Iowa State University Capstones, Theses and

Retrospective Theses and Dissertations

Dissertations

1998

An investigation of problem-solving theory and its

relationship to composition theory and pedagogy
Suzanne M. Kelsey
Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/rtd

Part of the Rhetoric and Composition Commons

Recommended Citation
Kelsey, Suzanne M., "An investigation of problem-solving theory and its relationship to composition theory and pedagogy" (1998).
Retrospective Theses and Dissertations. 198.
https://lib.dr.iastate.edu/rtd/198

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital
Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital
Repository. For more information, please contact [email protected].
 An investigation of problem-solving theory

and its relationship to

composition theory and pedagogy

by

Suzanne M. Kelsey

A Thesis Submitted to the

Graduate Faculty in Partial Fulfillment of the
Requirements for the Degree of

MASTER OF ARTS

Department: English
Major: English (Composition and Rhetoric)

Approved:

Signature redacted for privacy

In Charge of Major Work

Signature redacted for privacy

For the Major ,Department

Signature redacted for privacy

For the Graduate College

Iowa State University

Ames, Iowa
1988
 ii

TABLE OF CONTENTS

Page
CHAPTER I INTRODUCTION
CHAPTER II COGNITIVE PSYCHOLOGY: THE NEW APPROACH 5
TO LEARNING ABOUT HUMAN PROBLEM SOLVING
CHAPTER III RELATION OF PROBLEM-SOLVING THEORIES 20
TO COMPOSITION THEORIES
CHAPTER IV COGNITIVE STUDIES OF THE WRITING PROCESS 26
CHAPTER V ISSUES IN COGNITIVE RESEARCH ON THE 58
WRITING PROCESS
CHAPTER VI PEDAGOGICAL IMPLICATIONS 68
WORKS CITED 101
 CHAPTER I

INTRODUCTION

Composition as an academic field is still in its infancy.

While college students in the United States are now regularly

required to take composition classes, it wasn't until the

mid-ninteenth century that such classes began to be offered as a

standard part of the college curriculum (Underwood 9). Theorists

have borrowed concepts and ideas from other disciplines as they

have developed the discipline of composition, resulting in a

virtual explosion of conceptualization and research in composition

in the last twenty-five years.

One of the composition theorists borrowing from other

disciplines is Janice Lauer, who has advocated a return to

Aristotle's principles of rhetoric for invention strategies. She

has helped to revive Aristotle's topics, or methods of invention

(Lauer, Seminar). Young, Becker, and Pike's tagmemic theory,

which was borrowed from the discipline of linguistics, which

borrowed it from physics, is another example of discipline

transfer (Young and Becker, "Toward a Modern Theory" 130) • The

concept of "synectics," or using analogy as a way to explore a

topic, has been borrowed from a group of inventors, artists, and

psychologists trying to find creative solutions to problems

(Flower, Problem-Solving Strategies for Writing 85).

 2

Such invention strategies are commonly included in the

category of "heuristics." Heuristics are devices which people use

in their attempts to solve a problem, such as writing a paper.

Unlike rules and algorithms, which are completely systematic plans

(such as mathematical equations), heuristics are somewhat

systematic "in a clever way," in the middle of the continuum

between algorithms and trial and error (Lauer, Seminar). People

often intuitively use methods to solve problems without realizing

what these methods are. Explicating those intuitive methods makes

them heuristics ("Problem-Solving Strategies and the Writing

Process" 450-1). Heuristics also can be used in other stages of

the writing process besides invention, such as audience analysis

and revision.

While some composition theorists have turned to classical

rhetoric and various other disciplines for heuristics that can be

useful in the composing process, others have turned to more

scientific methods. Borrowing methods and terminology from the

field of cognitive psychology, these theorists attempt to discover

what actually goes on in a person's brain when that person is

writing; that is, they are studying the cognitive aspects of the

writing process. These cognitive theorists are using their

research to identify various heuristics that writers actually use.

They feel that the heuristics, once identified, can be used for

writing instruction.
 3

Linda Flower, who works closely with cognitive psychologist

John Hayes at Carnegie-Mellon University, is one of the most

prominent of the composition theorists who are applying cognitive

research to the study of writing. Flower and Hayes have defined

writing as a "problem-solving" process and have attempted,

primarily through protocol analysis, to find out what writers do

as they solve the "problem" of writing. The work of cognitive

theorists, to be sure, is not received without skepticism by many

in the field of composition. However, it is possible to draw,

from the work of Flower and Hayes and others, a set of

implications for pedagogy that can be useful in teaching the

composition process.

Chapter II of this study will first address the discipline of

cognitive psychology, and try to answer the following questions:

What is cognitive psychology? What is problem solving and why

does cognitive psychology study it? In Chapter III, the focus

will turn to the connection between cognitive psychology, problem

solving, and writing. What is the connection? How is it studied?

In Chapter IV, the researchers will be discussed: Who is

researching the cognitive processes involved in writing as a

problem-solving activity? and What are they finding out?

Chapter V will address some of the limitations: What are some of

the issues and problems with the cognitive theorists' works?

Finally, and probably most importantly, Chapter VI will try to

answer the questions: What does all this mean for instructors who
------------~- -------·---------- _ _ _ ___ -------------------------------
"

must face their composition classes on Monday morning? What

parts, if any, of current theory on the problem-solving process of

writing might they want to implement in their classrooms? And:

How do they do it?

 5

CHAPTER II
COGNITIVE PSYCHOLOGY:
THE NEW APPROACH TO LEARNING ABOUT HUMAN PROBLEM SOLVING

The field of psychology has undergone numerous changes within

this century. Howard Gardner notes that even in the early 1960s,

"psychology seemed a rather remote and sterile area to individuals

interested in the full and creative use of the mind." According

to Gardner, academic psychology, behaviorism, and psychoanalysis

were the main areas studied. Academic psychology featured the use

of contrived laboratory apparatus which studied "the perception of

visual illusions or the memorization of long lists of nonsense

syllables." Behaviorism based its theories on the idea of

reinforcement for behavior and consequently "denied inner life--no

thoughts, no fantasies, no aspirations." Thus, studies of human

thought processes and use of language were largely ignored by

behaviorists. Psychoanalysis emphasized human personality and the

unconscious, but said "little about rational thought processes or

conscious problem-solving" (Gardner 3).

The Cognitive Revolution

Gardner describes the cognitive revolution, which he says

came in two parts. "First, there was the frank recognition that

one could--one must--take seriously human mental processes,

including thinking, problem-solving, and creating." The study of

the mind was considered a more valid activity than some had
 6

considered it to be in the past. Second, researchers demonstrated

that "human thought processes were characterized by considerable

regularity and structure." While not all of these thought

processes could be directly observed, there was "structure to

thought processes, a structure the careful analyst could help lay

bare" (4). Cognitive psychologists concern themselves, then, with

the structure of thought processes.

John Hayes agrees with Gardner's assessment that a cognitive

revolution occurred in the late 1950s and early '60s. He

attributes the revolution, however, more specifically to two

publications which appeared in the late 1950s. One was "Elements

of a Theory of Human Problem Solving" (1958), a paper by Herbert

Simon and Allen Newe 11. The other was Noam Chomsky's "A Review of

B.F. Skinner's Verbal Behavior" (1959). These and subsequent

articles featured three factors which, says Hayes, were important

in the growth of cognitive psychology:

1. The growing awareness among psychologists that

complex mental structures were essential to
psychological theories as exemplified by Piaget.

2. The Behaviorists' failure to make progress in

describing language led many (such as Chomsky)
to believe that a new theoretical approach
was desirable.

3. The development of information-processing models

(such as that of Simon and Newell)
provided a new approach which facilitated the
description of complex mental structures.
(Paraphrased from Hayes, Cognitive Psychology 151)
 - - - - - --------------

Again, cognitive psychologists concern themselves with the

structure of thought processes, or the "description of complex

mental structures" in relation to human thought and behavior.

This concern has taken them into such areas as the psychological

processes underlying language generation, problem solving, and

creativity. For example, research in artificial intelligence by

cognitive psychologists has enhanced knowledge about language

generation, which in turn has caused researchers to examine "the

processes involved in understanding larger units of

language--paragraphs, stories, problem descriptions, scripts, and

even whole belief systems" (169).

Because cognitive researchers have extended their focus to

larger units of language, composition theorists have also begun to

apply cognitive research to the field of composition, causing a

revolution of their own. (Work of composition theorists will be

discussed in Chapter III.)

Problem Solving

One of the psychological processes studied widely by

cognitive psychologists is that of human problem-solving behavior.

Within this domain, three main models have been offered: (1)

information processing models; (2) models stressing human

abilities and factors; and (3) creative problem solving models

(Feldhusen and Guthrie 22). The cognitive process models used by

composition theorists are primarily related to the information
 8

processing models, upon which the remainder of this chapter will

focus.

The Information Processing Model

To date the most important information processing model comes

from Newell and Simon in their 1972 text, Human Problem

Solving. Because their work is drawn upon extensively by Flower

and Hayes in their cognitive model of writing, it will be useful

to lay out a few of Newell and Simon's concepts.

The fundamental components of Newell and Simon's information

processing model include:

1. a long-term memory function

2. a short-term memory function
3. serial processing, limited by individual rates of
processing
4. information about the environment acquired by the
different sensory systems
5. problem solving behavior affected by external memory
aids
6. goal-directed problem solving (Paraphrased from
Feldhusen and Guthrie 9)

Such an information processing model·, in addition to stating

generalities about human behavior, can "represent in some detail a

particular man at work on a particular task." This representation

is not a metaphor (e.g., computer for man), but a "precise

symbolic ~odel" which can help calculate aspects of human problem

solving behavior (Newell and Simon 5).

 ----- ---------~--~----------- --- ----

This model remains approximate, especially since its emphasis

on performance omits motivational and personality variables, but

Newell and Simon believe that i t "posits internal mechanisms of

great extent and complexity, and endeavors to make contact between

them and the visible evidences of problem solving" (10).

Feldhusen and Guthrie believe that the Newell/Simon model is a

valid representation of human problem solving because it clarifies

parts of the problem-solving process and begins to suggest

instructional activities.

Memory

But more crucial to our discussion here is the Newell/Simon

model's obvious reliance upon factors involving memory. The

ability to remember information is critical to the problem-solving

process, and, as we have seen, plays a primary role in the

information processing model. Psychologists who have studied the

structure of human memory theorize that people remember things in

"chunks," which are "package(s) of information that (are) treated

as a unit." Further, psychologists have determined that the

capacity of short-term memory is seven chunks, plus or minus two

(Hayes, Problem Solver 73). Chunking is an activity that people

perform, even if they are not aware that they are doing it (75);

people can make remembering easier if they consciously search for

chunks. Of course, writing these chunks down, or representing

them externally through a diagram or matrix, can free up our
 10

short-term memory for other tasks, thus making our problem-solving

efforts more efficient.

Frederiksen labels the short-term memory, "working memory."

He notes that there are "two kinds of information processing:

controlled and automatic." A controlled process requires

deliberate attention and is under the control of the person. It

involves the activation of chunks, or what Frederiksen also calls

"nodes." Consequently, "controlled processing • • • quickly uses up

the capacity of working memory." Automatic processing is carried

out without the person's attention, so it doesn't use much or any

of the capacity of working memory. Problem-solving capacity can

be "greatly increased by learning to use automatic processing for

the more routine elements of an activity, making available

controlled-processing resources for the novel aspects of problem

solving." Frederiksen gives the example of reading to illustrate

his point. The basic skills that we learn, such as "decoding

orthographic forms, translation into speech units, retrieving word

meanings, and establishing relationships among semantic

propositions," become automatic, which make it possible for us to

simultaneously process large amounts of information (364-366).

Long-term memory is also important to the problem-solving

process. The remembering of old patterns can aid us in new

problem-solving situations. Further, knowledge located in our

long-term memory is important in problem solving. Says Hayes, "If

you are missing relevant knowledge, an easy problem may appear

 11

difficult or impossible • • • • Much that passes for cleverness or

innate quickness of mind actually depends on specialized knowledge

(Problem Solver iv-v). While the short-term memory may handle

about seven chunks, the capacity of long-term memory

"is thought to be virtually limitless" (Frederiksen 364), a

proposition that will be vital when we come to applications of

problem-solving theory to the writing process.

Successful Problem Solving

Hayes, whose problem solving text builds on the work of

Newell and Simon, says there is a problem "whenever there is a gap

between where you are now and where you want to be, and you don't

know how to find a way to cross that gap" (Problem Solver i).

Another definition of a problem is that "A problem exists when

'what is' differs from 'what is desired'" (Pfeiffer eta!. 104).

Solving a problem, according to Hayes, involves finding a way to

cross the gap.

Success in "crossing the gap," says Hayes, depends on how

effectively the following activites are carried out:

1. Finding the problem

2. Representing the problem

3. Search, or planning the solution

4. Carrying out the plan

5. Evaluating the solution

---------------------------------- - - - -

12

6. Consolidating gains. (Problem Solver 1)

"Problem representation," "search" (planning the solution), and

"evaluating the solution" are three of the major activities

involved in finding answers to problems. Following is a

discussion of these three activities.

Problem Representation:

According to Hayes, there are four elements that people may

need to include when they internally represent problems:

1. The Goal--where we want to be when we are done.

2. The Initial State--where things are at the beginning

of the problem.

3. The Operators--the actions that change one problem

state into another.

4. The Restrictions on the Operators. (3-4)

While all problems involve at least a goal, one or more of the

other parts may be absent or omitted in a problem. Hayes uses the

example of a friend saying to another friend, "Get to my house at

ten o'clock" (4). This problem states the goal, but specifies no

initial state (where to start from), no operators (how to get

there), and no restrictions (how not to get there).

People represent problems to themselves in a variety of ways.

Some are better than others at filtering out irrelevant detail.

Some represent problems to themselves in visual imagery, others in

 13

sentences, others in auditory images (Hayes 7-8). How problems

are represented can make a big difference in the difficulty of the

problem (ii). Frederiksen says, "An inaccurate or incomplete

problem representation may make it difficult or impossible to

solve the problem" (367). For example, external representations,

such as sketches, lists, equations, and diagrams may be helpful,

if not necessary, in a problem-solving situation. These aids can

help the problem solver to remember information and to discover

new relations in the problem (Hayes, Problem Solver 11).

Hayes notes that people may change their internal and/or

external representations of the problem as they are solving them.

In fact, if they are experiencing difficulty in solving the

problem, it may be helpful to change the representation (14).

For instance, "Joe" might have difficulty in solving a writing

problem which he has internally represented as, "I need to write

an 'A' paper for my composition class." Joe might be better able

to solve his problem by changing his problem representation to

something like, "I need to write a paper on something I'm vitally

interested in so that I can enjoy myself while I research it and

write it, and so that I can hopefully get a better grade."

However, sometimes an inaccurate or incomplete problem

representation is a result of an ill-defined problem. Norman

Frederiksen explains ill-defined, or ill-structured problems, in

contrast to well-structured problems, as having at least three

qualities: (1) they are more complex and have less definite
 14

criteria for determining when the problem has been solved; (2)

they do not come equipped with all the necessary information with

which to solve the problem; and (3) they have no "legal move

generator," or something to indicate all the possibilities at each

step.

Processes for solving well- and ill-structured problems are

basically the same, but for ill-structured problems (which include

most writing tasks) a person's idea or conception of the problem

changes gradually as new elements are invoked from long-term

memory or other sources. Also, a "wide repertory of recognition

processes is necessary to evaluate whether one is 'getting warmer'

as a result of each altered state" (366).

Once the problem is at least initially represented, (though

the representation may change with time), the next major part of

solving comes into play: the search, or planning a solution.

Search:

Hayes says that problem solving can be thought of as a

process in which the problem solver "searches through the problem

space to find a solution path." If there are many "blind alleys"

for each solution path, then we can expect the problem to be

harder than if there are just a few blind alleys. Finding one

needle hidden in a large haystack is probably going to be harder

than finding the needle in a small haystack. But the size of the
 15

space through which the problem solver must searcp is not as

important as the method by which the search is conducted

(Cognitive Psychology 183).

Search methods can lie anywhere on a continuum from simple to

complex. "Trial and Error" search methods are fairly simple

methods and may or may not in the long run be the most efficient

(Problem Solver 30). Many novice writers attack a writing problem

with trial-and-error methods, jotting down phrases and sentences

as they occur to them, with no particular plan.

"Means-end" analysis is a more sophisticated method, and,

consequentially, more powerful. The person employing means-end

analysis uses an internalized means-end table in order to find a

means which will reduce the distance to the end, or the goal (32).

For instance, a writer trying to decide how to focus a topic for a

paper might use a brainstorming technique, listing a variety of

ways the paper could be focused. This technique may proved to be

more efficient for the writer, who might otherwise write several

drafts just trying to find a focus. Thus, the brainstorming

serves as a means which reduces the distance to the end.

Fractionation is another search m~thod which is very closely

related to means-end analysis, but even more sophisticated, with

less of the trial and error element. To apply the fractionation

method, people break the problem into parts and use subgoals to

guide themselves around the details. After breaking the problem

into parts, they can drop a restriction, solve a sub-problem, then

----------- ·----------------~-----···

16

add the restriction and fix it again. Many do this when faced

with a writing task. Because it would be overwhelming to consider

all the demands of the task at once, they break it into parts,

such as research, brainstorming, organization, writing a rough

draft, and revising (38-39), and work on one part at a time.

There are also knowledge-based search methods, of which

algorithms are an example: knowing how to perform long-division

entails having a formula that is guaranteed to solve the problem.

But people can also learn patterns, which, while not algorithmic,

may reduce the difficulty of the search (~0). For instance,

writers who have written a variety of letters-to-the-editor in the

past can base their next letter-to-the-editor on patterns which

they have probably internalized, thus reducing the difficulty of

the search for the correct pattern.

Evaluating the Solution:

In addition to "problem representation" and "search," Hayes

notes that evaluating the solution is an important part of the

problem-solving process. He gives an example of evaluation which

.
may strike a note with writing instructors. Sometimes, he says,

people need an outside opinion, such as with writing:

Testing writing out on the audience for which it

is intended is a very simple idea with obvious
advantages, but very few people do it. Major
corporations and governmental agencies put a great
deal of money and effort into producing instruction
manuals and other documents to inform the public, yet
 17

rarely consider testing those documents on the

audience for which they were intended. (48)

Hayes also notes the importance of consolidation, or

reflecting on the problem-solving experience and learning from it.

Consolidation can help people be more effective in solving future

problems.

In summary, successful problem solving depends on a good

problem representation (whether the problem is well- or

ill-defined), a good search, or plan of solution, and a final

evaluation of the solution.

Creativity

Creativity is a concept that often comes up in association

with problem-solving, particularly when ill-structured problems

are concerned. A person must possess elements of creativity in

order to have the insight to solve an ill-defined problem.

Frederiksen notes that theories of creativity are older than

theories of problem solving, yet simplistic in comparison, because

"creativity as such has so far received little attention from

cognitive psychologists" (384). Frederiksen classifies theories

of creativity into three groups:

(1) Stage Theories--the earliest stage theories were largely

based on introspections of poets, scientists, and mathematicians.

Helmholtz was one of the first to list stages, which were a)

saturation, b) incubation, and c) illumination. Poincare (1952)
 18

called the first step "preparation" and added a fourth step called

"verification." Other theories have named "inspirational" and

"elaborational" phases. Most of the stage theories imply that the

subconscious part of the mind continues to work on the problem

even when the conscious part of the mind doesn't. However, more

recent studies have indicated that this hypothesis is not valid.

(2) Creative individual theories. Studies of creative people

yield the information that research scientists have high energy

levels, persistence, curiosity, and independence. Other creative

people, such as architects, are self-confident, flexible,

self-accepting, and not concerned much with the opinions of other

people. Creative people also prefer complexity.

(3) Cognitive theories. These theories use more measurable

factors in researching creativity, such as cognition, memory,

divergent and convergent production, verbal comprehension,

cognitive flexibility, and knowledge. One finding of this

approach is that when problems are ill-structured, skill in

retrieving information from long-term memory is important, along

with reasoning (384-387).

John Hayes focuses on Frederiksen's third category, or

cognitive theories, in his discussion of creativity. He says a

creative act must satisfy two criteria: First, it must have "some

valuable consequence." That is, the act must have an interesting

or useful effect. Second, the creative act must be "novel and

surprising." If we find it difficult to understand how the act

 19

was performed, we are more likely to judge it creative (Cognitive

Psychology 215-216).

Hayes tells us that four cognitive processes underlie

creative acts, and we can see in these processes the parallels to

the problem solving elements of problem representation and search

indentified above. The processes are: (1) Problem finding, (2)

Idea generation, (3) Planning, and (4) Preparation. Hayes notes

that "extensive preparation is essential for acts of outstanding

creativity." For instance, studies show that musical composers

require about ten years of intense preparation before they are

able to produce excellent works. He warns us that if a person is

"basically lazy" and has decided to go into creative work, s/he

has made a mistake, as creative people typically work 70-80 hours

a week (Hayes, Problem Solver 199-215).

Summary

Cognitive psychology, a relatively new field of study, has

helped researchers to make new, exciting observations about such

things as human problem-solving behavior and creativity. Other

disciplines are borrowing methods and findings of cognitive

psychologists with which to make more observations. Composition

theorists are participating in this discipline concept-transfer.

Some of their findings will now be addressed.

 20

CHAPTER III
RELATION OF PROBLEM-SOLVING THEORIES
TO COMPOSITION THEORIES

As indicated, research in the area of human problem solving

is still in its infancy. Nevertheless, cognitive psychologists

are discovering some exciting things about the way we solve

problems, and other disciplines are borrowing some of their

findings, terminology, and methodology, in order to explore new

perspectives. Cognitive psychology's discoveries of human problem

solving techniques seem to lend themselves to studies of the

writing process, since the process of writing is an ill-defined

problem.

The writing process is implied in Frederiksen's description

(see Section II) of ill-structured problems. The well-known

metaphor, "writing as discovery" (which, we shall see later, may

have its disadvantages, but serves a purpose here), illustrates

the fact that a person's idea or conception of the problem changes

gradually as new elements are invoked. There are no algorithms

for producing a well-polished essay. Johq Hayes notes that in

order to solve ill-structured, or ill-defined problems, "we may

have to make gap-filling decisions and we may have to jump into

the problem before we understand it" (Problem Solver 22).

Certainly this is true of the writing process.

It should not be surprising, then, that composition theorists

and rhetoricians are studying the act of writing as a problem-

solving process, using protocol analysis and other methodology/

 21

terminology borrowed from cognitive psychology. This study of

writing ~ a problem-solving process is different from using

writing to solve problems, such as problems in business, society,

and personal relationships, though both borrow from cognitive

psychology. An brief discussion follows first of the latter

concept (writing to solve problems), then a more detailed one of

the former (writing as a problem-solving process).

Writing to Solve Problems

The idea that writing can and should be used as a tool for

analyzing problems in society and other areas of daily life is

discussed by Richard E. Young, Alton L. Becker, and Kenneth L.

Pike in the now classic textbook: Rhetoric: Discovery and

Change. An example of an assignment which stimulates

problem-solving activity is found at the end of their chapter

titled "Preparation: Identifying and Stating the Problem":

Spend the next fifteen minutes listing all the things

in your immediate surroundings that clash in some way
with features of your image; then state the problematic
situations. Ask yourself why, if these situations are
indeed problematic, you don't set about eliminating
them. ( 100)

Clearly, this assignment is an example of using writing to analyze

problems.

Many of Young, Becker, and Pike's statements about problem

solving have become familiar, such as: "Problems do not exist

 22

independent of men. There are no problems floating around in the

world out there waiting to be discovered; there are only problems

for someone." Further, problems come to the surface when there is

an inconsistency, or a clash: "when a person becomes aware that

two beliefs to which he is deeply committed are incompati-

ble • • • when he discovers something in the nature of the world

that doesn't 'fit' his conception of it." When a person becomes

cognizant of such an inconsistency, s/he finds him/herself in a

"problematic situation." The uneasy feeling that accompanies this

awareness is part of the earliest stage of inquiry; ignoring it

means that the process of inquiry may never begin (90-1).

Once a problem is recognized, it is important to make

explicit the unknown in a complete statement, for the statement is

"actually a partial description of the solution." (Recall that

Hayes would classify this statement as "problem representation.")

The statement acts as a guide to inquiry and also helps us to know

when we have found our solution. Sometimes a problem may actually

be a cluster of "subordinate" problems, which individually must be

solved in order for the larger one to be completely solved. A

well-stated unknown, then, "serves as an instruction for effective

investigation. It defines what is sought and guides but does not

construct inquiry" (92-96).

It is obvious that Young, Becker, and Pike's chapter on

identifying and stating problems through writing borrows concepts

from cognitive psychology. Richard Larson, in "Problem-solving,

 23

Composing, and Liberal Education," published just two years after

the Young, Becker, and Pike text, also suggests that methods

derived from cognitive psychology can be used in writing to solve

problems that people face in society. Problem-solving, he says,

is a way for a student to plan arguments on complex issues. It is

also a technique for reflecting on experience; for discovering

judgements, values, and desirable actions; and for evaluating the

contentions of others. If instructors of writing claim that their

courses "are more than laboratory sessions on linguistic and

social conformity," and if instructors believe that their classes

"have a place in the liberal studies curriculum more honorable

than that of servant to other disciplines," then teaching

problem-solving techniques may be of benefit (635).

The problem-solving process relates to writing in two ways:

it is ''at once an activity of mind and a principle of form capable

of organizing expository or argumentative writing." As an

activity of the mind, problem solving "represents an organized way

of confronting and dealing with issues and data--a way that

psychologists say is characteristic of many persons we think of as

'creative."' As a principle of form, the writer can carry on "a

systematic inquiry into his subject, using a controlled procedure

that dependably yields reasonable conclusions. The rhetorical

medium, in effect, is a major part of the message; it operates as

a form of ethical proof" (632). Larson's words, "activity of the

mind," pave the way for what comes later in the 1970s: the study
 24

of writing~ a problem-solving process, with guided inquiry into

the cognitive processes of the human mind.

Writing as a Problem-Solving Process

In 1971, Janet Emig published The Composing Processes

of Twelfth Graders, and this monograph was the first effort to

address "the texture of the composing experience." Emig's text

has caused an "explosion of research on composing processes in the

last decade" (Lunsford 155). Her monograph, which uses case

studies of high school seniors in order study the writing process,

seems to have established a new emphasis in

scholarship on composition that others
willingly continued: an emphasis on addressing the
texture of the composing experience, on what
practicing writers do when they write, on how (insofar
as one can get at the subject) their minds work.
(Larson, "Recent Research" 244)

Linda Flower has extended Emig's study of the writing

process, attempting "what are easily the most comprehensive

studies of the composing experience" (244). Shortly after Emig's

seminal text, Linda Flower and John Hayes published

"Problem-Solving Strategies and the Writing Process" (1977). This

article was the first, at least in the field of composition, to

put forward a theory of writing as a problem-solving process and

to study it as such. Flower and Hayes note that what many

 25

instructors teach in the classroom about writing does not

adequately cover the subject. Within the classroom,

writing appears to be a set of rules and models for the

correct arrangement of preexistent ideas. In contrast,
outside of school, in private life and professions,
writing is a highly goal-oriented, intellectual perform-
ance. It is both a strategic action and a thinking
problem. ( 449)

Flower and Hayes discuss problem solving as a new area in

cognitive psychology, one which has a "well-developed method for

studying thought processes." They wish to study writing as a form

of problem solving. That is, the activity of writing anything,

whether a problem-solving paper, a short story, a poem, or an

informational essay, is an activity composed of problem-solving

strategies. Flower and Hayes, in their study of writing as a

problem-solving activity, wish to discover and describe "some of

the basic heuristic procedures which underlie writing, and then to

translate these heuristics into teachable

techniques" (450). A discussion of the heuristics that Flower and

Hayes, and others, have identified will appear in Chapter VI,

after the following review of the research on the cognitive

processes of writing which has accumulated since the late 1970s.

 26

CHAPTER IV
COGNITIVE STUDIES
OF THE WRITING PROCESS

Overview

As indicated, there has been a proliferation of research into

the cognitive processes of writing since Janet Emig's and Flower

and Hayes' ground-breaking works. Flower and Hayes particularly,

in their identification of writing as a problem-solving proeess,

have opened up a new way of learning about the writing process--a

more empirical, scientific way. The approach offends many in this

humanistic profession who are accustomed to more intuitive ways of

thinking about writing. But cognitive studies of the writing

process, while we may not want to embrace them unconditionally,

may tell us things we don't know; or, they may confirm empir•ically

what our intuitions have told us is true.

Scardamalia and Bereiter call the cognitive science approach

"the only paradigm in town for investigating complex mental

processes, which all sides agree are of central concern in

writing." In fact, they pinpoint nine different strands of

research in composition that have emerged within the cogniti.ve

framework in the last decade, the most important of which for our

purposes here is research on the composing process itself. This

work has resulted in a "fairly coherent description of mental

processes that go on in writing" ("Research" 780).

 27

The efforts of cognitive theorists in investigating the

writing process can be categorized into at least three major

groups.

First, Flower and Hayes offer a cognitive model of the

composing process, which divides the writer's world into three

parts: the task environment, the writer's long-term memory, and

the writing process. (This model will be discussed shortly.)

Their model and research "provides us with extremely valuable

information on how writers set goals, how they solve problems, and

how they represent meaning to themselves" (Andrea Lunsford 155).

Second, Bereiter and Scardemalia incorporate Flower and

Hayes' model, along with models of speech production and of

language comprehension in order to advance a tentative model of

"stages in writing development." According to Lunsford, this

model is "even more tentative" than that of Flower and Hayes, and

"more seriously flawed." However, the model "focuses our

attention not only on describing the process but on relating that

process to instruction as it may or may not aid development of

writing ability" (156-157).

The goal of Bracewell, Frederiksen, and Frederiksen, the

third group of researchers, is to "develop a unified theory of

discourse production (writing) and comprehension (reading)."

Further, by linking the cognitive processes common to both reading

and writing, Bracewell et al. move toward a more general model of

cognition and communication (157).

 28

Research by all three groups, while "tentative and evolving,"

emphasizes "questions of process and representation and attempts

to develop models that would identify and define the cognitive

processes involved in reading and writing" (Lundsford 157).

It is important to recognize that identifying certain

strategies that can be taught from various models is certainly as

important, if not more so, than identifying the stages or models

themselves. However, as necessary background, there follows a

detailed discussion of the Flower and Hayes composing process

model, from which many of the most important pedagogical

strategies have been gleaned (see Chapter VI). Theories of other

researchers will be covered as well, usually in the context of the

Flower and Hayes model.

The Act of Writing: A Cognitive Process Model

To date, the most extensive cognitive process model for

writing is that developed by Flower and Hayes ("Identifying the

Organization of Writing Processes," 1980; and "A Cognitive Process

Theory of Writing," 1981). Elements of this model are presented

below, along with pertinent research by other theorists.

Flower and Hayes' "Structure of the Writing Model" appears

below. We must first understand that the multiple arrows in the

diagram indicate only that information flows from one box to

another, not that "such information flows in a predictable left to

right circuit, from one box to another as if the diagram were a

 29

Structure of the Writing Model

(From: Flower, Linda, and John R. Hayes. "A Cognitive Process

Theory of Writing." College Composition and Communication 32
(1981): 365-386.

TASK ENVIRONMENT

THE RHETORICAL TEXT

PROBLEM
Topic PRODUCED
Audience
Exigency· SO PAR

WRITING PROCESSES

1HE WRITER'S LONG-TERM PLANNING TRANSLATING REVIEWING

MEMORY --

~
Knowledge of Topic,
Audience,
I
ORGANIZING] I EVALUATING I
and Writins
Plans
Ill I GOAL
SETTING
I I REVISING
I
I 1 I
r MONITOR
I
 30

one-way flow chart." A flow chart would imply a stage model,

which is not the intent, for the writing process can not be

orchestrated in a linear series of stages ("Cognitive Process"

386-7).

The model is based on four key assumptions:

1. The process of writing is best understood as a set

of distinctive thinking processes which writers
orchestrate or organize during the act of composing.

2. These processes have a hierarchical, highly embedded

organization in which any given process can be
embedded within any other.

3. The act of composing itself is a goal-directed

thinking process, guided by the writer's own growing
network of goals.

4. Writers create their own goals in two key ways:

by generating both high-level goals and supporting
sub-goals which embody the writer's developing
sense of purpose, and then, at times, by changing
major goals or even establishing entirely new ones
based on what has been learned in the act of
learning. (366)

In their attempts to develop their model, Flower and Hayes have

focused on determining differences between "novice" and "expert"

writers.

As the model illustrates, the act of writing involves three

major elements: the task environment ("all the things outside the

writer's skin, starting with the rhetorical problem or assignment

and eventually including the growing text itself"); the writer's

long-term memory ("in which the writer has stored knowledge, not

only of the topic, but of the audience and of various writing

 31

plans"); and the writing processes ("specifically the basic

processes of Planning, Translating, and Reviewing, which are under

the control of a Monitor") (369). Following is a discussion of

these elements.

The Task Environment

The task environment consists of the rhetorical problem and

the growing text itself. Defining the rhetorical problem, which

is composed of (1) the rhetorical situation (topic), (2) audience,

and ·(3) the writer's own goals (exigency), is a major part of the

writing process. However, writers often reduce a rhetorical

problem to a simplistic level, such as "write another theme for

English class. 11 SUch ill-defined (or "under-developed")

rhetorical problems are unlikely to be "solved" by the writer

because the right questions cannot be addressed.

The act of internally formulating questions is called

"problem finding" by some, but Flower and Hayes say it is "more

accurate to say that writers build or represent such a problem to

themselves, rather than 'find it.'" A rhetorical problem is never

something a person discovers; rather, "it is an elaborate

construction which the writer creates in the act of composing."

Thus, Flower and Hayes caution against limiting writing to the

metaphor, "writing as discovery," because it implies that "hidden

stores of insight and ready-made ideas exist, buried in the mind

 32

of the writer, waiting only to be 'discovered"' ("Cognition of

Discovery" 21-22).

Flower and Hayes have used protocol analyses of novice and

expert writers to arrive inductively at a list of "the basic

elements of a writing problem which a given writer could actively

consider in the process of composing, if he or she chose to:

THE RHETORICAL PROBLEM

Elements of the Problem Protocol Examples

The Rhetorical Situation:

Exigency or Assignment "Write for Seventeen
magazine; this is
impossible."
Audience "Someone like myself,
but adjust for twenty
years."
The Writer's Own Goals
involving the:

Reader "I'll change their

notion of English
teachers ••• "
Persona or Self "I' 11 look like an
idiot if I say ••• "
Meaning "So if I compare those
two attitudes ••• "
Text "First we'll want an
introduction." (24)

Flower and Hayes note that these elements closely parallel the

four terms of the communication triangle, which are: reader,

writer, world, and word.

 33

Flower and Hayes' protocol analyses show that one of the big

differences between novice and expert writers is how many elements

of the rhetorical problem they consider and "how thoroughly they

represent any aspect of it to themselves." (Recall Hayes'

statement in Chapter II that problem representation is one of the

major activities employed in solving problems.) Further, it has

become apparent that experienced writers have more "stored

representations" of complex rhetorical problems (24-25). That is,

when faced with a particular rhetorical situation, such as "Write

an analysis of a newspaper article," experienced writers can draw

upon prior writing activities that engaged in similar analyses,

which help them with the present writing activity. (This strategy

is also known as a "knowledge-based search method," as discussed

in Chapter II. )

In addition to responding to more elements of the rhetorical

problem, Flower and Hayes note that good writers "create a

particularly rich network of goals for affecting the reader."

Also, good writers continue to more specifically develop their

image of the reader and other elements as they write. However,

poor writers often stick with a poorly developed problem

representation with which they started, and in general do not

change their conception of the audience or other rhetorical

concerns. Consequently, "good writers are simply solving a

different (rhetorical) problem than poor writers" (30).

 34

Along these same lines, Scardamalia and Bereiter suggest that

novice writers such as elementary and secondary students don't

even give themselves problems to solve when writing for an

assignment. Employing a strategy called "knowledge telling,"

these novice students reduce writing assignments to "topics," then

tell what they know about the topic. Since this strategy does not

involve explicitly formulating goals and representations for the

text, it "has the effect of eliminating much of the problem

solving observed in expert composing" ("Assimilative Processes

165). One reason for knowledge-telling behavior could be that

most instruction in writing encourages the perfection of the

knowledge-telling strategy. Is it possible, ask Scardamalia and

Bereiter, that "knowledge-telling strategy is reserved for

ordinary school tasks that students perceive as pointless and that

other strategies are available to them when a more personal goal

exists?" (167)

However, the reason for knowledge-telling behavior is

probably more complex than inadequate assignments. Goals are· a

form of knowledge, but that goal-related knowledge seems to come

rather late in writing development--later than other forms of

knowledge. Thus, a high school student may have a knowledge of a

subject, enough to produce a five-paragraph theme, but may not

have the kind of knowledge associated with goals that could help

him/her write a more meaningful essay. The instructor's job is

seen as helping students learn how to assimilate externally-

 35

assigned tasks to meaningful goals of their own (169-70).

We can see, then, that responding to the rhetorical problem

is in itself a complex set of tasks, involving the rhetorical

situation, audience, and the writer's own goals.

The second element in the task environment is "the written

text." Largely unexplored by process-oriented cognitive

researchers such as Flower and Hayes, the written text is an

element which places even more constraints upon the writer. "Just

as a title constrains the content of a paper and a topic sentence

shapes the options of a paragraph, each word in the growing text

determines and limits the choices of what can come next" (Flower

and Hayes, "Cognitive Process" 371).

In addition to the growing text, the writer's knowledge

stored in long-term memory and the writer's plans for dealing with

the rhetorical problem also direct the composing process.

The Writer's Long-Term Memory

The writer's long-term memory can exist not only in the mind,

but also in outside resources, such as books. The long-term

memory "is a storehouse of knowledge about the topic and audience,

as well as knowledge of writing plans and problem

representations." Sometimes an assignment, such as to write

something persuasive, can have a single cue which causes a writer

to remember "a stored representation" of a previous problem, thus

 ----~·----------·--~-~-~-·-~~-~- ··~ ·--·~·~-~~~~

36

bringing a whole set of writing plans into play ("Cognitive

Process" 371).

Use of long-term memory, a "relatively stable entity" with

its own internal organization, may pose two problems. One problem

can be retrieving things out of it, which entails finding the cue

which will let a network of useful knowledge be retrieved. The

second problem is reshaping the information retrieved to "fit the

demands of the rhetorical problem" (371).

The phenomenon of "Writer-Based" prose is an example of a

writing strategy based on memory retrieval, with no adaptation for

the rhetorical problem. Writer-based prose is organized so that

it "faithfully reflects the writer's own discovery process and the

structure of the remembered information itself," but it fails to

make the transformation or reorganization which is necessary to

meet the needs of the reader (371-2). Obviously, the writer must

be able to shape or adapt the information from long-term memory in

order to serve the reader's needs.

The Writing Process

The final major element in the writing model is called "the

writing processes," which Flower and Hayes break into further

groups: planning (with sub-categories of "generating,"

"organizing," and "goal setting"), translating, reviewing (with

the sub-categories, "evaluating" and "revising"), and the monitor.

It is crucial to remember that the writing process is not linear,

 37

with progression from one stage to another; rather, the writer

juggles the elements of planning, translating, and reviewing at

the same time, or slips from one to another and then back again.

Planning:

Planning is one of the most powerful heuristic procedures

because it can reduce a large problem to manageable size. (Recall

that planning a solution is discussed as "Search" in Chapter II,

and that search methods range from simple, trial-and-error

methods, to more complex methods, such as "means-end" or

"fractionation.") In fact, the ways writers coordinate their

various plans "affect the efficiency of their composing processes

and the effectiveness of their final products (Flower and Hayes

"Plans" 39).

While composing, expert writers use a variety of fractionated

plans: "plans for generating ideas" (generating), "plans for

producing a paper" (organizing), and "plans which guide the

composing process itself" (goal setting) ("Plans" 42). In

another article, Flower and Hayes call these three categories,

"plans to say," "plans to do," and "composing plans" ("Dynamics"

44). The latter categories will be used for ruther discussion.

Plans to say: Within the category of plans for generating

ideas ("plans to say"), writers make "procedural plans" and

"content-specific" plans. An example of a good procedural plan
 38

might be, "I'm just going to jot things down as they occur to me,"

or "I'll worry about the spelling later." A less defined plan,

and consequently less effective one, might be, "I'll just kind of

follow things along." These procedural plans are "content-free

decisions about how to control. •• idea-generating activities."

An advantage of good plans is that they help a writer to keep

focused on high-level goals, such as developing a broad set of

ideas (42-43).

Most idea-generating plans, however, are "content-specific,

acting on the information available immediately to the writer."

Flower and Hayes identify four content-specific plans exhibited by

their student writers:

1. Pursuing an Interesting Feature. Here, a writer

explores an interesting word, idea, or event by
using various "generating techniques," such as
searching memory, drawing inferences, and reasoning
from examples. This exploration leads to more
complex processes.

2. Thinking by conflict. Here, the writer either

asks him/herself specifically to find a
contradiction and pose questions, or more globally,
the writer may seek out conflicts in his/her own
thinking, or between ideas/intuitions/commonly
held notions. An example is writing parallel
columns of pros and cons as an aid in deciding
what to include in an essay.

3. Saying What I Really Mean, or WIRMI: "What I

Really Mean Is ••• " WIRMI is a plan writers use
when they want to reduce a body of complex
information to its essential features.

4. Finding a Focus. The protocols showed that writers

don't start with a focus. Instead, they start
with knowledge and goals, and create their
 39

focus. "Creating a focus is one of the crucial

acts that can bridge the gap between generating
ideas and turning them into a paper; it is also
a task with which many writers have trouble."
Many writers feel they can discover a "ready-
made" focus if they search their memories long
enough, rather than try to create a focus.
(Paraphrase from 44-45)

Plans to do: The second category of plans that Flower and

Hayes identify is "plans for producing a paper" (or "plans to

do"). This category also features both procedural plans and

content-specific plans. A procedural plan might be something

like, "I'll make an outline," or "Let's try and write something."

Some content-specific plans include:

1. Forming for Use. Here, the writer pulls an idea out

of her/his original context and reshapes it for the
written text.

2. Organizing. Organizing includes a global plan for

finding a focus, but also subplans for organizing
sections of the paper.

3. Reader-Based Plans. Here, writers consider the

audience and develop strategies based on "what the
reader might assume, object to, or need to know."
Sometimes writers do this by playing the role of
the audience to themselves. (As mentioned
above, many writers ignore this aspect of the
rhetorical situation and consequently produce
writer-based prose.)

4. Product-Based Plans. An example of a product-

based plan is "Here's an interesting point. I
think I'll write a paragraph on this." Product-
based plans occur "when the composing process
is governed by a concern for the form of the
finished product." A product-based plan, if
employed too early in the writing process, may
place too many "rigid constraints" on the writer
and "thwart the dynamics of the normal generating
process." (47-51)
 40

Composing plans: The first two categories of plans--"plans

to say," and "plans to do"--interact in the third category of

"composing plans," or plans controlling the composing act. There

are at least three sorts of interaction. One is "switching

plans." Major switches come when writers run into difficulty.

Another is "mapping plans onto one another." One plan may be

superimposed upon another, or may dictate the formation of the

second. A final interaction is "conflict between plans." A

writer may change a plan in favor of another, sometimes to the

detriment of the final products. An example of conflict between

plans is:

Pat finally entered in a very productive session of

brainstorming, generating a series of ideas and examples
that proved her point. Then suddenly in the middle,
she broke off brainstorming with the comment, "I'm
just listing things. This is a rip." • . • And yet,
these are some of the activities which are essential
to thoughtful and creative generation. (51-55)

While juggling all these types of plans can be difficult,

"working without plans can be more difficult still." Procedural

plans such as the ones described above are not "mysterious or

difficult," but they do entail a knowledge or consciousness about

the thinking/working process of writing and "an awareness of the

useful techniques writers use" (56-57).

In conjunction with their earlier-described work on

"knowledge-telling," Scardamalia and Bereiter maintain that

 41

formulating plans and goals is a form of knowledge, a competency

that children and novice writers just might not have yet:

The general pragmatic abilities of children permit

them to formulate plans for getting to already-
identified and fixed goals. They do not equip them
to engage in reflective planning, where goals take
shape through the planning process.

Further, a reflective planner transforms the writing task into one

of higher complexity, but the knowledge-teller, or novice writer,

transforms the task into one of lower complexity ("Assimilative

168).

Skillful, or reflective planning, is often "opportunistic," a

word employed originally by Barbara and Frederick Hayes-Roth ("A

Cognitive Model of Planning"). "Opportunistic" means that the

writer/planner recognizes when the "attainment of one subgoal

creates the opportunity for attaining another, and so he chooses

and arranges subgoals accordingly." Since opportunities are often

discovered in the course of writing, planning goes on throughout

the composing process. Even top-level goals may be changed in the

course of writing and be subsumed by other goals:

Compositions often turn out in ways unanticipated

by the writer, but this is not an indication of
planlessness. Rather it is an indication of a
very dynamic planning process that keeps adjusting
decisions at every level in the light of decisions
made at other levels. (Scardamalia and Bereiter,
"Recent Research" 789)
 42

In summary, at least three major categories of plans can be

ascertained. They are "plans to say," "plans to do," and

"composing plans." Plans must be changed from ideas into written

language in order to be useful for the writer. This is

accomplished through the act of translating, which will be

discussed next.

Translating:

Translating is "the process of putting ideas into visible

language." Information produced in planning may be represented in

key words or symbol systems other than language, such as imagery.

The writer must translate meaning, stored in key words or imagery,

into a "linear piece of written English" (Flower and Hayes,

"Cognitive Process" 373).

The process of translating requires the writer to juggle a

variety of demands of written English, which lie "on a spectrum

from generic and formal demands through syntactic and lexical ones

down to the motor tasks of forming letters." For children and

novice writers, "the extra burden may overwhelm the limited

capacity of short-term memory." Even so, considering the limited

capacities of any of us--novice ~ expert writers--it is

remarkable that we are able to deal with the following activities

simultaneously in writing, as Carl Bereiter points out:

"handwriting, spelling, punctuation, word choice, syntax, textual

 43

connections, purpose, organization, clarity, rhythm, euphony, the

possible reactions of various possible readers, and so on"

("Development in Writing 80).

Like "the written text," the element of "translating" is one

largely unexplored by composition theorists, probably because of

the lack of information available from protocol analysis. The

next element, however, which is "reviewing," has been the focus of

extensive research.

Reviewing:

Flower and Hayes' model of "reviewing" includes two elements:

"evaluating" and "revising." As discussed in Chapter II,

evaluating is an important element in the problem-solving process.

A positive evaluation, or judgement, usually leads the writer back

into "planning" or "translating." A negative evaluation, however,

leads the writer into "revising." Flower and Hayes have recently

collaborated with other researchers to further expand knowledge of

"revising." Their more recent findings are published in at least

two articles: "Diagnosis in Revision: The Expert's Opinion"

(1985); and "Detection, Diagnosis, and the Strategies of Revision"

(1986). Since both articles are extremely lengthy, only a brief

review of their findings is possible here.

Research suggests that the revision processes of experienced

and novice writers differ because they bring different assumptions

and language to the task. Some students face revision

 ljlj

like the grim reaper, prepared to tramp through a text

cutting and "slashing out." The goal, as they
describe it, is to fix errors rather than rethink;
the primary tool is deletion. In contrast, many
experienced writers conceptualize the task as
discovering content, structure, and voice.
("Detection" 17)

Another difference between experienced and novice writers is that

experienced writers make a good share of their revisions after

rereading their drafts, while poor writers make most of their

changes ~ they are writing. Flower and Hayes suspect that the

better writers are using rereading to "construct a sense of the

text's current gist and/or to form a rhetorical plan" to guide

revision. Yet another difference between experts and novices is

that "novices typically focus on convention and rule-governed

features," or "surface features," while experts use revisions to

alter meaning more often (17-18).

Flower and Hayes point out that our current idea of revision

is as an "extended, discovery-based process (which) reflects the

habits of the novelists, journalists, essayists, and academicians

usually studied." However, studies of writers in business

settings show many "first-time-final drafts." Instructors assume

that revisions make better papers, but:

The amount of rev~s~on is simply not the key

variable • • • • revision as an obligatory stage
required by teachers doesn't necessarily produce
improvement, especially if the writer has put effort
into the planning and writing. Revision, then, is
a strategic action, adapted to the necessities of
the task. ( 18)
 45

Scardamalia and Bereiter prefer the term "reprocessing" when

discussing revision. They say "revision is a special case of

reprocessing, applied to actual text." However, there can be

"internal revision," where writers try to discover and develop

what they have to say even before they write anything down
("Research" 790).

Along those lines, Stephen Witte, author of "Revising,

Composition Theory, and Research Design," says that internal

revision, or what he calls "pre-textual" revision, has been

neglected by researchers. He feels that the concept of revising

should include not only retranscribing but also revising a

pre-text, which is a "mental representation" that can take

"multiple forms--from sensory images, to concepts, to metaphors,

to feelings" (264). Witte suggests that many writers revise their

pre-texts extensively, even before anything is written down. He

grants that the Flower and Hayes model leaves room for

investigating pre-textual revision within its framework, but

suggests that researchers in general have ignored this important

area (278).

The Monitor:

As indicated by its name, the monitor "functions as a writing

strategist which determines when the writer moves from one process

to the next" (Flower and Hayes, "Cognitive Process" 374). Thus,

while writers write, they also monitor their process, deciding

 46

such things as how long to brainstorm for ideas, when to start

writing, and when to evaluate. Decisions made by the monitor are

influenced by the writer's goals and by his/her writing habits or

styles (374).

Summary of the Act of Writing Model

The structure of Flower and Hayes' writing model, as

indicated, consists of the task environment (which includes the

rhetorical problem and the text produced so far); the writer's

long-term memory; and the writing processes of planning

(generating, or "plans to say," organizing, or "plans to do," and

goal setting, or "composing plans"), translating, and reviewing

(evaluating and revising); and the monitor. Certain elements of

the model, such as planning and revising, have been researched

more extensively than others. Many researchers are now beginning

to investigate unexplored elements of the model and beyond. Some

of their findings will now be discussed.

Beyond the Act-of-Writing Model

Flower and Hayes have consistently reminded readers that

their model of the act of writing is a tentative one. Other

researchers have branched into areas unexplored by the model, many

with support and encouragement from Flower and Hayes. Some of

these areas have been those of pre-text, writer's block, the

affective element, and an account of reading along with writing.

 47

Pre-text

Witte, mentioned above, is one example of a researcher

attempting to expand knowledge of the Flower and Hayes model.

Since the publication of his article on revising, Witte has

written "Pre-Text and Composing," in which he expands his concept

of "pre-text" and associates it with other processes in addition

to revision. "Pre-text" refers, he says, to a writer's linguistic

representation of intended meaning. This meaning is tentative, a

"trial locution" produced in the mind, stored in the writer's

memory, and sometimes manipulated mentally before being

transcribed as written text. In this article, Witte confines

pre-text to semantic and syntactic components, such as phrases,

dependent clauses, sentences, or sentence sequences. Pre-texts

differ from more abstract plans like nonverbal images in that

pre-texts approximate written prose. A pre-text represents "the

writer's attempt to instantiate abstract plans and goals in

linguistic forms." Pre-texts are "the last cheap gas" before

writers commit themselves to extended written text" (397-98).

Witte has considerably modified his original definition of

pre-text in this article. He now feels that pre-texts "represent

critical points along a continuum of composing activities between

planning and transcribing written text" (397). Thus, pre-texts

are critical not only to revision, but other elements of the

writing process as well.

 --- --------~~-~---~----- ---------------- --~-------~------

48

Witte poses four observations about the "nature and function"

of pre-text during composing: (1) Pre-text may have direct

influence on written and unwritten text: (2) Pre-text may be

stored in memory in such a way that it may have a delayed, but

direct, effect on the written text; (3) Evaluating and revising

pre-text stored in memory can be based on criteria used to

evaluate and revise written text; and (4) Pre-text can function as

a critical link among written text, translating ideas intb

linguistic form, and transcribing ideas into visible language

(417).

Witte gives credit to Flower and Hayes, as well as

Scardamalia and Bereiter, for implying in their research the

possibilites of pre-text, but considers his research to be further

exploration of this largely ignored concept. Thus, his work can

be seen as an expansion of the Flower and Hayes model.

Writer's Block

Another researcher attempting to expand the Flower and Hayes

model is Mike Rose. Rose investigates the cognitive dimension of

the writing process for answers to the reasons for writer's block,

which he suggests is "one of the least studied dysfunctions of the

composing process" (Writer's Block 1). He defines writer's block

as "an inability to begin or continue writing for reasons other

than a lack of basic skill or commitment" (3).

 49

Rose feels that the Flower and Hayes model is "the most

detailed, multioperational, and comprehensive of composing models

constructed to date." However, he criticizes the model for its

lack of "fluidity." Rose says the model has a "top down,

hierarchically deductive perspective," which views the writer as

working in sequential way from, for instance, generation of ideas

sentence-production. Although Flower and Hayes admit that writing

is recursive and pose the notion of "priority interrupts," Rose

feels that "recursiveness" and "priority interrupts" could be

represented in a less mechanical way (8). These concepts, he

feels, are better represented by the label "opportunism." As

discussed earlier, opportunism suggests that attainment of one

subgoal creates the opportunity for attaining another, thus

subgoals are chosen and arranged accordingly. Sometimes these

decisions follow orderly paths, but sometimes the path for plan

development is disorderly.

Rose feels that the application of opportunism to writing

suggests that the goals, plans, discourse frames, and information

that must be juggled as a writer confronts a task are not always

hierarchically sequenced from most general to most specific.

These goals and plans can influence each other in a variety of

ways. For example:

While editing a paragraph, a writer may see that

material can be organized in a different way 0r as
a writer writes a certain phrase, it could cue
other information stored in memory.
 50

Rose says his scheme is not a complete model, but a "metaphorical

representation" that highlights parts of the composing process

relevant to his study, which is "the relation of writing to

high-level strategies and general problem-solving/

composing orientations, and to rules, plans, and discourse frames"

(8-10) 0

Working within this framework, Rose suggests some possible

reasons for writer's block:

1. The rules writers use for guiding their composing

processes may be "rigid, inappropriately invoked,
or incorrect."

2. Writers' assumptions about composing may be

misleading.

3. Writers may edit too early in the composing process.

4. Writers may "lack appropriate planning and discourse

strategies or rely on inflexible or inappropriate
strategies."

5. Writers may try to invoke "conflicting rules,

assumptions, plans, and strategies."

6. Writers may evaluated their writing with

inappropriate criteria." (Paraphrase from 3-4)

The model, or scheme, that Rose works with, then, is the

combination of the Flower and Hayes model and the concept of

opportunism. Some of Rose's implications for instruction about

writer's block will be discussed in Chapter VI.

 51

The Affective Element in the Writing Process

Because the Flower and Hayes model explores cognitive

variables to the virtual exclusion of affective variables, one

focus of recent research has been devoted to this possible gap in

the "Act of Writing" model. Alice Brand has been particularly

vocal about the exclusion of the affective/emotional element in

the writing process. Two recent articles address this exclusion:

"Hot cognition: Emotion and Writing Behavior" and "The Why of

Cognition: Emotion and the Writing Process" (1987).

"Hot cognition;" says Brand, is essentially "cognition

colored by feeling." What cognitive psychology needs is "some way

to heat up cognition--a theory that unites the cognitively blind

but arousing system of affect with the subtle cognitive apparatus

("Hot Cognition" 1985, 1). Cognitive researchers have not

admitted that "the problem solving that we associate with

conscious cognizing often relies on less-than-conscious,

intuitive, or prelogical thought" (4).

Consequently, says Brand, motivation and emotions such as

"apathy, anxiety, disequilibrium, alienation, despair, and

commitment" are "tucked into corners of work by James Britton and

by Linda Flower and John Hayes to be pulled out when other

explanations fail," even though affect is "repeatedly implicated"

in major human development theories like those of Jean Piaget

("The Why of Cognition" 438). Brand feels that the cognitive

process model of Flower and Hayes "provides no language to deal

 52

with emotion." Because of this lack of affective language, Brand

is concerned about what the model suggests about people:

The cognitive notions of monitors and Operators are

not people but incorporeal automatons. Disembodied
Editors, not humans, detect flaws. Disembodied
Inspectors evaluate performance. These entities have
decision-making powers through us, but they are not
us. We are circuits. We are transitors, fire alarms,
smoke detectors, switching yards, semaphore signals,
radar, and PC's. (440)

This approach, says Brand, encourages writing instructors to

assume that students should behave like the model, assuming a

"flat, uncomplicated objectivity," with no allowance for emotional

and/or motivational factors. Brand suggests that it is time for

cognitive theorists to add the "why of writing" to their

investigation of the "how of writing," in order to make the

theories more complete.

SUsan McLeod ("Some Thoughts about Feelings: The Affective

Domain and the Writing Process," 1987) offers a tentative theory

which fuses the affective element with a problem-solving view of

the writing process. She says that the "constructivist" views

offered by George Mandler in Mind and Body are compatible with the

cognitive theories by Flower and Hayes. Mandler asserts that

"emotional experience consists of two factors, one physiological,

the other cognitive." When an emotion occurs, the automatic

nervous system is activated and results in a response, such as

"the familiar 'gut' response: a knot in the stomach, a quickened

 ----------------- - - ---- --- ---

53

pulse, a heightened awareness of external stimuli." Our cognitive

response is to make sense of this physical agitation, evaluating

it either positively or negatively.

Mandler says that a major source of emotion is "the

interruption of an individual's plans or planned behavior, plans

which have a tendency toward completion." Thus, completing

something interrupted is a "positive, even a joyful experience."

McLeod connects this idea to Flower and Hayes' explanation of

"planning." Flower and Hayes say that plans are interrupted "with

disturbing frequency" by various constraints, so that goals and

strategies must be defined and redefined as writers write. So,

says McLeod, "interruption of plans, a major reason for emotions

to occur, is integral to the writing process. If we agree with

Mandler, we must admit that it its impossible to write without

some emotion occurring" (431-2).

While some writers interpret their agitation resulting from

interrupted plans as negative ("anxious, frustrated, blocked"),

others interpret their agitation as positive ("the excitement of

the chase"). Mandler's theory helps us to understand why students

may be reluctant to revise their written work: "The 'joy of

completion' which comes when an interrupted action is finally

completed is to them a signal of closure." Thus, the problem

solver relaxes and feels finished, even if the final solution is

inadequate (432).
 54

McLeod suggests that writing anxiety, motivation, and student

beliefs about writing can all be studied within this framework

which unites the affective and cognitive domains, thus at least

partially answering Alice Brand's call for a "hot cognition"

approach.

An Account of Writing and Reading

Another area possibly not covered by the Flower and Hayes

model is the link between discourse production and

comprehension, or between writing and reading. Bracewell,

Frederiksen, and Frederiksen say that research in literacy

normally investigates either reading or writing:

This split is characteristic both of theoretical and

empirical research and of instruction in the language
arts, even though one feels intuitively that reading
and writing comprise aspects of the same underlying
literacy skills. (146)

Further, in oral communication, comprehension and production

usually are mixed. Bracewell et al. reason that a theory of the

~·
development of literacy skills needs to treat both the process of

comprehension and production. Thus, their efforts are toward a

"unified and informative comparison of comprehension and composing

processes" (147).

Central to this approach are two categories of discourse

processing: "framing" and "regulating." Framing processes

 55

produce a structure, or frame, for the text. Thus, framing is

closely tied to the "planning" element in Flower and Hayes' model.

An example might be a child's writing that lacks a conceptual

frame that would make the fiction a story, rather than an event

sequence:

A monster comes down to earth, and climbs down

the ladder of the spaceship, and speeds away
in a supercharged car, and sees a witch, and kills
her with a laser gun, and then • • • • (151)

Regulating processes "access language structure, translate

conceptual structure into a text for production, and regulate

construction of the conceptual structure for comprehension" (146).

Regulating processes affect such things as flow, openings,

closing, and topic shifts. Research done by Bracewell et al.

shows that framing and regulating processes are involved in both

production and comprehension; thus, their research links the two

processes in its study of discourse processes.

Bracewell et al. give credence to the Flower and Hayes model,

saying that their research is an attempt to explicate the

relationships within the model, and expand the focus to

comprehension as well as production (151).

Stephen Kucer also explores the relationship between reading

and writing, using some of the terminology put forth by Flower and

Hayes. Kucer states it is unlikely that "language users have two

completely different and independent sets of cognitive mechanisms
 56

for reading and writing" ("The Making of Meaning" 319). Rather,

since both acts are "acts of meaning-making" (318), they probably

share similar cognitive mechanisms, such as long-term memory,

goals, plans, and strategies (320-328). Kucer suggests that by

establishing such "cognitive basics" as those just listed,

teachers should be able to develop literacy curricula which

integrates reading and writing instruction (332-3).

Section Summary

In 1985, Lester Faigley et al. (Assessing Writers' Knowledge

and Processes of Composing) suggested that a new generation of

cognitive research would be centered on issues raised in reaction

to the Flower and Hayes model. The new research would need to

center on such things as the degree to which individuals vary in

composing; the degree to which writing is the reverse process of

reading; the nature and scale of components in composing, and the

relationships among them; where "bottlenecks" occur in writing and

what causes "blocks"; and how learning occurs (8-9). The

above-described research by Stephen Witte, Mike Rose, Susan

McLeod, and Bracewell, Frederiksen and Frederiksen, seems to be a

product of that new generation of researchers. What these

researchers have in common, along with Flower and Hayes and

Scardamalia and Bereiter, is the fact that they are all working

within the cognitive framework of investigation.

 - - - - - - --- --------------

57

There are, however, other researchers and theorists who

disagree not only with the methodology, particularly protocol

analysis, of the cognitive researchers, but also with using the

cognitive framework to investigate the writing process. Some of

their concerns will now be addressed.
 58

CHAPTER V
ISSUES IN
COGNITIVE RESEARCH ON THE WRITING PROCESS

Dissent among composition theorists and rhetoricians

regarding cognitive research seems to revolve around (1) the

validity of protocol analysis and (2) the validity of the

cognitive framework for studying writing. Many dissenters, even

under article titles linked to a critique of protocol analysis,

largely wind up criticizing the theory behind the method, rather

than the method itself. However, here the two facets will be

addressed individually.

Protocol Analysis

Despite strong evidence that under certain situations

protocol analysis is a valuable tool for collecting data about

mental processes, the method is often criticized. Protocol

analysis is "a description of the activities, ordered in time, in

which a subject engages while performing a task" (Hayes,

Problem Solver 51). Three basic kinds of protocols are used by

cognitive researchers to describe psychological processes. They

are "motor protocols," (observers note such activities of the

subject as walking and picking things up); "eye-movement

protocols," (observers note where subjects fix their gaze as they

perform tasks); and "verbal protocols" (subjects are asked to say

aloud everything they think while performing a task, whether or

 59

not it seems trivial to them (52-53). Verbal protocols are most

widely used by those investigating the writing process.

Hayes suggests that while protocol analyses cannot be

expected to be complete recordings of what goes on in the mind,

they can give at least an indication. He likens protocol analysis

to "following the tracks of a porpoise":

Its brief surfacings are like the glimpses of the

underlying mental process which the protocol
affords us. Between surfacing, the mental process,
like the porpoise, runs deep and silent. OUr task
is to infer the course of the process from those
brief traces. (63-64)

Criticism of protocol analysis by composition theorists

ranges from delightfully satirical to more formally challenging.

Robert Gorrell humorously parodies protocol analysis:

If I want to learn about the process of composing a

stew, I can conduct research. I can interview various
stewmakers, for example. I can watch them in action
and take notes on what they do. Or I can put a
microphone in front of them, getting them to recite
what they think goes on in their minds as they peel
turnips or stir the broth or decide whether to
add cumin or oregano. I can check on whether a chef's
hat or apron seems to have much to do with the process.
I can analyze protocols and generalize about procedures
that seem to characterize many of the cooks. I can
perhaps identify different parts of the process--
selecting and preparing ingredients, putting them
together and cooking, tasting the results and
seasoning--preparing, cooking, and revision--invention,
arrangement, style. I can observe that these steps
are not necessarily sequential; that for instance the
cook may be selecting material even at the last stages,
tasting and deciding to use more garlic. ("Mulligan
Stew" 272)
 60

Gorrell's parody suggests that protocol analysis may be a somewhat

trivial procedure for analyzing the obvious.

On the other hand, Marilyn Cooper and Michael Holzman offer

more serious criticism of the protocol methodology employed by

Flower and Hayes: "Protocols, far from being 'extraordinarily

rich in data,' are exceedingly impoverished sources of information

on what writers are thinking about" ("Talking About Protocols"

284). Cooper and Holzman link protocols with "introspection," a

technique used by Wilhelm Wundt in the seventeenth century.

Wundt, regarded by some as the first modern psychologist, issued

complex rules to be followed by those using the introspective

method:

1. The observer must be able to determine when the

process is introduced.
2. He must be in a state of "strained attention."
3. The observation must be able to be repeated several
times.
4. The experiment must be able to be varied through
variations in stimuli. (285)

Wundt used highly trained "introspectors," usually research

assistants, in his investigations. His method was criticized at

the time for limiting observation to "static and relatively simple

mental states." Since the discovery of the unconscious, Wundt's

technique has also been attacked for its inability to give data

about unconscious activities of the mind. Also, many have been

skeptical about the applicability of results obtained from such a

small number of highly trained subjects (284-289).
 61

Cooper and Holzman are among those who remain skeptical about

introspection, and, consequently, protocol analysis. They contend

that it is rather odd for people to talk about what they are

thinking while they are doing something. Only those trained to

think aloud can supply data. Thoughts running through people's

heads are "diffuse, highly branched, visual as well as verbal."

Flower and Hayes' introspectors notice "virtually nothing other

than that which is to the point. Do these people never fantasize

about, say, lunch?" (289-90)

Flower and Hayes, however, object to the comparison of

protocol analysis with introspection. Wundt's method put

"constraints" on what and how subjects observed, asking them to

analyze their observations. For instance, instead of saying, "I

see a book," subjects were asked to report on "Wundtian elements,"

giving statements like, "I see a dark reddish patch shaped like a

parallelogram." The current protocol method, on the other hand,

asks subjects only to "Say whatever comes into your mind."

Another difference is that Wundt's introspectors made "thousands

of practicre observations" before actually participating in

experiments, whereas subjects participating in the protocol method

are purposely given very little training ("Response" 94-95).

Further, studies have shown that while introspection methods

sometimes distort the subject's thoughts, protocol analysis does

not. Ericsson and Simon (Protocol Analysis: Verbal Reports as

Data), ask the question, "Does the act of reporting about mental
 62

processes while they are going on alter the structure of those

processes in a significant way?" Their research indicates that

when subjects are asked to report on specific things while they

perform a task (i.e., introspection), their processes being

observed are likely to be modified. However, when subjects do not

have to report on specific things, but just on what they are

thinking (i.e., protocols), their processes may be slowed

somewhat, but are not altered (Flower and Hayes, "Response"

95-96).

Flower and Hayes admit that protocol analysis is incomplete,

but challenge Cooper and Holzman to come up with something better:

"Any research method has its limitations. That is not news. The

real question is, 'Is there a better method to do the job?'" (97)

In an article published almost two years after Flower and

Hayes' response to Cooper and Holzman, Erwin Steinberg also

defends protocol analysis ("Protocols, Retrospective Reports, and

the Stream of Consciousness" 1986). In response to Cooper and

Holzman's charge that protocol reports are too focused, not

reflecting diffuse and non-verbal thoughts, Steinberg explains

that when one's attention is focused on solving a problem,

"consciousness is more tightly focused" than it is during other

activities, such as daydreaming, in which we attend to other

non-verbal thoughts and images.

 63

Another criticism Steinberg addresses is that the presence of

the observer and tape recorder distorts the process. He says any

methodology has its limits and that in fact

Protocols may not reflect precisely what mental

activity goes on during problem solving; but in
thirty-five years they have enabled cognitive
psychologists and specialists in artificial
intelligence to model problem-solving methods quite
carefully, as a considerable body of literature
demonstrates. (700)

Steinberg also feels many "humanists" are threatened by protocol

analysis because it represents "empiricism," but he argues that

empirical techniques available from the social sciences and

computer science are also useful in humanistic research (710).

Nevertheless, the conflict over protocols goes on. David N.

Dobrin ("Protocols Once More," 1986), for example, criticizes

protocol analysis not because of the method itself, but because of

the model it presupposes. Dobrin's criticisms take us to the

second category of dissent: the use of a cognitive framework for

the study of writing.

The Validity of the Cognitive Framework

Dobrin's major objection to the cognitive model is that it

assumes writing is "a problem-solving symbol-manipulation"

procedure. To claim that the writing process is like solving math

problems, which are well-formulated and have a single solution, is

 64

to contradict our own writing experiences. Dobrin believes that

the cognitive model is invalid because the activity of composing

does not occur "at the focus of consciousness." Instead, during

writing we tend to pause, to detach ourselves, to let our thoughts

roam free, to perhaps look for something "upon the horizon of the

mind, not recognizable, yet in existence" (722).

Dobrin suggests that it is legitimate to interview other

writers about the way they write in order to research the writing

process, and that it is also legitimate to "revive data extracted

from onself." He believes that writing researchers should stop

"trying to lay their work off on the cognitive scientists and

instead listen to 'our own common sense' when it comes to studying

writing" (724).

Others also object to considering writing a problem-solving

process. Ann Berthoff has been vocal in her objections to both

the idea of writing ~ a problem-solving process and the idea of

using writing to solve problems.

Berthoff says that using writing to solve problems, while it

"has the sanction of educational psychologists, systems analysts,

defense intellectuals, and other technocrats because it promises

guidelines, "structures, (aniJ models," keeps us from seeing that

"certain contexts of the 'problem' are themsel'ves problematical."

She believes that problem solving hampers dialectics--questions

which generate answers and then further questions--in favor of

merely solving the "problem area" ("Theory of Imagination" 636-7).

 65

Berthoff also attacks the cognitive framework which considers

writing ~a problem solving process, and calls today's cognitive

theorists "the new brass instrument psychologists" who concern

themselves with "what can be plotted and quantified," not with the

things we truly want to know about the composing process, the

writer's mind, or modes of learning. Instead of allowing the

imagination to "be used and manipulated within the framework of

positivist assumptions," it should be "reclaimed," to use as a

"speculative instrument to focus on what it means to say that

composing is a process of making meaning ("Intelligent Eye"

192-196).

Anthony Petrosky echoes Berthoff's challenge of positivist

assumptions in his review of Linda Flower's textbook, Problem-

Solving Strategies for Writing. Flower, says Petrosky, "succeeds

not in presenting new and useful ideas • • • but in distorting even

the old chestnuts by couching them in narrow notions of mind that

smack of logical positivism." He further charges that the

problem-solving approach "ignores reflective, associative,

metaphoric, intuitive, and imaginative thinking" (233).

Flower, however, states that she has not ignored various

types of thinking:

As a teacher, I see no contradiction at all between

fostering the experience of discovery, of listening
to readers, of reseeing one's own ideas--things we
all value and teach toward--and asking students to
bring a more self-conscious, problem-solving approach
to their writing. ("Response" 96).
 66

Flower believes that the heuristics she discusses are not

positivist rules, but simply function to help students expand

their repertoire of strategies for dealing with the writing

process, whether they are writing a problem-solution paper, a

short story, or a poem. A strategic approach to writing empowers

students with an awareness of their own thinking process and a

sense of options.

Yet others feel that Flower's--and other cognitive

theorists'--emphasis on the writing process neglects

individual processes. Robert Gorrell believes that it neglects

process as it relates to product: "The writing process can be

thought of as a process of ordering and relating ideas, attempting

to achieve hierarchical patterns like some of those that can be

described by analysis of the product" ("Mulligan Stew" 272-277).

Perhaps the best answer to criticisms of t~e cognitive model

is a synthesis of several theoretical perspectives. Lester

Faigley, for example, credits cognitive theories, along with

expressive and social theories, with "giving student writing a

value and authority absent in current-traditional approaches":

Expressive theorists validate personal experience

in school systems that often deny it. Cognitive
theorists see language as a way of negotiating the
world • • • • And social theorists • • • have found
that children who are labelled remedial in
traditional cla·ssrooms can learn literacy skills by
studying the occurrences of writing in the familiar
world around them. ("Competing Theories" 541) ·
 67

As Faigley's synthesis suggests, cognitive research, with its

emphasis on process, illuminates certain aspects of the composing

process. But it may need to be united with other elements of

rhetorical theory which include the study of product. Cognitive

research may indeed isolate "the part from whole," in its focus on

process, but perhaps now the "whole" (process and product) must be

considered, without rejection of knowledge gained from that

necessary isolation of the "part" (542).

Even considering the need to merge cognitive research with

traditional rhetorical theories, writing instructors can benefit

from the pedagogical implications offered by cognitive research.

Next follows a discussion of such implications.

 68

CHAPTER VI
PEDAGOGICAL IMPLICATIONS

Chapter III briefly discussed the concept of using writing to

solve problems. Since the bulk of this paper has been devoted,

rather, to writing~ a problem-solving process, little space will

be used for detailing strategies employing the former concept,

even though there is a substantial body of research on the topic.

Please refer to the bibliography for a list of some of those

articles. Here I will simply describe one strategy offered by

Richard Larson.

In Chapter III, Larson was cited as suggesting that students

use problem-solving techniques to plan arguments on complex

issues. Larson adapts a model of the problem-solving process for

use in the classroom from a course offered at the Harvard Graduate

School of Business Administration. His model includes eight

steps, several of which echo John Hayes' breakdown of

problem-solving steps (in Chapter II):

1. Definition of the problem (note the parallel

to Haye;' category of "problem representation")

2. Determination of why the problem is a problem

3. An enumeration of the goals that must be served

by whatever action is taken

4. Determination, where possible, of the goals

which have highest priority
5. Invention of procedures that might attain the goals
(Hayes' "search" category)
6. Prediction of the results that will follow the
 69

taking of each possible action

1. Weighing of the predictions

8. Final evaluation of the choice that seems superior

(Hayes' category of "evaluating the solution")
( 629-630)

Larson suggests that students become familiar with these steps and

use them in writing about problems. For instance, perhaps the

student wishes to write a letter to a current roommate about a

problem with their living arrangement. The student can use these

steps to structure the letter and plan an argument for a solution.

In addition to planning arguments, Larson suggests that these

problem-solving techniques be used for reflection, discovery and

invention, and for evaluating the arguments of others (635).

But this chapter has more to do with pedagogical implications

for understanding writing as a problem-solving process, whether

the writing be in the form of a problem paper, a poem, a short

story, or an argumentative essay.

If writing instructors are willing to accept--even

·partially--the validity of the cognitive framework for studying

writing as a problem-solving process, then a variety of

pedagogical implications and heuristics are to be gleaned from

research in this area. But translating theory and models into

classroom techniques is not an easy task, for there are still no

definitive rules to be laid out about the composing process. And

there may never be, since the process of writing is so extremely

complex: "There can be no 'Skinner box' studies of the composing

 70

process, for isolating the composer and limiting his or her

activity to one segment of the whole process simply falsifies what

actually goes on" (Contemporary Rhetoric 17).

Therefore, the basis for any instruction "must finally be

subjective and founded upon one's own estimate of results" (16).

This section is just that: an informed but subjective estimate of

how cognitive theorists' findings--about writing as a

problem-solving process--can be applied to the writing classroom.

Some specific heuristics and approaches will be detailed first,

and then some general implications will be discussed.

Specific Heuristics and Approaches

In Chapter III, Linda Flower was cited as saying that she and

John Hayes study writing as a form of problem solving in order to

discover and describe "some of the basic heuristic procedures

which underlie writing, and then to translate these heuristics

into teachable techniques." Since their cognitive model was

detailed in Chapter IV, it is now pertinent to discuss some

specific heuristics that Flower and Hayes have gleaned from their

model. However, first a discussion of the term, "heuristics," as

it relates to the cognitive model of writing, is in order.

 71

Background

Heuristics can be defined as "alternative(s) to trial and

error," or the codification of useful techniques or cognitive

skills employed as discovery procedures or ways of getting to a

goal. Heuristics are not rules, but "alternative methods for

doing something." As mentioned in Chapter I, people often

intuitively use methods to solve problems without realizing what

these methods are. Explicating those intuitive methods enables us

to see them in terms of heuristics.

An example of a planning heuristic for writing offered by

Linda Flower is "Make a plan to do and a plan to say." This

heuristic differs from traditional outlines, or topic-based plans,

in that it urges the writer to set goal-based plans, such as "I

want to not only impress my audience of peers with my knowledge

about electrical engineering, but I also want to come across as

being aware of the ethical problems posed in this field." Once

these types of goals are established, it is easier for the writer

to plan what to say (Problem Solving Strategies for Writing).

Notice that this heuristic takes the writer further than the

traditional instructor's command to "Make sure your writing has a

purpose."

D. N. Perkins (The Mind's Best Work) urges caution in

teaching heuristics. Some of his reasons are: (1) Novices don't

use heuristics as much as they could. Studies show that novices

"have disclosed epidemic disorganization when the going gets

 72

rough," even though they've been taught certain heuristics. (2)

Students may know how to use a heuristic when asked to do so, but

may not think to use it on their own. Teaching organizational

strategies, along with heuristics, can help students remind

themselves of ways to solve the problem before them. (3) People

often change, or modify the heuristics they are taught.

Instructors may be fooling themselves if they think students will

carry intact the set of heuristics from their classroom to other

classes or writing situations in the future. Though he poses

these qualifications, Perkins would not have instructors quit

teaching heuristics. Rather, he would simply have the potential

of heuristics be put in perspective (195-206).

Others argue that the value of heuristics may be minimal.

Irvin Hashimoto states that a student's "cognitive style • • • age,

culture, intelligence, susceptibility to tedium, and beliefs,

attitudes, and intentions" all have an impact on how effective

heuristics will be. Shifts in perspective are not easily made by

most people. Some students are not "intrinsically motivated" to

develop "heuristic craving" ("Structured Heuristic Procedures"

74-76).

Flower and Hayes admit that the experienced writer "comes

equipped with many well-engrained, if counter-productive habits,"

which interfere with adopting new heuristics. They stress that a

heuristic must be presented in the classroom as an experience, one

which will enable the writer to actually learn how to use and
 --------------- ------------------ ------· -------------

73

apply the technique. Rather than simply present a formula, it is

necessary to first change the behavior of the student in order for

the student to assimilate the heuristic ("Problem-Solving

Strategies and the Writing Process" 461).

Virginia Allen Underwood, whose 1980 dissertation surveys

rhetorical heuristics, suggests that certain heuristics should be

used in certain situations, depending on the rhetorical problem.

However, she suggests, composition instructors should resist the

pervasive temptation to simplify the writing process into

overly-prescriptive heuristics (325-8):

We may yearn in vain ••• for a perfect heuristic--

simple, elegant, and complete--to bring bear on
any problem, and in the end we may be left with
Polya's (George Polya, author of How to Solve It:
A New Aspect of Mathematical Method, 1945) two
rules of discovery: "The first rule is to have
brains and good luck. The second rule is to sit
tight and wait until you get a bright idea."
(328)

It is apparent that composition theorists receive the idea of

teaching heuristics with varying degrees of enthusiasm. This

should not be surprising, considering the equally controversial

reception of protocol methods and the use of the cognitive

framework in investigating writing. Still, I hope the reader will

review heuristics generated by cognitive researchers with an open

mind, as the focus below turns to specific pedagogical strategies

that may be taught in the composition classroom.

 ~ -~---------

74

Cognitive-Based Heuristics

Linda Flower's textbook is a good place to start for

identifying specific heuristics that can be taught in the

composition classroom. The text is organized around nine steps

for the composing process, and each step contains strategies, or

heuristics, for writers to employ. The steps and strategies,

which closely parallel the cognitive model discussed in Chapter

IV, are listed below.

Steps and Strategies for the Composing Process

Step 1: Explore the Rhetorical Problem

Strategy 1: Test your image of the problem
Strategy 2: Explain the assignment to yourself

Step 2: Make a Plan

Strategy 1: Make a plan to do and a plan
to say
Strategy 2: Make your goals operational
Strategy 3: Reveal your plan to the reader

Step 3: Generate New Ideas

Strategy 1 : Turn off the editor and brainstorm
Strategy 2: Talk to your reader
Strategy 3: Systematically explore your topic
Strategy 4: Rest and incubate

Step. 4: Organize Your Ideas

Strategy 1: Expand your own code words
Strategy 2: Nutshell your ideas and teach them
Strategy 3: Build an issue tree
Step 5: Know the Needs of Your Reader
-Strategy 1: Analyze your audience
Strategy 2: Anticipate your reader's response
Strategy 3: Organize for a creative reader
Step 6: Transform Writer-Based Prose into Reader-Based
Prose
Strategy 1: Set up a shared goal
 75

Strategy 2: Develop a reader-based structure

Strategy 3: Give your reader cues
Strategy 4: Develop a persuasive argument

Step 7: Review Your Paper and Your Purpose

Strategy 1: Compare your paper to your plan
Strategy 2: Simulate a reader's response

Step 8: Test and Edit Your Writing

Strategy 1: Edit for economy
Strategy 2: Edit for a forceful style

Step 9: Edit for Connections and Coherence

Strategy 1: Transform listlike sentences
Strategy 2: Reveal the inner logic of your
paragraphs

Flower's text is basically an explication of these steps; it

translates the cognitive model of writing into specific heuristics

that writers can use to improve their writing skills. While most

of the steps and strategies listed above are parallel to Flower

and Hayes' cognitive model, a few are not.

Flower notes in her preface that her book attempts to combine

"traditional rhetorical concerns" with "current research in

composition" (vii). Thus, some of the strategies above reflect

traditional rhetorical heuristics. For instance, brainstorming

(Step 3, Strategy 1) has long been a technique taught not only in

composition, but also in other disciplines. "Systematically

explore your topic" (Step 3, Strategy 3) is a compilation of

well-known rhetorical heuristics for invention: (1) Aristotle's

topoi (which she discusses as Aristotle's "topics"); (2) Young,

Becker and Pike's tagmemics, and (3) William Gordon's synectics

(83-85). Flower also borrows methods from traditional audience
 76

analysis in Step 5, Strategy 1, when she suggests that students

analyze the knowledge, attitudes, and needs of their audience, and

also that students distinguish between their primary and secondary

audience (132-133). Steps 8 and 9, which include suggestions for

editing, are also based on well-known, traditional techniques.

The remaining heuristics, however, are based on Flower and

Hayes' cognitive research and are drawn directly from the material

discussed in Chapter IV. (The reader is encouraged to refer to

Flower's textbook for further explanation, which is very

"reader-based," geared well to novice and expert writers.) The

important fact here is that the three major elements of solving

problems that John Hayes discusses (see Chapter II) are visible in

these heuristics. The first element is "problem representation,"

which Flower incorporates into Step 1: "Explore the Rhetorical

Problem." The second element is "search," which is evident in

Steps 2 and 3: "Make a Plan," and "Generate New Ideas." The

third element that Hayes discusses is "evaluating the solution,"

which Flower incorporates into Step 6, "Transform Writer-Based

Prose into Reader-Based Prose," and Step 7, "Review Your Paper and

Your Purpose." A few heuristics from each category merit

explanation.

Problem Representation:

Within the category of "problem representation," Flower

defines a rhetorical problem as a "rather large, uncharted

 77

territory" that contains the writer, the reader, and the writer's

ideas and purpose. She gives the writer three heuristics, in the

form of questions, with which to explore his/her problem

representation. They are:

1. Do I have a mental image of my purpose or goal?

2. Do I have an image of my reader, in particular, an

image of what my reader might need or expect from my
paper?

3. Do I have an image of my own projected self as the

writer--a sense of how I am going to sound in
print?

Flower reminds the writer that the rhetorical problem will only be

a sketch--that details will be filled in and changed as the

process continues. But people only solve the problems they give

themselves to solve, so the picture of the problem should be as

thorough as possible. These heuristics will thus aid in picturing

the problem, or the problem representation (64-66).

Richard M. Coe and Kris Gutierrez ("Using Problem-Solving

Procedures and Process Analysis to Help Students with Writing

Problems") offer an excellent series of assignments designed to

help students with precise problem-representation in regard to

solving their own writing problems. Their three assignments help

students to (1) become more aware of their writing processes, (2)

define their strengths and weaknesses, and (3) set their own goals

for future writing projects--thus employing many of the heuristics

Linda Flower advocates. The assignments are spread over a period

 78

of time. Assignment A asks students to think of something they've

written recently and narrate everything they did from the time

they started to the time they finished. Assignment B asks

students to discuss their strengths and weaknesses by rereading

past assignments with comments for different instructors.

Assignment C builds on the first two, asking students to list

strengths and weaknesses and then establish goals for overcoming

some of their writing problems.

Students writing these assignments may end up setting such

goals as "I will make an issue tree after my first draft so that I

can see how my paper needs to be reorganized, in order to combat

my problem of disorganization," or "I will proofread my paper

several times before I hand it in, reading once for certain

misspelled words that I normally have problems with, once for

subject-verb agreement, and once to make sure I'm not

inappropriately switching verb tenses." Thus Coe and Gutierrez's

assignments can be useful as an exercise in problem representation

and solution, using such heuristics as improving problem

representation, making operational plans, and setting evaluation

criteria.

Search:

For the category of "search," Flower offers several

heuristics, such as the "Make a plan to do and to say" heuristic

 79

discussed above. Other search heuristics she suggests, in

addition to brainstorming, Aristotle's topoi, and synectics, are

to (1) talk to the reader, and (2) "rest and incubate." Since

people come up with some of their best ideas when engaging in a

face-to-face discussion with another person, talking to one's

reader can be a powerful strategy. If the reader is unavailable,

the writer can role-play, trying to anticipate the reader's

reactions and questions (82). Resting and incubating involves

formulating an unsolved problem, and then getting away from

thinking about it for awhile. Thus, starting a paper early is a

good tactic (86).

Another search heuristic is the "fractionation" method

discussed in Chapter II. Here, the writer breaks the task into

parts, rather than being overwhelmed by the entire task at once.

Thus, the task may be broken into such parts as research,

brainstorming, organization, writing a rough draft, and revising.

Some of these tasks may even be broken further. For instance,

revision may be split into such tasks as "evaluating," and

"diagnosing."

Again, search heuristics like "talk to the reader," "rest and

incubate" and "break the task into parts" have been gleaned from

cognitive research on expert and novice writers--they are

heuristics ·that work for many experts.

 80

Evaluation:

A heuristic offered for the category of "evaluation" is that

of transforming writer-based prose into reader-based prose. As

discussed in Chapter IV in conjunction with long-term memory,

writer-based prose "faithfully reflects the writer's own discovery

process and the structure of the remembered information itself,"

but fails to make the transformation or reorganization necessary

for the needs of the reader. Flower suggests that writers thus

transform writer-based prose into reader-based prose as part of

their evaluation process. Specific heuristics for accomplishing

this task include (1) setting up a shared goal, or finding a

reason for writing the paper and a reason for reading it; (2)

developing a reader-based structure, which may entail a more

logical organization; (3) giving the reader cues, such as

previews, summaries, and guides; and (4) developing a persuasive

argument, which includes serious thought about the reader's point

of view (160-180).
Roland Huff ("Teaching Revision: A Mode 1 of the Drafting

Process") employs many of Flower's evaluation heuristics--and

more--in his discussion of how to teach the drafting process to

students. Huff suggests a three-stage drafting process: (1)

zero-drafting (the student initially discovers a topic): (2)

problem-solving drafting (problems with the draft's conception and

organization are identified and resolved); and (3) final drafting

(students attempt to arrive at the best solution for their

 81

rhetorical problem). While these stages are presented

sequentially, Huff's intent is not to "define drafting as a

three-stage process" (802), but to help students develop an

awareness of strategies needed.

For each draft, Huff offers specific heuristics. For

instance, for the problem-solving draft he has students answer

such heuristic questions about audience as: "How old are members

of this audience?" "Where do they live?" "How well-educated are

they?" About the writer's stance, he asks, "How well-informed is

the writer about this subject?" "How well-educated in general?"

"How is the writer dressed?" Thus, each draft the student writes

is aided by heuristics like the above.

Indeed, some may well question what's so new about heuristics

such as Huff's and Flower's listed above. The heuristics may seem

sensible and even trite. Some people already employ many or most

of them. The key is that most of these heuristics are already

employed by expert writers, but not by novice writers. In other

words, the findings may simply verify what "good" writers and some

writing instructors have already intuited. Some intructors may

have been teaching some of these heuristics for years on their

own, without ever having heard of Linda Flower, and these

instructors should feel validated by Flower and Hayes' research.

Probably more teachers, however, have used some of these

heuristics in their own writing, without ever having

 82

conceptualized them in such a way as to be useful for their

students, who are novice writers.

For the instructor who feels frustrated by the lack of

teaching tools in the writing classroom, for those who feel

something lacking in the traditional modes of description,

exposition, narration, and argument, Flower's heuristics may be a

welcome teaching tool.

Personal Commentary:

I have had positive experiences with Flower's heuristics the

past two semesters in my freshman composition classroom at Iowa

State University. What I appreciate most about the heuristics is

that they provide a common language with which to work from. For

example, I consider it a powerful tool to be able to say to

students, "This needs to be more reader-based," knowing that they

will understand that I want them to set up shared goals and give

the reader cues, rather than be puzzled, and perhaps frustrated,

by a directive like "This needs coherence."

I have several goals when using Flower's text. First, I have

students read the entire text by mid-semester, with rereadings of

sections afterwards. This ensures time to incorporate a large

variety of heuristics in various writing assignments. Second, I

have students experience the heuristics offered by Flower, through

small-group activities or individual conferences, though some of

the heuristics are best illustrated with the class as a whole. If
 83

the student views an untried heuristic as very awkward or

unwieldy, s/he may not attempt it on her/his own or in a small

group. For example, the heuristic of building an issue tree seems

complicated to some, and may be more productively experienced in

the class as a whole. Third, I try to help students understand

that they will probably pick, choose, and modify the heuristics

they learn in my class. Fourth, I attempt to help students

realize that they can apply the heuristics they learn and modify

in my class to other writing situations, whether academic,

business-related, or otherwise. Thus, I try to aid students in

internalizing the heuristics by basing my evaluations and grades

on whether or not they use them in their rough drafts and

revisions.

My students' feedback regarding the text itself varies. Some

find it engaging, while others find it fairly tedious to read.

However, most feel that the strategies they learn are worthwhile,

and help them in their writing. Some have suggested that I teach

the strategies without requiring them to buy and read the book.

Since I also use a reader in my classroom, I have been tempted to

teach Flower's heuristics without the text, in order to lower the

student's cost of buying two texts. However, I have been

reluctant to discard Flower's text because--and I stress this to

my students--the book can be so helpful to them beyond my class.

Some students have objected to being put through the paces of

so many different strategies/heuristics. Perhaps they are the
 ---·-~--- ---- ----------- ---------- --- ------------ ----- ----~- --~ - ---

84

ones .who don't develop the "heuristic craving" that Hashimoto

talks about. But there is the danger of instruction becoming too

heuristic-oriented, thus interfering with the student's own

strategies, which may be good ones already. An instructor may

need to loosen up his/her heuristic hold on the classroom at

times, but a looser, cognitive-based approach can still be

effective. Some of these more general approaches will now be

discussed.

Cognitive-Based Approaches

Jack Selzer ("Exploring Options in Composing") warns that

teachers, in the rush to apply basic research on composing to the

classroom, often "attempt to impose a single, 'ideal' composing

style on to their students" (276). Further, teachers fail to

acknowledge that individual students may compose differently in

different situations. For instance, students "may plan and revise

freshman essays only superficially" (281) because they consider

the task unimportant, but may write several drafts of a letter for

a job application, which may be more important to them.

Selzer suggests several ways for instructors to help expand

their students' composing repertoires. First, they can expose

students to a "variety of composing styles and actions," through

descriptions of how other writers have responded to various

writing tasks. Journals kept by the students about their writing

processes, with results shared in the classroom, might help

 85

illustrate the differences in processes. Second, students need to

experience composing options themselves through a range of

assignments, such as "in-class essays and exercises, personal

letters, lengthy formal essays, and the kinds of writing required

in various disciplines" (282).

Certainly Selzer's concerns need to be heeded by instructors

teaching heuristics such as the ones Linda Flower offers. It is

important not to over-prescribe, to remember that writing is not a

well-defined problem, but an ill-defined one, not solvable by

algorithms. Heuristics provide no guarantee of a solution, and

instructors should not teach them as if they do.

In addition to Selzer, other researchers' works offer

implications for approaches in the writing classroom. Research on

writer's block, knowledge-telling, pre-text, the affective domain,

the link between reading and writing, and creativity (all

discussed in Chapters II and IV), yields information that

instructors may find pedagogically useful.

Writer's Block:

When students apply heuristics to their writing as if they

are algorithms, writer's block may well occur, as discussed in

Chapter IV. Mike Rose suggests that rather than freeing up

"stymied writers" with a list of techniques and tricks,

instructors need to investigate students' processes, as well as

their products, in order to discover what inflexible rules are
 86

causing the block. He suggests several ways to gain entry to

process. One is to have students bring every scrap of paper used

for a recent assignment and question them about previous writing

experiences, then more specifically about their written work.

Their answers may yield important information about rules that

they may have internalized. Second, students may compose in front

of the teacher, with the teacher asking questions immediately

after. Third, students may be asked to do their own protocol

analysis, with teacher and students together pinpointing specific

process problems. Once information is gathered, students'

problematic assumptions can be remedied by "conferences,

tutorials, lessons, readings, CAI programs, and small-group work"

( 84-87) 0

Knowledge-Telling:

In addition to the possibility of heuristics being used at

the wrong time in the writing process, heuristics may not be

beneficial for overcoming the knowledge-telling behavior discussed

in Chapter IV. Recall that knowledge-tellers reduce writing

assignments to "topics," with no explicit goals of their own for

the assignment except for telling what they know about the topic.

Goal-related knowledge, say Scardamalia and Bereiter, seems to

come rather late in writing development ("Assimilative Processes

165-170). Verbal strategy descriptions, or heuristics, may not

 87

help the knowledge-teller, because they presuppose a "high level

of metacognition" which allows one to assess and manipulate one's

cognitive strategies. Further, heuristics presuppose an "existing

executive strategy" already developed enough to incorporate the

heuristics (Scardamalia and Bereiter, The Psychology of Written

Composition 252).

Rather than teaching heuristics to combat knowledge-telling,

Scardamalia and Bereiter suggest directing instructional effort

toward "self-regulatory mechanisms," which may eventually result

in the student acquiring heuristics. Basically, a self-regulatory

mechanism is a type of a building block, which allows for later

internalization of specific heuristics or strategies. An example

of a self-regulatory mechanism is "goal concretization," where the

instructor substitutes "goals of a more concrete and stable type

than those naturally occurring in compositional tasks" (252-253).

For instance, the instructor may give an ending sentence and ask

students to write compositions leading up to the specified ending

sentence, which provides the concrete goal for the students (259).

Another example of supplying a concrete goal is shown in the

following description of an assignment:

Students are grouped into teams of three to discuss

interesting personal experiences. They choose one
person's experience, discuss it in detail, and each
student goes off to write about it as if it happened
to him or herself. The teacher may remind students
that their stories should contain no real names. Each
student in a team reads his or her story aloud and the
class tries to guess which of the three stories was
 88

written by the person the experience actually happened

to. The writer succeeds by convincing the class that
his or her narrative is the true account. (260)

The concrete goal in this example is to produce a fictitious

narrative that will pass for a true account. A more abstract and

intangible, and therefore less reachable, goal would be to "write

a realistic narrative." This goal does not provide the concrete

goal that supplying an ending sentence does. Providing

knowledge-telling students with concrete goals brings the students

closer to "the full act of dealing with emergent goals" (260).

Though Scardamalia and Bereiter's research focuses mostly on

students in elementary and secondary schools, certainly college

instructors are familiar with "knowledge-telling" and can benefit

from their suggestions for helping students move beyond it.

Pre-text:

More teaching approaches can be gleaned from the work of

other theorists discussed in Chapter IV as well. For instance,

Stephen Witte's findings on pre-text may as yet yield no specific

strategies, but his work helps to qualify the idea that heuristcs

for revision should happen toward the end of the composing

process. Recall that pre-text is "the writer's tentative

linguistic representation of intended meaning" which is stored in

the writer's memory and sometimes manipulated mentally prior to

being written down. Witte suggests that students should
 89

understand that they may have a powerful medium with which to

explore composing problems "without the added problem of juggling

all the constraints of formal written prose" (417). Implicit in

Witte's research is the idea that instructors can encourage

students to spend time thinking about their drafts before they

actually write. Activities might be designed which have students

talk their first drafts into a tape recorder, or with another

student, in an effort to acquaint students with the concept and

power of pre-text.

Affective Domain:

Susan McLeod's work with the affective domain in writing also

has implications for instruction. McLeod suggests that students

become aware that most writers are "agitated" as they compose, and

that students can "learn to find that agitation enabling rather

than debilitating." Further, students should realize that the

feeling of pleasure they have when a first draft is finished is

not necessarily an indication that the entire task is done. Says

McLeod:

We can work out specific coping strategies to help

students control their affective reactions--monitoring
their emotional state, allocating their energy, stopping
themselves when they are over-excited--so that their
emotions work for them rather than against them. (433).
 --- --------------~--------··- -----·------------- ·- --------------------

90

In order to include the affective domain in composition

instruction, writing instructors could have students monitor their

feelings of agitation, pleasure, anxiety, and excitement as they

write a paper. These feelings could then be discussed and

compared in class, which could help students to be more aware of

the affect their feelings have on their writing.

Reading and Writing:

Research by those such as Bracewell, Frederiksen, and

Frederiksen and Kucer on the link between discourse production and

comprehension, or between writing and reading, is still in a very

theoretical stage. Thus, pedagogical implications are sparse. At

the very least, their research implies what most teachers are

already aware of--that writing cannot escape the context of

reading. Therefore, instruction in writing must also involve

instruction in reading. This implies that writing instructors

must be aware of research done in both writing and reading.

Creativity:

Cognitive research in creativity can also shed light on

composition instruction. In answer to the question, "Can an

individual become more creative?" John Hayes says there are at

least three things that people can do to increase their chances of

being creative. The most difficult is to develop knowledge bases
 91

in the fields they are interested in. He also recommends that

people "create the right atmosphere for creativity." This

involves such techniques as brainstorming or searching for

analogies. Because people "often fail to notice that a new

problem is really an old problem" that they already know how to

solve, they may spend "unnecessary time solving it anew."

Analogies can help people recognize that they are solving old

problems. Hayes, like Flower, cites "synectics" as a way to

invent analogies, with its classification of analogies into four

categories: personal, direct, symbolic, and fantasy (237-240).

D. N. Perkins is another researcher who has focused on

cognitive theories of creativity. Author of The Mind's Best Work,

Perkins has used protocol analysis to study the creative processes

of various artists. He offers a set of heuristics for those

interested particularly in the invention stage of creativity.

Following is a paraphrase of his advice:

1. Try to be original.
2. Find the problem. Explore the alternatives, and be
flexible.
3. Strive for objectivity.
4. Search as necessary and prudent. Explore
alternatives when the present routine fails, or
even explore alternatives routinely.
5. Try, to don't expect, to be right the first time.
You may have to "satisfice" (or settle for something
that is "good enough").
6. Make use of noticing. Use your ability to notice
patterns which are relevant to a problem.
7. When stuck, change the problem. "Any problem can be
solved--if you cange the problem into a related one
that solves the real problem."
8. When confused, use concrete representations. Making
 92

thoughts concrete can help to clear confusion.

9. Practice in a context. Practice creativity with an
activity you often undertake and try hard to be more
creative in that.
10. Invent your behavior. Criticize, revise, and devise
ways you do things important to you. (214-219)

Again, many of these heuristics employ the previously identified

important elements of problem-solving: problem representation,

search, and evaluation.

In cognitive research, writing instructors have at their

disposal some exciting and concrete heuristics and promising

approaches with which to approach the classroom. These may be

methods that some instructors have already intuited and used in

their classrooms; for most, however, these methods are a viable

alternative to the "Here, let me do it," approach that many, for

lack of a better method, find themselves using in conferences and

written evaluations.

General Implications

Norman Frederkisen says, "The primary missions of educational

institutions, from elementary to graduate and professional

schools, are to impart knowledge and to teach cognitive skills.

One of the most important cognitive skills is no doubt

problem-solving ability" (Frederiksen 363).

Frederiksen says that instructors do know enough to teach

specific skills such as the use of algorithms in mathematics, but

 93

in the area of ill-structured problems, instructors know little

about how to teach students to develop representations of

ill-structured problems, to make plans for solving such problems,

or to use the appropriate strategies or heuristic approaches

(396).

Much needs to be learned about how ill-structured problems

are solved, however, because

Many of the problems we face in real life, and all

the important social, political, economic, and
scientific problems in the world, are ill-structured
(Simon, 1973). Schools seldom require students to
solve such fuzzy problems--problems that are not
clearly stated, where the needed information is not
all available, there is no algorithm, and there may not
be a single answer that can be demonstrated to be
correct. (363)

Frederiksen feels that methods of solving ill-structured problems

do exist, and can be "taught in such a way that they can be used

in new domains where they are relevant" (363).

Kenneth Pfeiffer, Gregory Feinberg, and Steven Gelber agree

that these methods do exist, but say that "there is little

evidence that general problem-solving skills have been taught

effectively" ("Teaching Productive Problem-Solving Attitudes" 99).

Their hypothesis is that "researchers have focused too much

attention on what to teach students to make them better problem

solvers, instead of how to teach students" (100). They feel that

educators need to think less in terms of causing specific changes

in the student, "as though there (isj a direct, one-on-one

 94

relationship between what is taught and what is learned," and more

in terms of facilitating "practice and active participation by the

student" (99-100). (There is a clear parallel here to teaching

writing as process, rather than product.) They use the analogy of

riding a bicycle: it requires active participation by the

student, with a lot of practice, rather than sitting passively,

trying to learn how from the teacher. Thus, they feel that

"general problem solving can be taught if the teacher provides not

only facts, but also practice, feedback, and motivation" (102).

The teacher needs to play the role of a coach, or motivator and

facilitator, rather than merely a giver of knowledge.

If instructors "buy" the idea that writing is a

problem-solving process, explorable by methods such as protocol

analysis, and that the writing process (not that there is only one

process, but a rough sketch of elements common to various writing

processes) can be conceptualized into a type of a cognitive model,

then there are at least a few general implications that can be

made about teaching writing. Those buying into this idea

certainly must be aware of the controversy that surrounds

cognitive research into such a complicated mental process such as

writing; that controversy has been touched upon in Chapter v.

Buyers must also be aware that a cognitive model, such as the one

offered by Flower and Hayes, may look deceivingly simple--that the

intent is not to show writing to be a stage-like process,

beginning with one step and ending with another. Rather, writing
 --------~ ~~--~- ~~~~~~~

95

consists of a complicated set of thinking processes which have a

hierarchical organization based on processes embedded within other

processes. Writing is a "goal-directed thinking process, guided

by the writer's own growing network of goals." These goals may be

L. changed or substituted by new goals as the process continues

(Flower and Hayes, "Cognitive Process" 386-7).

If instructors buy this parcel of cognitive goods, then

certain implications become apparent. First, the concept that

writing is a mysterious process and thus unteachable may not be

accurate. This concept was influenced by the Romantic movement

and is held at some level, implicitly or explicitly, by many

instructors today. Rather, it is possible that while writing may

not be an algorithmic process which is teachable through strict

rules, there may be some basic, teachable heuristics to use for

instruction. Thus, writing, or at least certain elements of

writing, is teachable.

Second, results of cognitive research of writing as a

problem-solving process imply that information about the writing

process can be used as subject matter in the classroom. This is

not to say that traditional literature subject-oriented classrooms

will be a thing of the past. Reading literature and writing about

it will always be a thought-provoking, viable activity in the

classroom. However, cognitive research enables instructors to

discuss and present writing as a process, and to design activities

(such as those employing heuristics) around that process. Most
 96

instructors are at least vaguely aware of the "process" theory of

writing and use texts which give lip service to the stages of

"prewriting," "writing," and "rewriting." But this stage theory

is not the same concept as the cognitive research model.

Cognitive research goes several steps beyond the stage theory in

its assertion that writing is not orchestrated in stages, but

recursively, embedded hierarchically. Thus, cognitive research

gives instructors a subject and focus to talk about regarding the

writing process itself.

Third, cognitive research gives instructors something to do

besides talk in the classroom. The traditional method of

lecturing has long been criticized, but some writing instructors

have been at a loss of what to do instead. Again the analogy of

learning to ride a bicycle is useful in teaching problem-solving

attitudes. Learning to ride a bicycle requires active

participation by the student, with a lot of practice, rather than

sitting passively, trying to learn how from the teacher. The

writing instructor, like the bicycle instructor, needs to play the

role of a coach, or motivator, rather than merely a giver of

knowledge. This is possible when the instructor uses heuristics

derived from cognitive research to "coach" with, such as the ones

described in this chapter. In fact, it is more helpful for the

instructor to encourage the students to experience the heuristics

than it is to merely present them in lecture-fashion.

 97

Happily, this approach meshes well with the "environmental

mode," or approach, that George Hillocks, Jr. asserts is most

effective in the composition classroom, in contrast to several

other approaches. Hillocks uses meta-analysis to compare results

of countless experiments and research on composition. His 1984

article, "What Works in Teaching Composition: A Meta-analysis of

Experimental Treatment Studies," and his more recent book,

Research ~Written Composition: New Directions for Teaching

(1986) contain a detailed discussion of the process of

meta-analysis. In both these works, Hillocks identifies and

compares the effectiveness of four modes of instruction. They are

presentational, natural process, environmental, and

individualized.

The presentation mode is perhaps most familiar to those who

have been educated in a traditional college or university. It can

be characterized by the following: clear and specific objectives,

lecture and teacher-led discussion dealing with concepts to be

learned and applied,·the study of models, assignments that usually

involve imitating a pattern, and feedback coming mostly from the

instructor. Hillocks found that presentation treatments in his

study were three times more frequent than any other ("Teaching

Composition" 143).
The natural process mode includes: general objectives, free

writing as a way of exploring a subject, writing for peers as

audience, generally positive feedback from peers, frequent

 98

opportunities for revision, and frequent student interaction. In

this mode, there is a low level of structure, and the instructor

is generally nondirectional about the qualities of good writing

(143).

Characteristics of the environmental mode include clear and

specific objectives, materials and activites aimed at engaging

students with each other in specific processes important to

particular aspects of writing, and activities conducive to high

levels of peer interaction on specific tasks. In this mode,

teachers minimize lecture and teacher-led discussion, usually

having students work on specific tasks in small-groups and then

independently. Principles are not just announced and illustrated

by the instructor; rather, the students are engaged in working

through them. Hillocks suggests that while the presentational

mode emphasizes the role of the teacher as presenter of knowledge

about writing, and the natural process mode emphasizes the student

as the generator of ideas, the environmental mode "appears to

place teacher and student more nearly in balance"; thus the term,

"environmental" (144-145).

The individualized mode is fairly self-explanatory. Here,

students are instructed on an individualized basis through

tutorials or programmed materials (or a combination of both) on

such things as mechanics, researching, planning, and writing

papers (146).
 99

Through his meta-analysis, Hillocks found that "the

environmenta+ mode is responsible for higher gains than the other

modes." The differences (based on pretest-to-posttest results),

among the presentational, natural proces, and individualized modes

are not significant, but the environmental gain is three times the

gain for the others (147-9). Hillocks says,

In contrast to the presentational mode, (the

environmental mode) places priority on high levels of
student involvement. In contrast to the natural
process mode, the environmental mode places high
priority on structured problem-solving activities
(emphasis mine), with clear objectives, planned to
enable students to deal with similar problems
in composing. (160)

Thus the environmental mode, which proves, at least according to

Hillocks' study, to be the most effective mode of instruction, is

also the mode which places high priority on problem-solving

activities, of which the writing process is one. The

environmental mode, combined with teaching writing as a

problem-solving process, can give instructors something to do

besides lecture to a class full of half-interested students.

If instructors believe that at least parts of the writing

process can be taught in terms of problem-solving, and if they

have tools, or heuristics, with which to teach that are based on

legitimate cognitive research, and if they are freed from the

lecture-based approach with which they were most likely taught,

then perhaps writing instructors will no longer speak quite so

 100

often of working "in the trenches" as they teach composition

courses.
 -------- -------~--------

101

WORKS CITED

Bereiter, Carl. "Development in Writing." Cognitive Processes

in Writing. Ed. Lee Gregg and Erwin Steinberg. Hillsdale,
NJ: Lawrence'Erlbaum, 1980. 73-93.

Berthoff, Ann E. "From Problem-Solving to a Theory of

Imagination." College English 33.6 (1972): 636-649.

"The Intelligent Eye and the Thinking Hand." In The

Writer's Mind: Writing as a Mode of Thinking. Ed. Janice
N. Hayes, Phyllis A. Roth, John R. Ramsey, Robert D. Foulke.
Urbana, IL: NCTE, 1983.

Bracewell, Robert J., Carl H. Frederiksen, and Janet D.

Frederiksen. "Cognitive Processes in Composing and
Comprehending Discourse." Educational Psychologist
17.3 (1982) 146-64.

Brand, Alice. "Hot Cognition: Emotions and Writing Behavior."

Journal of Advanced Composition 6 (1985): 236-277.

Brand, Alice. "The Why of Cognition: Emotion and the Writing

Process." College Composition and Communication 38.4
(1987): 436-443.

Coe, Richard M. and Kris Gutierrez. "Using Problem-Solving

Procedures and Process Analysis to Help Students with
Writing Problems." College Composition and Communication
32.3 (1981): 262-271.

Contemporary Rhetoric: A Conceptual Background with Readings.

Ed. w. Ross Winterowd. New York: Harcourt, Brace,
Jovanovich, 1975.

Cooper, Marilyn, and Michael Holzman. "Reply." College

Composition and Communication 36.1 (1985): 97-100.

"Talking About Protocols." College Composition and

Communication 34.4 (1983): 284-93.

Dobrin, David N. "Protocols Once More." College English

48 (1986): 713-25.

Emig, Janet. The Composing Processes of Twelfth Graders.

Champaign: NCTE, 1971.
·~-

102

Ericsson, K. A., and H. A. Simon. Protocol Analysis: Verbal

Reports as Data. Cambridge, MA: MIT Press, 1984.

Faigley, Lester. "Competing Theories of Process: A Critique

and a Proposal." College English 48 ( 1986): 527-420

Faigley, Lester, Roger D. Cherry, David A. Jollife, Anna M.

Skinner. Assessing Writer's Knowledge and Processes
of Composing. Norwood, NJ: Ablex, 1985.

Feldhusen, John, and Virginia Guthrie. "Models of Problem

Solving Processes and Abilities." Journal of Research
and Development in Education 12.2 (1979): 22-32.

Flower, Linda. Problem-Solving Strategies for Writing.

San Diego: Harcourt Brace Jovanovich, 1985. 2nd ed.

"Response to Anthony Petrosky, 'Review of Linda

Flower, Problem-Solving Strategies for Writing."'
College Composition and Communication 35.1 (1984): 96-7.

"Writer-Based Prose: A Cognitive Basis for Problems

in Writing." College English 41. 1 ( 1979) : 19-37.
Flower, Linda, and John R. Hayes. "A Cognitive Process Theory
of Writing." College Composition and Communication
32 (1981): 365-386.

"The Cognition of Discovery: Defining a Rhetorical

Problem." College Composition and Communication 31
(1980): 21-32.
"The Dynamics of Composing: Making Plans and Juggling
Constraints." Cognitive Processes in Writing. Ed. Lee
Gregg and Erwin Steinberg. Hillsdale, NJ: Lawrence
Erlbaum, 1980. 31-50.

"Identifying the Organization of Writing Processes."

In Cognitive Processes in Writing. Ed. Lee Gregg and
Erwin Steinberg. Hillsdale, NJ: Lawrence Erlbaum, 1980.
3-30.
"Plans that .Guide the Composing Process." In Writing:
The Nature, Development, and Teaching of Written
Communication. Ed. Carl H. Frederiksen and Joseph F.
Dominic. Vol. 2. Hillsdale, NJ: Lawrence Erlbaum, 1981.
39-58.
 103

"The Pregnant Pause: An Inquiry into the Nature of

Planning." Research in the Teaching of English 15.3
(1981): 229-2 3.

"Problem-Solving Strategies and the Writing Process."

College English 39.4 (1977): 449-461.

"Response to Marilyn Cooper and Michael Holzman,

'Talking About Protocols.'" College Composition and
Communication 36.1 (1985): 94-97.

Flower, Linda, Linda Carey, and John R. Hayes. "Diagnosis in

Revision: The Expert's Option." CDC Technical Report No.
27. Pittsburg, PA: Carnegie-Mellon Univeristy
Communications Design Center. November, 1985.

Flower, Linda, John R. Hayes, Linda Carey, Karen Schriver, and

James Stratman. "Detection, Diagnosis, and the Strategies
of Revision." College Composition and Communication
37.1 (1986): 16-55.

Frederiksen, Norman. "Implications of Cognitive Theory for

Instruction in Problem Solving." Review of Educational
Research 54.3 (1984): 363-407.

Gardner, Howard. Art, Mind, and Brain. New York: Basic

Books , Inc • , 1982 •

Gorre 11 , Robert M. "How to Make Mulligan Stew: Process and

Product Again." College Composition and Communication
34.3 (1983): 272-77.

Hashimoto, Irvin. "Structured Heuristic Procedures: Their

Limitations." College Composition and Communication
36. 1 ( 1985) : 73-81.

Hayes, John R. Cognitive Psychology: Thinking and Creating.

Homewood, Illinois: The Dorsey Press, 1978.

The Complete Problem Solver. Philadelphia, Penn.:

The Franklin Institute Press, 1981.
Hayes-Roth, Barbara, and Frederick Hayes-Roth. "A Cognitive
Model of Planning." Cognitive Science 3 (1979): 275-310.

Hillocks, George, Jr. Research on Written Composition:

New Directions for Teaching. Urbana, Il: National
Conference on Research in English and ERIC Clearinghouse
on Reading and Communication Skills, 1986.
 104

Huff, Roland K. "Teaching Revision: A Model of the Drafting

Process." College English45.8 (1983): 800-816.

Kucer, Stephen L. "The Making of Meaning: Reading and Writing

as Parallel Processes." Written Communication 2.3 (1985):
317-336.

Larson, Richard L. "Problem-solving, Composing, and Liberal

Education." College English 33.6 ( 1972): 628-635.

"The Writer's Mind: Recent Research and Unanswered

Questions." The Writer's Mind: Writing as a Mode
of Thinking. Ed. Janice N. Hays, Phyllis A. Roth,
Jon R. Ramsey, Robert D. Foulke. Urbana, Ill: NCTE,
1983.

Lauer, Janice. Seminar on Modern Composition Theory. West

Lafayette, Indiana: Purdue University, June, 1987.

Lunsford, Andrea. "Cognitive Studies and Teaching Writing."

Perspectives in Research and Scholarship in Composition.
Ed. Ben w. McClelland and Timothy R. Donovan. New York:
The Modern Language Association of America, 1985. 145-161.

Mandler, George. Mind and Body: Psychology of Emotion and

Stress. New York: Norton, 1984.

McLeod, Susan. "Some Thoughts about Feelings: The Affective

Domain and the Writing Process." College Composition
and Communication 38.4 (1987): 426-435.

Newell, Allen, and Herbert A. Simon. Human Problem Solving.

Englewood Cliffs, New Jersey: Prentice-Hall, 1972.

Perkins, D.N. The Mind's Best Work. Cambridge, Massachusetts:

Harvard University Press, 1981.

Petrosky, Anthony R. "Review--Problem-Solving Strategies for

Writing." College Composition and Communication 34.2
.
(1983): 233-235 •
Pfeiffer, Kenneth, Gregory Feinberg, and Steven Gelber.
"Teaching Productive Problem-Solving Attitudes." In
Applications of Cognitive Psychology: Problem Solving,
Education, and Computing. Ed. Dale E. Berger, Kathy
Pezdek, William P. Banks. Hillsdale, New Jersey:
Lawrence Erlbaum Associates, 1987.
 105

Rose, Mike. "Rigid Rules, Inflexible Plans, and the Stifling

of Language: A Cognitivist Analysis of Writer's Block."
College Composition and Communication 31.4 (1980):
389-400.

Writer's Block: The Cognitive Dimension.

Carbondale: Southern Illinois University Press, 1984.

Scardamalia, Marlene, and Carl Bereiter. "Assimilative Processes

in Composition Planning." Educational Psychologist 17.3
( 1982): 165-71.

"Research on Written Composition." Handbook of Research

on Teaching. 3rd ed. Ed. M.C. Wittrock. New York:
Macmillan, 1986. 778-803.

Selzer, Jack. "Exploring Options in Composing." College

Composition and Communication 35.3 (1984): 276-84.
Steinberg, Erwin R. "Protocols, Retrospective Reports, and
the Stream of Consciousness." College English 48
(1986): 697-712.

Underwood, Virginia Allen. A Survey of Some Rhetorical

Heuristics and Their Implications for the Teaching of
Composition. The Florida State University, 1980 Ph.D.

Witte, Stephen P. "Revising, Composition Theory, and Research

Design." The Acquisition of Written Language: Response
and Revision. Ed. s.w. Freedman. Norwood, NJ: Ablex,
1985. 250-84.

Young, Richard E., and Alton L. Becker. "Toward a Modern Theory

of Rhetoric: A Tagmemic Contribution." In Contemporary
Rhetoric: A Conceptual Background with Readings. Ed. w.
Ross Winterwod. New York: Harcourt, 1970: 123-143.

Young, Richard E., Alton L. Becker, and Kenneth L. Pike.

Rhetoric: Discovery and Change. New York: Harcourt Brace
Jovanovich, Inc., 1970. Chapter 5.
 106

Works on Pedagogy for Writing to Solve Problems:

Larson, Richard L. "Problem-solving, Composing, and Liberal

Education." College English 33.6 (1972): 628-635.

MacDonald, Susan Peck. "Problem Definition in Academic

Writing." College English 49.3 (1987): 315-331.
Nugent, Harold E., and Susan Monroe. Critical Thinking and
Problem Solving: The Evolution of a Course. Bedford, NH:
Annual Meeting of the New England Association of Teachers
of English, October 8-10, 1982.

Nugent, Susan Monroe. Teaching Critical Thinking: Protocols

in the Classroom. Detroit, MI: Conference on College
Composition and Communication, March 17-19, 1983.

Rosen, Joan. "A Model for Teaching is a Model for Writing."

Teaching English in the Two-Year College (1985):
274-279.

Rosen, Joan G. "Problem-Solving and Reflective Thinking:

John Dewey, Linda Flower, Richard Young." Journal
of Teaching Writing 6.1 (1987): 69-78.

Teich, Nathaniel. "Rhetoric and Problem-Solving Strategies

in Advanced Composition: A Pluralistic Approach."
New York, NY: Conference on College Composition and
Communication, March 29, 1984.