and the mental, inner speech of the adult share three
important features in common. First, they are highly elliptical. In talkingto oneself the
psychological subject of discourse (the old information to which we are adding new
predicates) is always known. Therefore, explicit subjects and referents disappear. Five
people straining to glimpse the bus need only say, "Coming!" Secondly, inner speech
frequently deals in the sense of words, not their more specific or limited public
meanings. Words become "saturatedwith sense" in much the way a key word in a poem
can come to represent its entire, complex web of meaning. But unlike the word in the
poem, the accrued sense of the word in inner speech may be quite personal, even
idiosyncratic; it is, as Vygotsky writes, "the sum of all the psychological events aroused
in our consciousness by the word" (Thoughtand Language,p. 146).
Finally, a third feature of egocentric/inner speech is the absence of logical and causal
relations. In experiments with children's use of logical-causal connectives such as
because,therefore,and although,Piaget found that children have difficulty managing such
relationships and in spontaneous speech will substitute a non-logical, non-causal con-
nective such as then. Piaget described this strategy for relating things asjuxtaposition:
"the cognitive tendency simply to link (juxtapose) one thought element to another,
rather than to tie them together by some causal or logical relation."4
One way to diagnose this problem with sophisticated relationships is to say, as
Vygotsky did, that young children often think in complexesinstead of concepts.5
When people think in complexes they unite objects into families that really do share
common bonds, but the bonds are concrete and factual rather than abstract or logi-
cal. For example, the notion of "college student" would be a complex if it were
based, for the thinker, on facts such as college students live in dorms, go to classes,
and do homework.

2Jean Piaget, TheLanguageand Thoughtof the Child, trans. Majorie Gabin (New York: Harcourt, Brace,
1932), p. 49.
3Herbert Ginsberg and Sylvia Opper, Piaget'sTheoryof IntellectualDevelopment(Englewood Cliffs, N.J.:
Prentice-Hall, 1969), p. 89.
4John Flavell, TheDevelopmental PsychologyofJean Piaget (New York: D. Van Nostrand, 1963), p. 275.
For these studies see the last chapter of Piaget's Languageand Thoughtof the Child and Judgmentand
Reasoningin the Child, trans. M. Warden (New York: Harcourt, Brace, 1926).
5Thoughtand Language,p. 75. See also the paper by Gary Woditsch which places this question in the
context of curriculum design, "Developing Generic Skills: A Model for a Competency-Based General
Education," available from CUE Center, Bowling Green State University.
22 COLLEGEENGLISH

Complexes are very functional formations, and it may be that many people do
most of their day-to-day thinking without feeling the need to form more demanding
complex concepts. Complexescollect related objects; concepts,however, must express
abstract, logical relations. And it is just this sort of abstract, synthetic thinking that
writing typically demands. In a child's early years the ability to form complex con-
cepts may depend mostly on developing cognitive capacity. In adults this ability
appears also to be a skill developed by training and a tendency fostered by one's
background and intellectual experience. But whatever its source, the ability to move
from the complexes of egocentric speech to the more formal relations of conceptual
thought is critical to most expository writing.
Piaget and Vygotsky disagreed on the source, exact function, and teleology of
egocentric speech, but they did agree on the features of this distinctive phenome-
non, which they felt revealed the underlying logic of the child's thought. For our
case, that may be enough. The hypothesis on which this paper rests is not a de-
velopmental one. Egocentric speech, or rather its adult written analogue, Writer-
Based prose, is not necessarily a stage through which a writer must develop or one
at which some writers are arrested. But for adults it does represent an available
mode of expression on which to fall back. If Vygotsky is right, it may even be
closely related to normal verbal thought. It is clearly a natural, less cognitively de-
manding mode of thought and one which explains why people, who can express
themselves in complex and highly intelligible modes, are often obscure. Egocentric
expression happens to the best of us; it comes naturally.
The work of Piaget and Vygotsky, then, suggests a source for the cognitive pat-
terns that underlie Writer-Based prose, and it points to some of the major features
such a prose style would possess. Let us now turn to a more detailed analysis of
such writing as a verbal style inadequately suited for the needs of the reader.

WRITER-BASEDPROSE:A CASE STUDYOF A TRANSFORMATION. As an introduction

to the main features of Writer-Based prose and its transformations, let us look at two
drafts of a progress report written by students in an organizational psychology class.
Working as consulting analysts to a local organization, the writers needed to show
progress to their instructor and to present an analysis with causes and conclusions to
the client. Both readers-academic and professional-were less concerned with what
the students did or saw than with why they did it and what they made of their
observations.
To gauge the Reader-Based effectiveness of this report, skim quickly over Draft 1
and imagine the response of the instructor of the course, who needed to answer
these questions: As analysts, what assumptions and decisions did my students
make? Why did they make them? At what stage in the project are they now? Or,
play the role of the client-reader who wants to know: How did they define my
problem, and what did they conclude? As either reader, can you quickly extract the
information the report should be giving you? Next, try the same test on Draft 2.

1:
DRAFT
GROUP REPORT

(1) Work began on our project with the initial group decision to evaluate the Oskaloosa
Brewing Company. Oskaloosa Brewing Company is a regionally located brewery man-
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 23

ufacturing several different types of beer, notably River City and Brough Cream Ale.
This beer is marketed under various names in Pennsylvania and other neighboring states.
As a group, we decided to analyze this organization because two of our group members
had had frequent customer contact with the sales department. Also, we were aware that
Oskaloosa Brewing had been losing money for the past five years and we felt we might
be able to find some obvious problems in their organizational structure.

(2) Our first meeting, held February 17th, was with the head of the sales department,
Jim Tucker. Generally, he gave us an outline of the organization from president to
worker, and discussed the various departments that we might ultimately decide to
analyze. The two that seemed the most promising and most applicable to the project
were the sales and production departments. After a few group meetings and discussions
with the personnel manager, Susan Harris, and our advisor Professor Charns, we felt it
best suited our needs and the Oskaloosa Brewing's to evaluate their bottling department.

(3) During the next week we had a discussion with the superintendent of production,
Henry Holt, and made plans for interviewing the supervisors and line workers. Also, we
had a tour of the bottling department which gave us a first hand look into the production
process. Before beginning our interviewing, our group met several times to formulate
appropriate questions to use in interviewing, for both the supervisors and workers. We
also had a meeting with Professor Charns to discuss this matter.

(3a) The next step was the actual interviewing process. During the weeks of March
14-18 and March 21-25, our group met several times at Oskaloosa Brewing and inter-
viewed ten supervisors and twelve workers. Finally during this past week, we have had
several group meetings to discuss our findings and the potential problem areas within the
bottling department. Also, we have spent time organizing the writing of our progress
report.

(4) The bottling and packaging division is located in a separate building, adjacent to
the brewery, where the beer is actually manufactured. From the brewery the beer is
piped into one of five lines (four bottling lines and one canning line), in the bottling
house where the bottles are filled, crowned, pasteurized, labeled, packaged in cases, and
either shipped out or stored in the warehouse. The head of this operation, and others, is
production manager, Phil Smith. Next in line under him in direct control of the bottling
house is the superintendent of bottling and packaging, Henry Holt. In addition, there
are a total of ten supervisors who report directly to Henry Holt and who oversee the
daily operations and coordinate and direct the twenty to thirty union workers who oper-
ate the lines.

(5) During production, each supervisor fills out a data sheet to explain what was
actually produced during each hour. This form also includes the exact time when a
breakdown occurred, what it was caused by, and when production was resumed. Some
supervisors' positions are production staff oriented. One takes care of supplying the raw
material (bottles, caps, labels, and boxes) for production. Another is responsible for the
union workers assignment each day.
These workers are not all permanently assigned to a production line position. Men
called "floaters"are used filling in for a sick worker, or helping out after a breakdown.

(6) The union employees are generally older than 35, some in their late fifties. Most
have been with the company many years and are accustomed to having more workers per
a slower moving line. They are resentful to what they declare "unnecessary" production
changes. Oskaloosa Brewery also employs mechanics who normally work on the produc-
tion line, and assume a mechanics job only when a breakdown occurs. Most of these men
are not skilled.
24 COLLEGE ENGLISH

DRAFT 2:
MEMORANDUM
TO: Professor Martin Charns

FROM: Nancy Lowenberg, Todd Scott, Rosemary Nisson, Larry Vollen

DATE: March 31, 1977

RE: ProgressReport:The OskaloosaBrewingCompany

WHY OSKALOOSABREWING?

(1) Oskaloosa Brewing Company is a regionally located brewery manufacturing sev-

eral different types of beer, notably River City and Brough Cream Ale. As a group, we
decided to analyze this organization because two of our group members have frequent
contact with the sales department. Also, we were aware that Oskaloosa Brewing had
been losing money for the past five years and we felt we might be able to find some
obvious problems in their organizational structure.

INITIALSTEPS: WHERETO CONCENTRATE?

(2) Through several interviews with top management and group discussion, we felt it
best suited our needs, and Oskaloosa Brewing's, to evaluate the production department.
Our first meeting, held February 17, was with the head of the sales department, Jim
Tucker. He gave us an outline of the organization and described the two major depart-
ments, sales and production. He indicated that there were more obvious problems in the
production department, a belief also implied by Susan Harris, personnel manager.

NEXT STEP
(3) The next step involved a familiarization of the plant and its employees. First, we
toured the plant to gain an understanding of the brewing and bottling process. Next,
during the weeks of March 14-18 and March 21-25, we interviewed ten supervisors and
twelve workers. Finally, during the past week we had group meetings to exchange in-
formation and discuss potential problems.

THE PRODUCTION
PROCESS
(4) Knowledge of the actual production process is imperative in understanding the
effects of various problems on efficient production; therefore, we have included a brief
summary of this process.

The bottling and packaging division is located in a separate building, adjacent to the
brewery, where the beer is actually manufactured. From the brewery the beer is piped
into one of five lines (four bottling lines and one canning line) in the bottling house where
the bottles are filled, crowned, pasteurized, labeled, packaged in cases, and either
shipped out or stored in the warehouse.

PEOPLEBEHINDTHE PROCESS
(5) The head of this operation is production manager, Phil Smith. Next in line under
him in direct control of the bottling house is the superintendent of bottling and packag-
ing, Henry Holt. He has authority over ten supervisors who each have two major re-
sponsibilities: (1) to fill out production data sheets that show the amount produced/hour,
and information about any breakdowns-time, cause, etc., and (2) to oversee the daily
operations and coordinate and direct the twenty to thirty union workers who operate the
lines. These workers are not all permanently assigned to a production line position. Men
called "floaters"are used to fill in for a sick worker or to help out after a breakdown.
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 25

(6) The union employees are highly diversifiedgroup in both age and experience.
They are generallyolder than 35, some in their late fifties. Most have been with the
company many years and are accustomed to having more workers per a slower moving
line. They are resentful to what they feel are unnecessary production changes. Oskaloosa
Brewing also employs mechanics who normally work on the production line, and assume
a mechanics job only when a breakdown occurs. Most of these men are not skilled.

PROBLEMS

Throughextensiveinterviewswith supervisorsand union employees,we have recog-

nized four apparentproblemswithin the bottle house operations.First, the employees'
goals do not matchthose of the company.This is especiallyapparentin the union em-
ployeeswhose loyalty lies with the unioninsteadof the company.This attitudeis well-
foundedas the unionensuresthem of job securityand benefits. ...

In its tedious misdirection, Draft I is typical of Writer-Based prose in student

papers and professional reports. The reader is forced to do most of the thinking,
sorting the wheat from the chaff and drawing ideas out of details. And yet, although
this presentation fails to fulfill our needs, it does have an inner logic of its own. The
logic which organizes Writer-Basedprose often rests on three principles: its underly-
ing focus is egocentric, and it uses either a narrative framework or a survey form to
order ideas.
The narrative framework of this discussion is established by the opening an-
nouncement: "Work began. . ." In paragraphs 1-3 facts and ideas are presented in
terms of when they were discovered, rather than in terms of their implications or
logical connections. The writers recount what happened when; the reader, on the
other hand, asks, "Why?" and "So what?" Whether he or she likes it or not the
reader is in for a blow-by-blow account of the writers' discovery process.
Although a rudimentary chronology is reasonable for a progress report, a narra-
tive framework is often a substitute for analytic thinking. By burying ideas within
the events that precipitated them, a narrative obscures the more important logical
and hierarchical relations between ideas. Of course, such a narrative could read like
an intellectual detective story, because, like other forms of drama, it creates interest
by withholding closure. Unfortunately, most academic and professional readers
seem unwilling to sit through these home movies of the writer's mind at work.
Narratives can also operate as a cognitive "frame"wrhichitself generates ideas.6 The
temporal pattern, once invoked, opens up a series of empty slots waiting to be filled
with the details of what happened next, even though those details may be irrelevant.
As the revision of Draft 2 shows, our writers' initial narrative framework led them
to generate a shaggy project story, instead of a streamlined logical analysis.
The second salient feature of this prose is its focus on the discovery process of the
writers: the "I did/I thought/I felt" focus. Of the fourteen sentences in the first three
paragraphs, ten are grammatically focused on the writers' thoughts and actions
rather than on issues: "Work began," "We decided," "Also we were aware . . . and
we felt. ..."

6The seminal paper on frames is M. Minsky's "A Framework for Representing Knowledge" in P.
Winston, ed., The Psychologyof ComputerVision (New York: McGraw Hill, 1973). For a more recent
discussion of how they work see B. Kuipers, "A Frame for Frames"in D. Bowbow and A. Collins, eds.,
Representationand Understanding:Studiesin CognitiveScience(New York: Academic Press, 1975), pp. 151-
184.
26 COLLEGE ENGLISH

In the fourth paragraphthe writers shift attention from their discovery process to
the facts discovered. In doing so they illustrate a third feature of Writer-Based
prose: its idea structure simply copies the structure of the perceived information. A
problem arises when the internal structure of the data is not already adapted to the
needs of the reader or the intentions of the writer. Paragraph five, for example,
appears to be a free-floating description of "What happens during production." Yet
the client-reader already knows this and the instructor probably does not care.
Lured by the fascination of facts, these writer-based writers recite a litany of per-
ceived information under the illusion they have produced a rhetorical structure. The
resulting structure could as well be a neat hierarchy as a list. The point is that the
writers' organizing principle is dictated by their information, not by their intention.
The second version of this report is not so much a "rewrite"(i.e., a new report) as
it is a transformationof the old one. The writers had to step back from their experi-
ence and information in order to turn facts into concepts. Pinpointing the telling
details was not enough: they had to articulate the meaning they saw in the data.
Secondly, the writers had to build a rhetorical structure which acknowledged the
function these ideas had for their reader. In the second version, the headings, topic
sentences, and even some of the subjects and verbs reflect a new functional structure
focused on Process, People, and Problems. The report offers a hierarchical organiza-
tion of the facts in which the hierarchy itself is based on issues both writer and
reader agree are important. I think it likely that such transformations frequently go
on in the early stages of the composing process for skilled writers. But for some
writers the under-transformed Writer-Based prose of Draft 1 is also the final prod-
uct and the starting point for our work as teachers.
In the remainder of this paper I will look at the features of Writer-Based prose
and the ways it functions for the writer. Clearly, we need to know about Reader-
Based prose in order to teach it. But it is also clear that writers already possess a
great deal of intuitive knowledge about writing for audiences when they are stimu-
lated to use it. As the case study shows, the concept of trying to transform Writer-
Based prose for a reader is by itself a powerful tool. It helps writers identify the
lineaments of a problem many can start to solve once they recognize it as a definable
problem.

WRITER-BASEDPROSE: FUNCTION,STRUCTURE,AND STYLE. While Writer-Based

prose may be inadequately structured for a reader, it does possess a logic and struc-
ture of its own. Furthermore, that structure serves some important functions for the
writer in his or her effort to think about a subject. It represents a practical strategy
for dealing with information. If we could see Writer-Based prose as a functional
system-not a set of random errors known only to English teachers-we would be
better able to teach writing as a part of any discipline that asks people to express
complex ideas.
According to Vygotsky, "the inner speech of the adult represents his 'thinking for
himself' rather than social adaptation [communication to others]: i.e., it has the
same function that egocentric speech has in the child" (Languageand Thought,p. 18).
It helps him solve problems. Vygotsky found that when a child who is trying to
draw encounters an obstacle (no pencils) or a problem (what shall I call it?), the
incidence of egocentric speech can double.
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 27

If we look at an analogous situation-an adult caught up in the complex mental

process of composing-we can see that much of the adult's output is not well
adapted for public consumption either. In studies of cognitive processes of writers as
they composed, J. R. Hayes and I observed much of the writer's verbal output to be
an attempt to manipulate stored information into some acceptable pattern of mean-
ing.7 To do that, the writer generates a variety of alternative relationships and trial
formulations of the information she has in mind. Many of these trial networks will
be discarded; most will be significantly altered through recombination and elabora-
tion during the composing process. In those cases in which the writer's first pass at
articulating knowledge was also the final draft-when she wrote it just as she
thought it-the result was often a series of semi-independent, juxtaposed networks,
each with its own focus.
Whether such expression occurs in an experimental protocol or a written draft, it
reflects the working of the writer's mind upon his material. Because dealing with
one's material is a formidable enough task in itself, a writer may allow himself to
ignore the additional problem of accommodating a reader. Writer-Based prose, then,
functions as a medium for thinking. It offers the writer the luxury of one less con-
straint. As we shall see, its typical structure and style are simply paths left by the
movement of the writer's mind.
The structureof Writer-Based prose reflects an economical strategy we have for
coping with information. Readers generally expect writers to produce complex
concepts-to collect data and details under larger guiding ideas and place those ideas
in an integrated network. But as both Vygotsky and Piaget observed, forming such
complex concepts is a demanding cognitive task; if no one minds, it is a lot easier to
just list the parts. Nor is it surprising that in children two of the hallmarks of
egocentric speech are the absence of expressed causal relations and the tendency to
express ideas without proof or development. Adults too avoid the task of building
complex concepts buttressed by development and proof, by structuring their infor-
mation in two distinctive ways: as a narrative of their own discovery process or as a
survey of the data before them.
As we saw in the Oskaloosa Brewing Case Study, a narrative structured around
one's own discovery process may seem the most natural way to write. For this
reason it can sometimes be the best way as well, if a writer is trying to express a
complex network of information but is not yet sure how all the parts are related. For
example, my notes show that early fragments of this paper started out with a narra-
tive, list-like structure focused on my own experience: "Writer-Based prose is a
working hypothesis because it works in the classroom. In fact, when I first started
teaching the concept. ... In fact, it was my students' intuitive recognition of the
difference between Writer-Based and Reader-Based style in their own thought and
writing. ... It was their ability to use even a sketchy version of the distinction to
transform their own writing that led me to pursue the idea more thoroughly."
The final version of this sketch (the paragraph numbered 2 on p. 20) keeps the
reference to teaching experience, but subordinates it to the more central issue of

7L. Flower and J. Hayes, "Plans That Guide the Composing Process," in Writing: The Nature, Develop-
ment and Teachingof Written Communication,C. Frederikson, M. Whiteman, and J. Dominic, eds. (Hills-
dale, N.J.: Lawrence Erlbaum, in press).
28 COLLEGE ENGLISH

why the concept works. This transformation illustrates how a writer's major propo-
sitions can, on first appearance, emerge embedded in a narrative of the events or
thoughts which spawned the proposition. In this example, the Writer-Based early
version recorded the raw material of observations; the final draft formed them into
concepts and conclusions.
This transformation process may take place regularly when a writer is trying to
express complicated information which is not yet fully conceptualized. Although
much of this mental work normally precedes actual writing, a first draft may simply
reflect the writer's current place in the process. When this happens rewriting and
editing are vital operations. Far from being a simple matter of correcting errors,
editing a first draft is often the act of transforming a narrative network of informa-
tion into a more fully hierarchical set of propositions.
A second source of pre-fabricated structure for writers is the internal structure of
the information itself. Writers use a surveystrategy to compose because it is a power-
ful procedure for retrieving and organizing information. Unfortunately, the original
organization of the data itself (e.g., the production process at Oskaloosa Brewing)
rarely fits the most effective plan for any given piece of focused analytical writing.
The prose that results from such a survey can, of course, take as many forms as
the data. It can range from a highly structured piece of discourse (the writer repeats
a textbook exposition) to an unfocused printout of the writer's memories and
thoughts on the subject. The form is merely a symptom, because the governing
force is the writer's mental strategy: namely, to compose by surveying the available
contents of memory without adapting them to a current purpose. The internal struc-
ture of the data dictates the rhetorical structure of the discourse, much as the pro-
ceedings of Congress organize the Congressional Record.As an information processor,
the writer is performing what computer scientists would call a "memory dump":
dutifully printing out memory in exactly the form in which it is stored.
A survey strategy offers the writer a useful way into the composing process in
two ways. First, it eliminates many of the constraints normally imposed by a speech
act, particularly the contract between reader and writer for mutually useful dis-
course. Secondly, a survey of one's own stored knowledge, marching along like a
textbook or flowing with the tide of association, is far easier to write than a fresh or
refocused conceptualization would be.
But clearly most of the advantages here accrue to the writer. One of the tacit
assumptions of the Writer-Based writer is that, once the relevant information is
presented, the reader will then do the work of abstracting the essential features,
building a conceptual hierarchy, and transforming the whole discussion into a func-
tional network of ideas.
Although Writer-Based prose often fails for readers and tends to preclude further
concept formation, it may be a useful road into the creative process for some writ-
ers. The structures which fail to work for readers may be powerful strategies for
retrieving information from memory and for exploring one's own knowledge net-
work. This is illustrated in Linde and Labov's well-known New York apartment
tour experiment.8 Interested in the strategies people use for retrieving information

8C. Linde and W. Labov, "Spatial Networks as a Site for the Study of Language and Thought,"
Language,51(1975), 924-939.
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 29

from memory and planning a discourse, Linde and Labov asked one hundred New
Yorkers to "tell me the layout of your apartment" as a part of a "sociological sur-
vey." Only 3% of the subjects responded with a map which gave an overview and
then filled in the details; for example, "I'd say it's laid out in a huge square pattern,
broken down into 4 units." The overwhelming majority (97%) all solved the prob-
lem by describing a tour: "You walk in the front door. There was a narrow hallway.
To the left, etc." Furthermore, they had a common set of rules for how to conduct
the tour (e.g., you don't "walk into" a small room with no outlet, such as a pantry;
you just say, "on the left is . . ."). Clearly the tour structure is so widely used
because it is a remarkably efficient strategy for recovering all of the relevant infor-
mation about one's apartment, yet without repeating any of it. For example, one
rule for "touring"is that when you dead-end after walking through two rooms, you
don't "walk" back but suddenly appear back in the hall.
For us, the revealing sidenote to this experiment is that although the tour strategy
was intuitively selected by the overwhelming majority of the speakers, the resulting
description was generally very difficult for the listener to follow and almost impos-
sible to reproduce. The tour strategy-like the narrative and textbook structure in
prose-is a masterful method for searching memory but a dud for communicating
that information to anyone else.
Finally, the style of Writer-Based prose also has its own logic. Its two main stylis-
tic features grow out of the private nature of interior monologue, that is, of writing
which is primarily a record or expression of the writer's flow of thought. The first
feature is that in such monologues the organization of sentences and paragraphs
reflects the shifting focus of the writer's attention. However, the psychological sub-
ject on which the writer is focused may not be reflected in the grammatical subject
of the sentence or made explicit in the discussion at all. Secondly, the writer may
depend on code words to carry his or her meaning. That is, the language may be
"saturated with sense" and able to evoke-for the writer-a complex but unex-
pressed context.
Writers of formal written discourse have two goals for style which we can usefully
distinguish from one another. One goal might be described as stylistic control, that
is, the ability to choose a more embedded or more elegant transformation from
variations which are roughly equivalent in meaning. The second goal is to create a
completely autonomous text, that is, a text that does not need context, gestures, or
audible effects to convey its meaning.
It is easy to see how the limits of short-term memory can affect a writer's stylistic
control. For an inexperienced writer, the complex transformation of a periodic
sentence-which would require remembering and relating a variety of elements and
optional structures such as this sentence contains-can be a difficult juggling act.
After all, the ability to form parallel constructions is not innate. Yet with practice
many of these skills can become more automatic and require less conscious atten-
tion.
The second goal of formal written discourse-the complete autonomy of the
text-leads to even more complex problems. According to David Olson the history
of written language has been the progressive creation of an instrument which could
convey complete and explicit meanings in a text. The history of writing is the trans-
formation of language from utterance to text-from oral meaning created within a
30 COLLEGE ENGLISH

shared context of a speaker and listener to a written meaning fully represented in an

autonomous text.9
In contrast to this goal of autonomy, Writer-Based prose is writing whose mean-
ing is still to an important degree in the writer's head. The culprit here is often the
unstated psychological subject. The work of the "remedial"student is a good place
to examine the phenomenon because it often reveals first thoughts more clearly than
the reworked prose of a more experienced writer who edits as he or she writes. In
the most imaginative, comprehensive and practical book to be written on the basic
writer, Mina Shaughnessy has studied the linguistic strategies which lie behind the
"errors"of many otherwise able young adults who have failed to master the written
code. As we might predict, the ambiguous referent is ubiquitous in basic writing:
he's, she's and it's are sprinkled through the prose without visible means of support.
It frequently works as a code word for the subject the writer had in mind but not on
the page.' As Professor Shaughnessy says, it "frequently becomes a free-floating
substitute for thoughts that the writer neglects to articulate and that the reader must
usually strain to reach if he can."'10
With all the jobs available, he will have to know more of it because thire is a great
demand for it.

For the writer of the above sentence, the pronoun was probably not ambiguous at
all; it no doubt referred to the psychological subject of his sentence. Psychologically,
the subject of an utterance is the old information, the object you are looking at, the
idea on which your attention has been focused. The predicate is the new informa-
tion you are adding. This means that the psychological subject and grammatical
subject of a sentence may not be the same at all. In our example, "college knowl-
edge" was the writer's psychological subject-the topic he had been thinking about.
The sentence itself is simply a psychological predicate. The pronoun it refers quite
reasonably to the unstated but obvious subject in the writer's mind.
The subject is even more likely to be missing when a sentence refers to the writer
herself or to "one" in her position. In the following examples, again from Errorsand
Expectations,the "unnecessary"subject is a person (like the writer) who has a chance
to go to college.
Even if a person graduated from high school who is going on to college to obtain a
specific position in his career [ ] should first know how much in demand his possible
future job really is.
[he]

If he doesn't because the U.S. Labor Department say's their wouldn't be enough jobs
opened, [ ] is a waste to society and a "cop-out" to humanity.
[he]

Unstated subjects can produce a variety of minor problems from ambiguous refer-
ents to amusing dangling modifiers (e.g., "driving around the mountain, a bear came
into view"). Although prescriptive stylists are quite hard on such "errors," they are

9David R. Olson, "From Utterance to Text: The Bias of Language in Speech and Writing," Harvard
EducationalReview, 47 (1977), 257-281.
'0Mina Shaughnessy, Errorsand Expectations(New York: Oxford University Press, 1977), p. 69.
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 31

often cleared up by context or common sense. However, the controlling but un-
stated presence of a psychological subject can lead to some stylistic "errors"that do
seriously disrupt communication. Sentence fragments are a good example.
One feature of an explicit, fully autonomous text is that the grammatical subject is
usually a precise entity, often a word. By contrast, the psychological subject to
which a writer wished to refer may be a complex event or entire network of infor-
mation. Here written language is often rather intransigent; it is hard to refer to an
entire clause or discussion unless one can produce a summary noun. Grammar, for
example, normally forces us to select a specific referent for a pronoun or modifier: it
wants referents and relations spelled out.ll This specificity is, of course, its strength
as a vehicle for precise reasoning and abstract thought. Errors arise when a writer
uses a one clause to announce his topic or psychological subject and a second clause
to record a psychological predicate, a response to that old information. For example:

The jobs that are listed in the paper,I feel you need a collegedegree.
The job that my motherhas, I know I could never be satisfiedwith it.

The preceding sentences are in error because they have failed to specify the gram-
matical relationship between their two elements. However, for anyone from the
Bronx, each statement would be perfectly effective because it fits a familiar formula.
It is an example of topicalization or Y-movement and fits a conventionalized, Yid-
dish influenced, intonation pattern much like the one in "Spinach-you can have
it!" The sentences depend heavily on certain conventions of oral speech, and insofar
as they invoke those patterns for the reader, they communicate effectively.12
However, most fragments do not succeed for the reader. And they fail, ironically
enough, for the same reason-they too invoke intonation patterns in the reader
which turn out to be misleading. The lack of punctuation gives off incorrect cues
about how to segment the sentence. Set off on an incorrect intonation pattern, the
thwarted reader must stop, reread and reinterpret the sentence. The following
examples are from Maxine Hairston's A Contemporary Rhetoric(Boston: Houghton
Mifflin, 1974):
The authoritiesdid not approveof theiracts. These acts beingconsidereddetrimentalto
society. (society, they . . .)

Young people need to be on their own. To show their parentsthat they are reliable.
(reliable, young people . . .)
(p. 322)

Fragments are easy to avoid; they require only minimal tinkering to correct. Then
why is the error so persistent? One possible reason is that for the writer the frag-
ment is a fresh predicate intended to modify the entire preceding psychological sub-

1""Pronounslike this, that, whichand it should not vaguely refer to an entire sentence or clause," and
"Make a pronoun refer clearly to one antecedent, not uncertainly to two." Floyd Watkins, et al., Practical
EnglishHandbook(Boston: Houghton Mifflin, 1974), p. 30.
121 am
greatly indebted here to Thomas Huckin for his insightful comments on style and to his work in
linguistics on how intonation patterns affect writers and readers.
32 COLLEGE ENGLISH

ject. The writer wants to carry out a verbal trick easily managed in speech. For the
reader, however, this minor grammatical oversight is significant. It sets up and vio-
lates both intonation patterns and strong structural expectations, such as those in the
last example where we expect a pause and a noun phrase to follow "reliable." The
fragment, which actually refers backward, is posing as an introductory clause.
The problem with fragments is that they are perfectly adequate for the writer. In
speech they may even be an effective way to express a new idea which is predicated
on the entire preceding unit thought. But in a written text, fragments are errors
because they do not take the needs of the reader into consideration. Looked at this
way, the "goodness" of a stylistic technique or grammaticalrule such as parallelism,
clear antecedents, or agreement is that it is geared to the habits, expectations, and
needs of the reader as well as to the demands of textual autonomy.
Vygotsky noticed how the language of children and inner speech was often "satu-
rated with sense." Similarly, the words a writer chooses can also operate as code
words, condensing a wealth of meaning in an apparently innocuous word. The fol-
lowing examples come from an exercise which asks writers to identify and transform
some of their own pieces of mental shorthand.
The students were asked to circle any code words or loaded expressions they
found in their first drafts of a summer internship application. That is, they tried to
identify those expressions that might convey only a general or vague meaning to a
reader, but which represented a large body of facts, experiences, or ideas for them.
They then treated this code word as one would any intuition-pushing it for its
buried connections and turning those into a communicable idea. The results are not
unlike those brilliant explications one often hears from students who tell you what
their paper really meant. This example also shows how much detailed and percep-
tive thought can be lying behind a vague and conventional word:
First Draft: "By having these two jobs, I was able to see the business in an entirely
(different perspective)" (Circle indicates a loaded expression marked by the writer.)

Second Draft with explanation of what she actually had in mind in using the circled
phrase: "By having these two jobs, I was able to see the true relationship and relative
importance of the various departments in the company. I could see their mutual de-
pendence and how an event in one part of the firm can have an important effect on
another."
The tendency to think in code words is a fact of life for the writer. Yet the
following example shows how much work can go into exploring our own saturated
language. Like any intuition, such language is only a source of potential meanings,
much as Aristotle's topics are places for finding potential arguments. In this ex-
tended example, the writer first explores her expression itself, laying out all the
thoughts which were loosely connected under its name. This process of pushing our
own language to give up its buried meanings forces us to make these loose connec-
tions explicit and, in the process, allows us to examine them critically. For the
writer in our example, pushing her own key words leads to an important set of new
ideas in the paper.
Excerpt from an application for the
National Institute of Health Internship Program
First Draft: "I want a career that will help other people while at the same time be
 Writer-Based Prose: A Cognitive Basisfor Problems in Writing 33

challenging scientifically. I had the opportunity to do a biochemical assay for a

neuropsychophamocologist at X- Clinic in Chicago. Besides learning the scientific
procedures and techniques that are used, I realized some of the(organizational, finan-
cial and people problems)which are encountered in research. This internship program
would let me pursue further my interest in research, while concurrently exposing me
to(relevant and diverse)areasof bioengineering."
Excerpt from Writer's Notes Working on the Circled Phrases
Brainstorm
How did research of Sleep Center tie into overall program of X- Clinic? Not everyone
within dept. knew what the others were doing, could not see overall picture of efforts.
Dr. 0.-dept. head-trained for lab yet did 38-40 hrs. paperwork. Couldn't set up
test assay in Sleep Center because needed equip. from biochem.
Difficulties in getting equipment
1. Politics between administrators
Photometer at U. of - even though Clinic had bought it.
2. Ordering time, not sufficient inventory, had to hunt through boxes for chemicals.
3. Had to schedule use by personal contact on borrowing equipment-done at time of
use and no previous planning.
No definite guidelines had been given to biochem. people as to what was "going on"
with assay. Partner who was supposed to learn assay was on vacation. Two people were
learning, one was on vac.
No money from state for equipment or research grants.
Departments stealing from each other.
Lobbying, politics, included.
My supervisor from India, felt prejudices on job. Couldn't advance, told me life story
and difficulties in obtaining jobs at Univ. Not interested in research at Clinic per se,
looking for better opportunities, studying for Vet boards.
Revision (additions in italics)
"As a biomedical researcher, I would fulfill my goal of a career that will help other
people while at the same time be challenging scientifically. I had exposure to research
while doing a biochemical assay for a neuropsychopharmocologist at X- Clinic in
Chicago. Besides learning the scientific procedures and techniques that are used, I
realized some of the organizational, financial and people problems which are encountered
in research. Theseproblemsincludeda lackof fundsand equipment,disagreements amongresearch
staff, and the extensiveamountsof time, paperworkand stepsrequiredfor testinga hypothesis
which
was only one verysmall but necessarypart of the overallproject.But besidesknowingsomeof the
frustrations,I alsoknowthat manymedicaladvancements, suchas the cardiacpacemaker,artificial
limbs and curesfor diseases,exist and benefitmany peoplebecauseof the effortsof researchers.
Therefore I would like to pursue my interest in research by participating in the NIH
Internship Program. The exposure to many diverseprojects,designedto betterunderstandand
improvethe body'sfunctioning,would helpme to decidewhich areasof biomedicalengineeringto
pursue."
We could sum up this analysis of style by noting two points. At times a Writer-
Based prose style is simply an interior monologue in which some necessary informa-
tion (such as intonation pattern or a psychological subject) is not expressed in the
text. The solution to the reader's problem is relatively trivial in that it involves
adding information that the writer already possesses. At other times, a style may be
Writer-Based because the writer is thinking in code words at the level of intuited but
34 COLLEGEENGLISH

unarticulated connections. Turning such saturated language into communicable

ideas can require the writer to bring the entire composing process into play.

FOR WRITERSAND TEACHERS. From an educational

IMPLICATIONS perspective,
Writer-Based prose is one of the "problems" composition courses are designed to
correct. It is a major cause of that notorious "breakdown" of communication be-
tween writer and reader. However, if we step back and look at it in the broader
context of cognitive operations involved, we see that it represents a major, functional
stage in the composing process and a powerful strategy well fitted to a part of the
job of writing.
In the best of all possible worlds, good writers strive for Reader-Based prose from
the very beginning: they retrieve and organize information within the framework of
a reader/writer contract. Their top goal or initial question is not, "What do I know
about physics, and in particularly the physics of wind resistance?" but, "What does
a model plane builder need to know?" Many times a writer can do this. For a
physics teacher this particular writing problem would be a trivial one. However, for
a person ten years out of Physics 101, simply retrieving any relevant information
would be a full-time processing job. The reader would simply have to wait. For the
inexperienced writer, trying to put complex thought into written language may also
be task enough. In that case, the reader is an extra constraint that must wait its turn.
A Reader-Based strategy which includes the reader in the entire thinking process is
clearly the best way to write, but it is not always possible. When it is very difficult
or impossible to write for a reader from the beginning, writing and then transform-
ing Writer-Based prose is a practical alternative which breaks this complex process
down into manageable parts. When transforming is a practiced skill, it enters natu-
rally into the pulse of the composing process as a writer's constant, steady effort to
test and adapt his or her thought to a reader's needs. Transforming Writer-Based
prose is, then, not only a necessary procedure for all writers at times, but a useful
place to start teaching intellectually significant writing skills.
In this final section I will try to account for the peculiar virtues of Writer-Based
prose and suggest ways that teachers of writing-in any field-can take advantage of
them. Seen in the context of memory retrieval, Writer-Basedthinking appears to be
a tapline to the rich sources of episodic memory. In the context of the composing
process, Writer-Based prose is a way to deal with the overload that writing often
imposes on short term memory. By teaching writers to use this transformation pro-
cess we can foster the peculiar strengths of writer-based thought and still alert writ-
ers to the next transformation that many may simply fail to attempt.
One way to account for why Writer-Based prose seems to "come naturally" to
most of us from time to time is to recognize its ties to our episodic as opposed to
semantic memory. As Tulving describes it, "episodic memory is a more or less
faithful record of a person's experiences." A statement drawn from episodic memory
"refers to a personal experience that is remembered in its temporal-spatialrelation to
other such experiences. The remembered episodes are . . . autobiographical events,
describable in terms of their perceptible dimensions or attributes."13

'3Edel Tulving, "Episodic and Semantic Memory," in Edel Tulving and Wayne Donaldson, eds.,
Organizationof Memory(New York: Academic Press, 1972), p. 387.
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 35

Semantic memory, by contrast, "is the memory necessary for the use of language.
It is a mental thesaurus, organized knowledge a person possesses about words and
other verbal symbols, their meaning and referents, about relations among them, and
about rules, formulas, and algorithms for the manipulation of these symbols, con-
cepts, and relations." Although we know that table salt is NaCl and that motivation
is a mental state, we probably do not remember learning the fact or the first time we
thought of that concept. In semantic memory facts and concepts stand as the nexus
for other words and symbols, but shorn of their temporal and autobiographical
roots. If we explored the notion of "writing" in the semantic memory of someone we
might produce a network such as this:
teachers
stone tablets
penmanship

WRITING ---- reading, writing, arithmetic

wronging
rhetoric
composition
texts
In an effort to retrieve what she or he knew about stone tablets, for example, this
same person might turn to episodic memory: "I once heard a lecture on the Rosetta
stone, over in Maynard Hall. The woman, as I recall, said that . . . and I remember
wondering if...."
Writers obviously use both kinds of memory. The problem only arises when they
confuse a fertile source of ideas in episodic memory with a final product. In fact, a
study by Russo and Wisher argues that we sometimes store our ideas or images (the
symbols of thought) with the mental operations we performed to produce these
symbols.14 Furthermore, it is easier to recall the symbols (that fleeting idea,
perhaps) when we bring back the original operation. In other words, our own think-
ing acts can serve as memory cues, and the easiest way to recover some item from
memory may be to reprocess it, to reconstruct the original thought process in which it
appeared. Much Writer-Based prose appears to be doing just this-reprocessing an
earlier thinking experience as a way to recover what one knows.
Writing is one of those activities that places an enormous burden on short-term or
working memory. As George Miller put it, "The most glaring result [of numerous
experiments] has been to highlight man's inadequacy as a communication channel.
As the amount of input information is increased, the amount of information that the
man transmits increases at first but then runs into a ceiling. . . . That ceiling is
always very low. Indeed, it is an act of charity to call man a channel at all. Com-
pared to telephone or television channels, man is better characterized as a
bottleneck."15

14J. Russo and R. Wisher, "Reprocessing as a Recognition Cue," Memoryand Cognition,4 (1976), 683-
689.
5George Miller, The Psychologyof Communication(New York: Basic Books, 1967), p. 48.
36 COLLEGE ENGLISH

The short-term memory is the active central processor of the mind, that is, it is
the sum of all the information we can hold in conscious attention at one time. We
notice its capacity most acutely when we try to learn a new task, such as driving a
car or playing bridge. Its limited capacity means that when faced with a complex
problem-such as writing a college paper-we can hold and compare only a few
alternative relationships in mind at once.
Trying to evaluate, elaborate, and relate all that we know on a given topic can
easily overload the capacity of our working memory. Trying to compose even a
single sentence can have the same effect, as we try to juggle grammatical and syntac-
tic alternatives plus all the possibilities of tone, nuance, and rhythm even a simple
sentence offers. Composing, then, is a cognitive activity that constantly threatens to
overload short-term memory. For two reasons Writer-Based prose is a highly effec-
tive strategy for dealing with this problem.
1. Because the characteristic structure of Writer-Based prose is often a list (either
of mental events or the features of the topic) it temporarily suspends the additional
problem of forming complex concepts. If that task is suspended indefinitely, the
result will fail to be good analytical writing or serious thought, but as a first stage in
the process the list-structure has real value. It allows the writer freedom to generate
a breadth of information and a variety of alternative relationships before locking
himself or herself into a premature formulation. Furthermore, by allowing the
writer to temporarily separate the two complex but somewhat different tasks of
generating information and forming networks, each task may be performed more
consciously and effectively.
2. Taking the perspective of another mind is also a demanding cognitive opera-
tion. It means holding not only your own knowledge network but someone else's in
conscious attention and comparing them. Young children simply can't do it.16
Adults choose not to do it when their central processing is already overloaded with
the effort to generate and structure their own ideas. Writer-Based prose simply elim-
inates this constraint by temporarily dropping the reader out of the writer's delibera-
tions. 17
My own research suggests that good writers take advantage of these strategies in
their composing process. They use scenarios, generate lists, and ignore the reader,
but only for a while. Their composing process, unlike that of less effective writers,
is marked by constant re-examination of their growing product and an attempt to
refine, elaborate, or test its relationships, plus an attempt to anticipate the response
of a reader. Everyone uses the strategies of Writer-Based prose; good writers go a
step further to transform the writing these strategies produce.
But what about the writers who fail to make this transformation or (like all of us)
fail to do it adequately in places? This is the problem faced by all teachers who
assign papers. I think this study has two main and quite happy implications for us as
teachers and writers.

'6Marlene Scardamalia, "How Children Cope with the Cognitive Demands of Writing," in Writing:
The Nature, Developmentand Teachingof Written Communication,C. Frederikson, M. Whiteman, and J.
Dominic, eds.
"Linda Flower and John R. Hayes, "The Dynamics of Composing: Making Plans and Juggling Con-
Approach,Lee Gregg and Irwin Steinberg,
straints," in CognitiveProcessesin Writing: An Interdisciplinary
eds. (Hillsdale, N.J.: Lawrence Erlbaum, 1979).
 Writer-BasedProse:A CognitiveBasisfor Problemsin Writing 37

The first is that Writer-Based prose is not a composite of errors or a mistake that
should be scrapped. Instead, it is a half-way place for many writers and often repre-
sents the results of an extensive search and selection process. As a stage in the
composing process it may be a rich compilation of significant thoughts which cohere
for the writer into a network she or he has not yet fully articulated. Writer-Based
prose is the writer's homework, and so long as the writer is also the audience, it may
even be a well-thought-out communication.
The second happy implication is that writing Reader-Based prose is often simply
the task of transforming the groundwork laid in the first stage of the process.18
Good analytical writing is not different in kind from the writer-based thought that
seems to come naturally. It is an extension of our communication with ourselves
transformed in certain predictible ways to meet the needs of the reader. The most
general transformation is simply to try to take into account the reader's purpose in
reading. Most people have well-developed strategies for doing this when they talk.
For a variety of reasons-from cognitive effort to the illusion of the omniscient
teacher/reader-many people simply do not consider the reader when they write.
More specifically, the transformations that produce Reader-Based writing include
these:
Selecting a focus of mutual interest to both reader and writer (e.g., moving from
the writer-based focus of "How did I go about my research or reading of the as-
signment and what did I see?" to a focus on "What significant conclusions can be
drawn and why?").
Moving from facts, scenarios, and details to concepts.
Transforming a narrative or textbook structure into a rhetorical structure built on
the logical and hierarchical relationships between ideas and organized around the
purpose for writing, rather than the writer's process.
Teaching writers to recognize their own Writer-Based writing and transform it
has a number of advantages. It places a strong positive value on writing that repre-
sents an effort and achievement for the writer even though it fails to communicate to
the reader. This legitimate recognition of the uncommunicated content of Writer-
Based prose can give anyone, but especially inexperienced writers, the confidence
and motivation to go on. By defining writing as a multistage process (instead of a
holistic act of "expression")we provide a rationale for editing and alert many writers
to a problem they could handle once it is set apart from other problems and they
deliberately set out to tackle it. By recognizing transformation as a special skill and
task, we give writers a greater degree of self-conscious control over the abilities they
already have and a more precise introduction to some skills they may yet develop.

'8For a study of heuristics and teaching techniques for this transformation process see L. Flower and J.
Hayes, "Problem-Solving Strategies and the Writing Process," CollegeEnglish, 39 (1977), 449-461.