00001.tif

PDF

Download

Open document

Transcript (if available)

Content A PHENOMENOLOGICAL CRITIQUE
OF PROTOCOL ANALYSIS
IN COMPOSING PROCESS RESEARCH
by
John R. Edlund
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(English)
August 1991
Copyright 1991 John R. Edlund
UMI Number: DP23156
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author.
Dissertation Publishing
UMI DP23156
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106-1346
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-4015
This dissertation, written by
Jo h n \R, Ed 1 u n d '.........................................
under the direction of k . i s Dissertation
Committee, and approved by all its members,
has been presented to and accepted by The
Graduate School, in partial fulfillm ent of re
quirements fo r the degree of
D O C TO R O F P H ILO S O P H Y
Dean of Graduate Studies
Date v -
DISSERTATION COMMITTEE
Chairperson
Ph.D.
E
’9 1
E23
CHAPTER ONE
THE IMPLICATIONS OF PHENOMENOLOGY
FOR RESEARCH INTO THE COMPOSING PROCESS . 1
Composition and Scientism .................. 4
The Mathematization of the World............ 11
The "Nuts and Bolts" of Phenomenology .... 21
Phenomenology in the Social Realm .......... 3 9
Signs and the Intersubjective Boundary .... 64
CHAPTER TWO
JANET EMIG AND PAUL RICOEUR:
TOWARD A PHENOMENOLOGICAL UNDERSTANDING
OF THE COMPOSING PROCESS........ 77
The Phenomenology of the W i l l .............. 79
The Composing Processes of Twelfth Graders:
Background of the Study ........ 89
"Reflexive" and "Extensive" Writing ........ 102
The Composing Process and the Project .... 110
Emig's Categories: Context and Stimulus . . . 116
Science and Method............................121
Prewriting, Planning, Starting, Stopping ... 129
Inscription and Reformulation .............. 138
Decision, Motive, and Value ................ 142
A Redefinition of Emig's Categories ........ 152
Habit and the Accessibility
of the Composing Process.......... 155
CHAPTER THREE
LINDA FLOWER AND JOHN R. HAYES:
DISCOVERING THE LIMITS OF THE COGNITIVE APPROACH 163
Protocols and the Cognitive Model .......... 189
Bridge Principles and Models ................ 2 06
Segmentation, Coding and Analysis .......... 214
CHAPTER FOUR
A PRACTICAL CRITIQUE OF
THE FLOWER AND HAYES MODEL...... 227
The Interface with the Empirical World .... 231
Describing Individual Differences .......... 238
The Task Environment......................241
Other Elements of the Model..............248
Implications of the Model for Teaching .... 263
Representations and the Mind..............277
CHAPTER FIVE
FROM SIGN TO SEMANTICS:
THE PROBLEMS OF REFERENCE AND MEANING . . 286
Reflective Methodologies ................... 294
From Reflection to Critical Consciousness . . 3 09
The Contextualist Perspective . . .
Ricoeur and the Problem of Reference
314
317
BIBLIOGRAPHY
323
i i i
LIST OF FIGURES
Figure 1: The Intentional Structure of a Decision . 147
Figure 2: Temporal and Non-temporal Categories . . 155
Figure 3: The Structure of the Flower and Hayes
Model.................................... 229
Figure 4: The Structure of Generating........... 2 60
CHAPTER ONE
THE IMPLICATIONS OF PHENOMENOLOGY
FOR RESEARCH INTO THE COMPOSING PROCESS
Though in the post-modern world it has become
unfashionable to attempt to ground scientific,
philosophical or literary study in any sort of
foundational principles, in practice most scholars and
researchers are uncomfortable without benchmarks of some
kind. Literary study is no exception. There is a need
for a way to determine which of competing
interpretations or readings has a greater claim to
validity than another, or at least to correlate their
arguments. Though attempts to ground literary
interpretations in psychology, history, and linguistic
principles have been made, in practice most literary
study is grounded on the text as the one invariant
feature to which an appeal can be made. Traditional
j product oriented composition theory as it has developed
! in English departments is also grounded on texts, making
it compatible with literary practices. Methods which
count T-units, or errors, or sentence varieties, or
1
types and tokens are examples of this text-based
science. When, in about 1970, compositionists began to
focus on the composing process, it was a radical
departure from previous practice, because the object of
I
study was no longer a text or a body of texts, but a {
series of psychic events and acts. However, {
compositionists were not initially aware of the radical J
nature of the change in focus. The predominant method |
l
by which the composing process has been investigated has |
i
been through think-aloud "protocols,” in which writers !
are asked to verbalize their thoughts as they compose a
text. Protocol analysis is essentially a way of 1
bringing a representation of psychic processes, a J
semiotic residue, into the objective world where it can
be studied as data through empirical methods, or
interpreted as a footnote or subtext to an actual text.
Thus the detour into the psychic realm is momentary and
incomplete.
Janet Emig, who initiated the shift in focus from
I
the compositional product to the process, and introduced I
i
the method of protocol analysis into the field of
I
»
composition, was aware from the beginning of the
limitations of the method. In her seminal study, The
Composing Processes of Twelfth Graders, she writes,
i
For the purposes of this inquiry, eight
twelfth graders attempt to compose aloud.
The assumption here is that composing !
aloud, a writer's effort to externalize !
his process of composing, somehow
reflects, if not parallels, his actual
inner process (40-41).
Emig enters the new study with appropriate caution, both
in her language and in her later interpretation of her
results, a caution that was not emulated by many who
I
followed her in using this method. She clearly labels
as an assumption the idea that there is some motivated
i
connection between the protocols and mental processes, j
and she indicates, with the word "somehow," that the j
i
nature of this connection is neither understood nor .
under investigation at this time. In effect, she
brackets the philosophical questions that lie behind her
method. As we shall see, behind the metaphor of "inner"
and "external," and the words "reflect" and "parallel,"
are complex philosophical assumptions that require
elaboration and scrutiny.
I
\
; This is not to say that The Composing Processes of !
i
Twelfth Grader's is a naive or wrong-headed study.
Protocol analysis was an exciting and thought-provoking j
i
beginning, and the field of composition has explored a
wealth of theoretical and pedagogical perspectives along
the path blazed by this work, but in research into the
I
composing process itself we have not progressed much j
I
beyond Emig's original innovation. i
With the composing process as a new object of
study, composition needed a new ground and a new method.
In this study I hope to show that phenomenological
investigation, as originally developed by Edmund Husserl
and practiced by philosophers such as Paul Ricoeur, will
illuminate the problems implicit in Emig's cautious
words and will allow an exploration of whether composing
aloud externalizes, reflects, parallels or has some
I
other relationship to actual mental processes. I will
also investigate the worthiness of cognitive science,
especially as practiced by Linda Flower and John Hayes,
as a scientific, objective alternative to the
phenomenological perspective. The focus of this study
j will be on a re-evaluation, from the point of view of
J phenomenology, of some of the most influential research
i on composition that has been done up to this time.
i
Composition and Scientism
Maxine Hairston has argued that the movement to a
process-centered theory of composition that began in the
early 1970's with the publication of Janet Emig's work
was equivalent to a Kuhnian paradigm shift, a
"replacement of one conceptual model for another"
(Hairston "Winds of Change" 76). For Louise Phelps,
however, the shift of focus from product to process was
4
a mere harbinger of a greater shift which is only now
taking place. In Composition as a Human Science. Phelps
characterizes composition as a young discipline
awakening to find itself already situated in a
postmodern world, already "embedded in a matrix of
concerns with which it has complex and unexamined
relationships . . . preoccupied with its adolescent
growth while trapped by historical accident into an
exclusive and dependent relationship with literary j
studies" (p. 4). She argues, "Composition comes to
maturity at just the moment when discourse (especially
writing) and its interpretation stand at the epicenter j
i
of a great change, a fundamental crisis in human j
I
I
consciousness some regard as one of the revolutionary j
transitions that transform history" (p. 4). This crisis I
i
is a result of the grand battle between the proponents j
! and practitioners of object oriented approaches, j
| j
| including positivism and empirical science, and those of
subject oriented approaches— hermeneutics, and
phenomenology.
A change in focus is not enough to indicate a
paradigm shift. What is required is a change in the
concept of knowledge— what it is, how to get it, and
what it means to have it. For this century and much of
the last, empirical science on the model of physics has
5
permeated and dominated the thought of our culture to
the extent that it is widely believed that unless
something can be investigated through the scientific
method, it cannot be known at all. The academic version
of this scientistic view can be characterized as a
belief that "the methodology of the positive sciences is
in principle capable of answering all meaningful
questions and that philosophy is a pre-scientific stage
in the thrust toward positive science and will wither
away in a scientific culture" (Heelan "Hermeneutics of
Experimental Science" 9). Due to the dominance of this
view, researchers and scholars in all academic
disciplines live in fear of the subjective, the
impressionistic, and the irrational. Much of what we
want to know about, however, and this is especially true
for compositionists, is difficult to investigate
indirectly through empirical methods. As Husserl says,
all the puzzling problems of the natural sciences, and
even the mathematical sciences, "lead back to the enicrma
of subjectivity and are thus inseparably bound to the
enigma of psychological subject matter and method"
(Crisis 5). Composition, and the social sciences in
general, are caught on the horns of this dilemma.
In literary studies themselves, the scientistic
dogma has produced a number of paradoxes. This is not
the place to present a history of scientism in the study
of literature, but because of the powerful if indirect |
I
influence of literary studies on composition, a few
examples will clarify the nature of the environment in
which composition developed. One of the most
influential proponents of the scientistic view in
literary studies was I.A. Richards, who attempted, in
Practical Criticism, to found literary criticism on
psychology. His experiments mimicked certain aspects of
i
the scientific method, but not others. He found that j
l
J his students had great difficulty telling good poems
from bad ones, but he decided which ones were which. He j
i
was also very unclear about basic facts concerning his
subjects. For example, he says "Men and women were
probably included in equal numbers" (p.4). Richards
exhibits a blind faith in the notion of scientific
i
i
i progress when he notes, "It should be borne in mind that |
the knowledge which the men of A.D. 3000 will possess,
if all goes well, may make all our aesthetics, all our
psychology, all our modern theory of value, look
pitiful" (Richards Principles 4). Richards believes
that psychology is an empirical science, and that the
painstaking accumulation of knowledge typical of such a
1
t science will allow the men of the future to have greater
insight into literature than men of his own time. It is
clear from his own methods, however, that he is not
, practicing the scientific method and thus is not adding
I
to that stock of knowledge.
Wellek and Warren, in Theory of Literature, define
the problem as "one of how, intellectually, to deal with
I
I art and with literary art specifically. Can it be done?
i
And how can it be done?" They note that one answer to j
this question has been to use the methods developed by |
the natural sciences, "which need only be transferred to
I the study of literature." They distinguish several j
I
J kinds of such transfers.
One is the attempt to emulate the general j
scientific ideals of objectivity,
impersonality, and certainty, an attempt
which on the whole supports the collecting
of neutral facts. Another is the effort
to imitate the methods of natural science
j through the study of causal antecedents j
and origins; in practice, this "genetic |
method" justifies the tracing of any kind !
of relationship as long as it is possible .
on chronological grounds. Applied more j
rigidly, scientific causality is used to j
explain literary phenomena by the
assignment of determining causes to ;
economic, social, and political
conditions. Again, there is the I
introduction of the quantitative methods j
appropriately used in some sciences, i.e, j
statistics, charts, and graphs. And
finally there is the attempt to use
biological concepts in the tracing of the
evolution of literature (16).
Wellek and Warren are not satisfied with the results of
these "scientific" investigations, however. Theory of
Literature is divided into separate discussions of
extrinsic and intrinsic methods of literary study, and
as New Critics, Wellek and Warren favor the latter !
I i
approach (139) . All of the scientific methods outlined j
i , I
above clearly belong to the extrinsic category. Thus |
i
they argue:
Most promoters of this scientific invasion
into literary study have either confessed
failure and ended with scepticism or have
comforted themselves with delusions
I concerning the future successes of the
I scientific method. Thus, I.A. Richards
j used to refer to the future triumphs of j
: neurology as insuring the solutions of all j
, literary problems (16).
!
This anti-scientific stance exhibits considerable I
i
insecurity, however.
Literary scholarship has its own valid
methods which are not always those of the
natural sciences but are nevertheless
intellectual methods. Only a very narrow
conception of truth can exclude the
achievements of the humanities from the
i realm of knowledge. Long before modern
scientific development, philosophy,
history, jurisprudence, theology and even
philology had worked out valid methods of
knowing. Their achievements may have
become obscured by the theoretical and
practical triumphs of the modern physical
sciences; but they are nevertheless real
and permanent and can, sometimes with some
modifications, easily be resuscitated or
renovated. It should simply be recognized
that there is this difference between the
methods and aims of the natural sciences
and the humanities (17). j
Thus, even for New Critics, science has threatened the 1
t ,
i traditional methods of the humanities so thoroughly that j
they are in need of "resuscitation" or "renovation" i
9
under its gaze. This uncertainty lurks in every dark
corner of Theory of Literature.
Northrop Frye, in Anatomy of Criticism, is
j attracted, as was I.A. Richards, to the systematic
i
j accretion of knowledge that characterizes a natural
J science. He asks, "What if criticism is a science as
well as an art?" He says that in some literary
criticism
Evidence is examined scientifically;
previous authorities are used
scientifically; fields are investigated
scientifically; texts are edited
scientifically. Prosody is scientific in
structure; so is phonetics; so is
philology. Either literary criticism is
scientific, or all these highly trained
and intelligent scholars are wasting their
time on some kind of pseudo-science like
phrenology. Yet one is forced to wonder
whether scholars realize the implications
of the fact that their work is scientific.
In the growing complication of secondary
sources one misses that sense of
consolidating progress which belongs to a
science (Anatomy 7-8).
By "scientific," however, Frye means "systematic." He
wants to distinguish "systematic criticism" from what is
merely part of "the history of taste." Frye also
considers history, as a discipline, to be scientific in
the same sense in its handling of evidence. But history
is full of contingencies and accidents that do not lend
themselves to expression in theorems, and is not
scientific in the way that physics is.
10
Frye so wants literary criticism to be a science
that he is willing to make uninvestigated assumptions to
make it possible.
It is clear that criticism cannot be a
systematic study unless there is a quality
in literature which enables it to be so.
We have to adopt the hypothesis, then,
that just as there is an order of nature
behind the natural sciences, so literature
is not a piled aggregate of "works," but
an order of words (Frye Anatomy 17,
I emphasis mine).
Frye notes that there are those for whom the notion of
science "conveys emotional overtones of unimaginative j
!
barbarism," (p. 8) and indeed, literary studies have j
i
always ultimately backed off from any reduction of j
literature to scientific principles. Those literary j
scholars who had reservations about diminishing the
i
i
aesthetic and imaginative dimensions of literature
through scientific methods, however, were not likely to
I
feel the same way about composition, and in fact,
i |
compositionists have felt compelled to be more
i
"scientific" than literary critics, promoting a
particularly insecure attachment to scientism. Thus, 1
i
there is a history to the scientistic tendencies of Emig j
i
and other composition researchers.
The Mathematization of the World
Husserl's main interest in his last work, The
11
Crisis of European Sciences, was to articulate how the
hegemony of empirical science came to be and to argue
for phenomenology as a corrective. This work is
important not only as a critique of natural science but
also as a way into Husserl’s philosophy.
For Husserl, European culture began to take this
path with the work of Galileo, who began the
mathematization or geometrization of the world that led
to modern natural science. Aron Gurwitsch, in
Phenomenology and the Theory of Science, outlines
Husserl's argument in the following way: Geometry,
originally an art of measurement, developed toward the
concept of ideal limits, the absolutely smooth plane,
the perfect circle. Such ideal geometrical figures can
be visualized, but cannot be perceived, because they do
j not exist in reality, though they are based on real
i
j world experience. After these concepts have been
developed, things encountered in the world can be seen
as approximations of these ideal shapes. Geometry
became increasingly formalized through progressive
algebrization and logicization, a tradition was
established, the accomplishments of one generation added
to those of the previous, and the foundations of
geometry and mathematics in the life-world were
forgotten (41-43).
1 2
Gurwitsch points out that Husserl's scholarly
career began in mathematics, and that pointing out
problems in the interpretation of science is quite
j different from being hostile to science. Husserl was
| not a historian, however, and undoubtedly attributes
j
ideas to Galileo that actually belong to his
predecessors or his successors. For Husserl, Galileo is
more of a symbol for an approach than a historical
figure. This, however, does not invalidate his critique
(Gurwitsch 38-39). Husserl characterizes Galileo's
geometrical reasoning in this way:
Wherever such a methodology is developed,
there we have also overcome the relativity
of subjective interpretations which is,
after all, essential to the empirically
intuited world. For in this manner we
attain an identical, nonrelative truth of
which everyone who can understand and use
this method can convince himself (Crisis
29) .
Husserl goes on to point out, however, that, "all this
pure mathematics has to do with bodies and the bodily
world only through an abstraction, i.e., it has to do
only with abstract shapes within space-time, and these,
furthermore, as purely 'ideal' limit-shapes" (Crisis
29) .
Husserl says of this geometrical view of the world:
In place of real praxis— that of action or
that of considering empirical
possibilities having to do with actual and
really [i.e., physically] possible
13
empirical bodies— we now have an ideal
praxis of "pure thinking" which remains
exclusively within the realm of pure
limit-shapes. Through a method of
idealization and construction which ;
historically has long since been worked j
out and can be practiced intersubjectively j
in a community, these limit-shapes have
become acquired tools that can be used ,
habitually and can always be applied to
something new. (Crisis 26) . j
The point is that even geometry, which is the basis for j
natural science, has become such a cultural artifact j
! i
! that we have lost sight of the idealization and the j
! process through which it occurred. For Husserl, natural j
science can only be meaningful if the scientist has the
ability to inquire back into the original meaning of all
his meaning-structures and methods, into the historical
meaning of their primal establishment, and their
i !
inherited meanings. Husserl argues that responsible ■
philosophers and scientists must remain !
"presuppositionless" and that all things taken for (
granted are prejudices. However, as one might imagine, j
the scientific establishment of Husserl's time was not
receptive to his critique. Husserl writes with some
bitterness, "Every attempt to lead the scientist to such
reflections, if it comes from a non-mathematical,
nonscientific circle of scholars, is rejected as
'metaphysical'" (Crisis 56-57). ,
j
Gurwitsch notes that a "declaration of j
14
independence" of physics from philosophy was published
*
by Leonhard Euler, a Swiss mathematician, in 1748. The
I claim to autonomy was based on the accomplishments of
i
i
physics in the preceding century and a half. Gurwitsch
calls this the beginning of the second phase of the
philosophy of science, a phase in which science is
simply fact and no longer needs justification or
validation (Gurwitsch 36-37).
| The purpose of Husserl's Galileo analysis is to j
j
j problematize this acceptance of science as ultimate j
t I
fact. Husserl is not questioning the intrinsic validity |
of science, but rather the interpretation of what
scientific knowledge means. Galilean science is founded |
on a presupposition, the conception of nature as
possessing a mathematical structure. The Galilean
analysis has created "an infinite and yet self-enclosed
world of ideal objects as a field for study" (Crisis j
26). This conception orients scientific procedures, and \
I
determines the sense of the scientific, explanation of !
the universe. Husserl argues that because this |
presupposition has not been made explicit and has always j
i I
been taken for granted, the sense of what science means j
has been obscured (Gurwitsch 38).
Thus the "nature" of the physicist, as opposed to
nature as it presents itself in perception, is a mental
15
| accomplishment, an Idea in the Kantian sense. The
j mental processes of mathematization, idealization and
j formalization from which this accomplishment is derived
| have been forgotten, or "sedimented" as Husserl is fond
J
i
of saying. Physics has lost sight of the life-world
upon which it is based (Gurwitsch 46). j
Natural science has not only forgotten its origins
in the life-world, but also obfuscates its present j
i
i
connections.
The biophysical sciences, those which at
first concentrate, in a one-sided fashion,
purely on what pertains to the physical
body, still find it necessary to begin by
grasping the concrete entities
descriptively, analyzing and classifying
them intuitively; but the physicalistic
view of nature makes it obvious that a
further-developed physics would in the end
"explain" all these concrete entities in a
physicalistically rational way (Crisis
63) .
Husserl believes that the natural sciences themselves
employ methods that are subjective and intuitive, but
then present only empirical results. Of course, biology
has always been the most "suspect" of the natural
sciences, because to many the theory of evolution seems j
i
)
to include not only the survival of the fittest, which j
one might turn into a predictive scientific principle, j
but also the survival of the fortunate or lucky, which ]
involves contingencies and accidents. ,
i
For these reasons, Husserl sees natural science as
a cloak of mathematical ideas obscuring the life-world.
He argues:
Mathematics and mathematical science, as a
garb of ideas, or the garb of symbols of
the symbolic mathematical theories,
encompasses everything which, for
scientists and the educated generally,
represents the life-world, dresses it up
as "objectively actual and true" nature.
It is through the garb of ideas that we
take for true being what is actually a
method— a method which is designed for the
purpose of progressively improving, in
infinitum, through "scientific"
predictions, those rough predictions which
are the only ones originally possible
within the sphere of what is actually
experienced and experiencable in the
life-world. It is because of the disguise
of ideas that the true meaning of the
method, the formulae, the "theories,"
remained unintelligible, and, in the naive |
formation of the method, was never !
understood (Crisis 51-52).
The idea that science is a mere method which has been
t
confused with the ultimate truth is a common theme in |
I
the Crisis. Husserl argues that what he calls the
"life-world" is the only truly real world for all of us,
i
I
including scientists, but that the life-world has been J
i
relegated by natural science to the inferior status of a J
merely subjective phenomenon. Science refuses to accept j
i
the perceptual world at face value, instead believing }
that reality contains, embodies and conceals a j
mathematical structure. Husserl's project is to restore j
the life world, and to do so "the sediment of sense I
which has accrued to it in the course of and because of !
17
the development of modern science must be removed"
(Gurwitsch 17-18).
The problem is that natural science reduces the
world to a mathematical model because there is no method
of measuring all real properties and real-causal
relations of the intuitable world. In other words,
there are qualities which can be experienced but cannot
be measured. Specific sense qualities, such as color
and texture, cannot be treated in the same way as
qualities such as mass and shape, and so are reduced to
a secondary subjective status. Science has created an
ideal mathematical universe, a model of the life-world, j
which has come to be considered as more real than the j
observable world. Husserl compares science to the
operation of a machine and notes that if by an
"understanding" of science we mean no more than the j
successful application of methodological procedures, j
there is no problem. But a philosopher cannot be |
satisfied with the actual working of the machine, it is i
I
his job to understand the principles and assumptions |
that underlie its construction (Gurwitsch 39-40). j
In this section I have suggested that although
compositionists have chosen the composing process, a
series of psychic events and acts, as an object of
study, they have, with rare exceptions, been led by the
18 i
overwhelming predominance of natural science in our
culture to investigate that process through objectivist
methods. While natural science is an appropriate method
I
| for investigating numerous phenomena in the life-world,
I
j from Husserl's point of view, it is only a method, and
I
is not appropriate for investigating the nature and
character of subjective mental experience and mental
acts.
I
It is not my intent to argue that phenomenological j
\
f
methods are a substitute or replacement for those of j
i
natural science, however. Don Ihde wonders, quite ;
rightly, if phenomenology is naive about phenomena that
have no correlate in human experience. He asks "Is it
I entirely accidental that we have today a
phenomenological psychology but no phenomenological j
neurology?" (Ihde Hermeneutic Phenomenology 17). Alfred j
Schutz says, "The results of phenomenological research |
j
cannot and must not clash with the tested results of the ]
mundane sciences, or even with the proved doctrines of
so-called philosophies of the sciences ....
Phenomenology has its field of research in its own
right, and hopes to end where the others begin" (Schutz
Problem of Social Reality 115). What is important is
I
that the method be appropriate to the object of !
investigation. Natural science has intruded into areas '
19
!
I
where it distorts, disconnects, or discards the very I
phenomena we want to study.
i
I should also note that the natural sciences in our j
own time are by no means as resistant to subjectivity as 1
*
they were in Husserl's. Einsteinian science is very j
different from the Galilean and Newtonian science that ;
1
is the object of Husserl's polemic. Don Ihde argues 1
that the Newtonian view considers space and time as
"containers" into which the configurations and motions
i of matter must fit, whereas the Einsteinian revolution ,
I •
I reverses this conception and generates space-time out of
,
the interplay of concrete complexes of matter. This
produces a correlationalist, gestaltist, and
contextualist model for physical reality (Ihde
i
Consequences 161). And quantum mechanics has led
physics, the very model of the natural sciences, into
I I
paradoxes where mere empiricism simply will not do. j
Patrick A. Heelan says:
j
Quantum mechanics arose as the outcome of !
Werner Heisenberg's reflection that, if
the elements of a classical space-time
model of physical reality cannot be I
observed or measured, then the model must 1
yield to some theory whose elements are
observable and measurable, even though
these are not objectifiable in the
classical spatio-temporal sense ....
"Natural science," wrote Heisenberg, "does
not simply describe and explain nature: it
describes nature as exposed to our method
of questioning" (Heelan "Hermeneutics of I
Experimental Science" 42).
I
20 !
However, although the discipline of physics has had its
own paradigm shift, and literary studies has experienced
the somewhat parallel phenomenon of post-modernist
thought, as we will see, composing process research has
remained for the most part firmly rooted in the
empiricist perspective.
Implicit in Husserl's critique of the natural
sciences of his day are alternatives and correctives to
the objectivist position. In the next section I will
outline Husserl's phenomenological alternative to the
kind of knowledge provided by objective science. This
will be done as background for phenomenology as an
general attitude toward research, and to open the way
for Ricoeur's phenomenological hermeneutics.
The "Nuts and Bolts" of Phenomenology
As I noted above, Husserl wanted to develop a
philosophy that was "presuppositionless," in that
nothing was taken for granted. This does not mean that
the philosopher proceeds without assumptions, but rather
that everything that is normally taken for granted by
ordinary as well as theoretical thinkers must be
admitted, examined, and accounted for (Natanson Edmund
Husserl 12). The most important tool of Husserl's
phenomenology is the epoche, the suspension or putting
21
aside of portions of knowledge or experience so that the
functioning of another part can be made clear. Emmanuel
Levinas provides a good introductory answer to the
i
question, "What is phenomenology?"
Husserl was primarily concerned with
establishing and perfecting phenomenology
as a method, that is, as an
epistemological method of describing how j
our logical concepts and categories emerge
and assume an essential meaning. What is
the relation between our logical judgments
and our perceptual experience? This was
Husserl's question and phenomenology was
his method of responding by means of
rigorous and exact descriptions of our
intentional modes of consciousness. j
Phenomenology was thus a way of suspending j
our preconceptions and prejudices in order !
to disclose how essential truth and j
meaning are generated; it was a methodical i
return to the beginnings, to the origins
of knowledge (Kearney Dialogues 51-52). !
I
What Husserl calls epoche is often called !
I phenomenological bracketing. With the brackets on, we 1
1 i
J choose not to consider our experience in terms of causal
relations of objects conceived as discreet entities
i
somehow "out there," and the question of existence or j
nonexistence is suspended. We consider objects "as '
i
meanings presenting themselves to a consciousness" ;
(Ricoeur Freedom and Nature xiii-xiv). Don Ihde puts it j
t
this way:
The term epoche in its broadest sense
means to "suspend" or "put out of play."
But what is to be suspended is to be a |
certain set of taken-for-granted-beliefs. i
It is a suspension of "presuppositions"
rather than a reduction of (primary)
experience . . . . It is a rule which
excludes, "brackets," "puts out of play,"
all factors which may not be noted as
"bodily present" or actually fulfillable
(intuitable) within ongoing experience
(Ihde Listening and Voice 28).
| Husserl himself says of phenomenological bracketing,
I
"that which is parenthesized is not erased from the
phenomenological blackboard but only parenthesized, and
thereby provided with an index" (Ideas 171).
The value of epoche even to the practitioners of
natural science is well-illustrated in a review of
Stephen Jay Gould's new book, Wonderful Life: The
Burgess Shale and the Nature of History. The book
concerns the discovery and interpretation of the Burgess
Shale, a deposit of fossil-bearing rock about 530
million years old in Yoho National Park in the Canadian
Rockies. According to Gould, all the basic body plans
of the invertebrate animals now known appeared almost
instantaneously (geologically speaking) at the beginning
of the Cambrian period, about 550 million years ago.
Charles Doolittle Walcott, who discovered the shale in
1909, saw the organisms embedded in the shale as a
collection of the conventional types, and reconstructed
the squashed remains to fit this hypothesis. In 1970,
however, the shale fossils were reexamined by a team led
by Harry Whittington at Cambridge University, who
discovered among the Burgess Shale organisms a number of
body plans that are no longer seen and are not known
I
from other deposits of fossils, including "Qpabina. with j
its five eyes and front hose, a sort of vacuum cleaner J
with paddles, Wiwaxia. made up of overlapping plates
from which issue long spines, and Anomalocaris. whose
1
circular mouth, jointed front arms, and leaflike j
i
appendages make it unlike anything else ever seen" ;
(Lewontin "Fallen Angels" 6). Gould argues that i
Walcott's misapprehension of the forms of the organisms
in the shale represents the effects of ideology on
science, while Lewontin (the author of the review) ,
argues that it is more a matter of career needs, because '
I i
| i
i Walcott was at the top of his profession and had no need j
to overturn established truth, while the Cambridge team '
stood to benefit from discovering something entirely 1
new. Whatever motivations lie behind these efforts, it
is clear that Walcott "shoehorned" his specimens into !
presupposed categories. It is this sort of mistake
Husserl has in mind when he cautions against the
I
unexamined use of sedimented knowledge and belief.
i
On a similar theme, Ihde gives a hypothetical j
i
example of two seers, a Cartesian seer and a Druidic
seer, describing a tree. The Cartesian seer describes
i
I
| the appearance of the tree in bright sunlight, all other
24
conditions being dismissed as less than ideal. The
Druidic seer describes the tree as it appears on misty
nights and in the half-light at dawn. For this seer
i
! these are the ideal or normative conditions. Ihde draws
I
I this conclusion: "each seer sees what he already
I
believes is ’out there'; his seeing confirms him in his
metaphysics. Phenomenology holds that reality belief
must be suspended in order to allow the full range of
appearances to show themselves" (Ihde Experimenta1
37-3 8). One way to accomplish this is through fantasy
variations (discussed below).
A second fundamental concept of phenomenology is
that of "intentionality." The familiar phrase is "All
consciousness is consciousness of something." Again
Emmanuel Levinas provides a good introductory
explanation:
The most fundamental contribution of
Husserl's phenomenology is its methodical
disclosure of how meaning comes to be, how
it emerges in our consciousness of the
world, or more precisely, in our becoming
conscious of our intentional rapport . . .
with the world. The phenomenological
method enables us to discover meaning
within our lived experience; it reveals
consciousness to be an intentionality
always in contact with objects outside of
itself, other than itself. Human
experience is not some self-transparent
substance or pure cogito; it is always
intending or tending towards something in
the world which preoccupies it. The
phenomenological method permits
I consciousness to understand its own
25
preoccupations, to reflect upon itself and }
thus discover all the hidden or neglected '
horizons of its intentionality (Kearney
Dialogues 50).
i
A good deal of the confusion surrounding phenomenology j
I
i i
comes from the apparent paradox created by the epoche. j
the suspension of belief in the world, and *
intentionality, which teaches that our mental acts are 1
always "about" something, always directed toward
objects. Phenomenology is not simple introspection.
Ihde describes the phenomenology of the period of
Husserl's Ideas: ;
!
Within experience overall there is that !
which is experienced, that called the I
obiect-correlate or noematic correlate, |
and, in strict correlation with the noema, j
there is the act of experience or the j
experiencing which was the
i "subject-correlate" or the noetic act. j
I Here, as a correlative rule, it is :
maintained by intentionality simply that j
for every object of experience there is an j
act of "consciousness" which apprehends
that object, and for every act there is an
"intended" correlate, although some may
not be fulfilled (empty) (Ihde Listening 1
and Voice 35). j
The noetic act is about the noema, the appearance of the '
object, the way it presents itself. The introduction of
the noema into the correlation between the object in the !
world and the intentional act that apprehends it is the
solution to the apparent paradox. The real-world object
of an intentional act need not be present, and may in
I
fact be imaginary, so its reality is not a necessity. j
A series of simple diagrams will make the concept
clearer. (I am drawing here on Don Ihde's book
Experimental Phenomenology, which offers a very clear
and persuasive introduction to this topic.) Here is the
i
intentional correlation, including the "I," the
experiencer or subject: j
(I)noesis------- >noema !
i
(experiencer)experiencing-experienced i
What is experienced, as experienced, the object of the i
intentional act, Husserl calls the noema or noematic
i
correlate. For the mode of experiencing which is j
detected through reflection, he used the term noesis or j
noetic correlate (Ihde Exoerimenta1 42-43). "Noesis" is j
|
the act itself. The "ego" or subject has been !
i
problematic in phenomenology. Ihde asks, "How is the
'I' •constituted1 as Husserlian language would have it?"
It is constituted as the terminus of the noetic act, the
bearer of the experience, and not its origin. Idhe
provides the following diagram to illustrate this:
(I) noesis-------- >noema |
(Ihde Experimental 50) j
Analysis moves from the appearance of the object of I
experience to the noetic act, the how of experience, and j
terminates in the constitution of the "I" as a
correlation of the noema.
The "I" is a late arrival in the
phenomenological analysis. In this
respect, phenomenological analysis is the
inverse of introspective analysis. The
"I" is arrived at not directly, but by way
of reflexivity. An introspective ego or
"I" claims direct, immediate and
full-blown self-awareness as an initial
and certain given (Ihde Experimental 50).
Phenomenology is often dismissed as mere introspection,
an exercise in metaphysical solipsism. The point that
the ego is discovered by working backward through an
analysis of the object and the noetic act is thus an
important defense of the phenomenological method.
Husserl himself described this process in terms of
making the noetic act "thematic."
As subjects of acts (ego-subjects) we are
directed toward thematic objects in modes
of primary and secondary, and perhaps also
peripheral, directedness. In this
preoccupation with the objects, the acts
themselves are not thematic. But we are
capable of coming back and reflecting on
ourselves and our current activity: it now
becomes thematic and objective through a
new act, the vitally functioning one,
which itself is now unthematic (Crisis
109) .
Husserl claimed that intentionality. the
directedness of consciousness which makes an object
thematic, is an invariant feature of overall experience,
the fundamental structure which makes possible the way
j in which phenomena can and do appear. Intentionality is
i
the directional shape of experience, and functions as a
correlational rule. Husserl created a problem, however,
when he reasoned that reflecting on the noetic act,
making it thematic, required the existence of a
transcendental ego, an "I" looking down upon an "I."
j Ihde diagrams the move this way:
J Transcendental Ego (I")
\ l /
Ego(11)noesis---->noema
The problem is that in the above, the whole intentional
structure is the object for the transcendental ego. If
the intentional correlation is itself the ultimate
i
i structural feature of human experience, this move is
questionable, at least to some analysts, because it
introduces yet another level of correlation (Ihde
Experimental 45-4 6). Jacques Derrida states the problem
this way
Husserl specifies, for example, that my
transcendental ego is radically different
from my natural and human ego; and yet it
is distinguished by nothing, nothing that
can be determined in the natural sense of
distinction. The (transcendental) ego is
not an other. It is certainly not the
formal or metaphysical phantom of the
empirical ego. Indeed this leads us to
take the ego— as absolute spectator of its
own psychic self— to be but a theoretical
image and metaphor (Derrida Speech and
Phenomena 12).
This argument is worth pursuing, because post
structuralism critiques both empiricism and
phenomenology. Derrida is not playing fair here. It is
odd indeed to have Derrida talking about the "natural
29
sense" of anything, much less a "distinction." Husserl
explains the distinction very clearly, albeit in the \
I
I
terms of his own system. The realm of transcendental j
consciousness is the realm of "absolute" being, provided !
to us by the phenomenological reduction. |
It is the primal category of all being j
. . . the one in which all other regions J
of being are rooted, to which, according ;
to their essence, they are relative and on J
which they are therefore all essentially !
dependent. The theory of categories must j
start entirely from this most radical of ;
ontological distinctions— being as j
consciousness and being as something which j
becomes "manifested" in consciousness !
(Ideas 171). I
i
The transcendental ego is not in the world. It is ;
I
revealed by the reduction, m that the basic structure j
of intentionality remains; mental acts belong to an ego
even when the world is bracketed. Husserl says
The Ego living in mental processes is not
something taken for itself and which can !
be made into an Object proper of an I
investigation. Aside from its "modes of j
relation" or "modes of comportment," the
Ego is completely empty of essence-
components, has no explicable content, is
undescribable in and for itself: it is
pure Ego and nothing more (Ideas 191).
It is somewhat surprising that Derrida would reject a
distinction based on relations rather than qualities,
given his own argument that the linguistic system is
based on relations between signifiers rather than
reference to objects.
30
I 1
! Most post-Husserlian phenomenologists have rejected j
I i
the transcendental ego, leading to existentialist
philosophies, such as those of Heidegger, Merleau-Ponty,
j
and Sartre, and hermeneutic forms of phenomenology, such
i
as those practiced by Gadamer and to some extent Ricoeur
I
I
and Habermas. In the practice of a life-world j
l
phenomenology, as we must do in order to investigate the j
composing process, I think the controversy is something J
I
of a moot point, though important at a theoretical |
i
level.
The most important methodological steps toward a
phenomenological method have been discussed. Now it
remains to talk about how to "do, , phenomenology, and how
i
to apply the method to actual investigations. j
Don Ihde summarizes the method with a series of j
l
"hermeneutic" rules, the purpose of which is to enable
the observer to eliminate pre-conceived notions,
presuppositions and prejudices which arise from
sedimented knowledge and cultural biases. The first
rule is to attend to the phenomena of experience as they
appear. or, as Husserl says, return "to the things
themselves." A second, closely related rule, which
I
makes attention more rigorous, is: describe, don't ]
f
explain. To explain is to go beyond or behind the i
phenomena, to use theories, ideas, concepts, or !
31
constructions to account for what is being observed in
terms other than what appears. This is a difficult rule
to practice; for scholars the temptation to explain is
overwhelming, but must be delayed until the evidence is
in. Ihde's third hermeneutic rule is horizontalize or
i
equalize all immediate phenomena. In other words, do
not assume an initial hierarchy of "realities'* or
attempt to judge what is "real" or "more real." Epoche j
t
requires that looking precede judgment. (Ihde \
i
I
Experimental 34-36). !
Ihde's first three rules specify the field and
eliminate certain methodological choices. This is the j
first level of phenomenological investigation. But we J
could end up describing forever without establishing any
relevance. At the next stage, Husserl asks
i
1 phenomenologists to look, not just at particularities, ]
t I
1
! but for essential features (essences) of phenomena [
I
(structural features or invariants). The fourth
hermeneutic rule is: Seek out structural or invariant
features of the phenomena. One way to accomplish this
is through a practice called the variational method. j
I
which Husserl also calls eidetic analysis. In the I
example of the Cartesian and the Druidic seers related
above, each seer had chosen to privilege certain
observational conditions and certain presentations of i
32
the object. The phenomenologist seeks out all the
variations of appearances to find what it is that does
not change, looking for the structures of things that
appear in the way in which they appear. We must observe
the tree under all possible conditions, in all possible
i
! contexts, in order to discover its essence. Repeated
patterns are significant and must be actively
scrutinized and tested. Husserl preferred the technique
of "fantasy" variations, a kind of mental experiment. \
It is not necessary to proceed in this manner, however.
(Ihde Experimental 38-39). The variational method is
j
similar in practice, if not in philosophical derivation,
to the tagmemic theory of invention developed by Richard
E. Young, Alton L. Becker, and Kenneth Pike. Tagmemics,
inspired by Einsteinian physics, which I have already
I
characterized as a contextualist physics, sees an object J
in three different aspects: contrast, variation, and j
distribution; and from three different perspectives: j
i
particle, wave, and field (Winterowd Contemporary
| Rhetoric 124-25). It is another effective method for !
"loosening the givenness" of things. The question asked j
in tagmemics under variation, for example, is, "How much j
can this object change and still remain itself?" If you
know the answer to this question you have intuited an
essence.
33
Another tool familiar to rhetoricians that can be
compared to the variational method is the dramatistic
program describe by Kenneth Burke in A Grammar of
Motives. Burke's "pentad" consists of five terms: act,
agent, agency, scene, and purpose. Each term is a
starting point, or a perspective, for the analysis of
human actions and the motives that inspire them. Burke
is not interested in making rigid distinctions between
his terms, which he says "arise out of a great central
moltenness, where all is merged" (Grammar xix). His
purpose is to work his way back to that moltenness in
order to return with new distinctions, again and again.
Let one of these crusted distinctions
return to its source, and in this alchemic
I center it may be remade, again becoming
molten liquid, and may enter into new
combinations, whereat it may be again
thrown forth as a new crust, a different
distinction. So that A may become non-A.
But not merely by a leap from one state to
another. Rather, we must take A back into
the ground of its existence, the logical
substance that is its causal ancestor, and
on to a point where it is consubstantial
with non-A; then we may return, this time
emerging with non-A instead fGrammar xix).
Burke argues that such concepts as freedom and
necessity, activity and passivity, cooperation and
competition, and cause and effect are merged in the
central alembic, thus anticipating some of both Ricoeur
and Derrida.
In Experimental Phenomenology. Ihde demonstrates
34
the principle of variation through a series of figures
which are often used in psychological tests, among them
!
the Necker Cube. These drawings create an illusion of
three dimensionality through tricks of perspective, but !
I
also exhibit "multi-stability" in that it is possible to |
see each particular drawing in a number of ways. The 1
i
cube can be seen as a three-dimensional figure observed !
i
from above, but it can be flipped, mentally to be seen j
as if from below. Standard theories of perception admit j
i
only these two perspectives. Ihde demonstrates that it j
is possible to see more variations of the figures than |
i
the psychological texts normally allow, sometimes ;
through establishing a narrative or linguistic context I
i
for the new view of the figure. For example, with the
right story as context one can see the three dimensional
cube as a two dimensional drawing of a headless robot,
or an insect in a hexagonal cell (Ihde Experimental
67-108). These demonstrations are intended to be a
learn-by-doing introductory course in phenomenological
investigation. Certain variations are hard to see at
first, but once achieved are easy to return to. Ihde
explains his findings in this way:
In ordinary experience a certain
inflexibility is assumed to belong to
givenness; a hallway and a cube appear
naturally and obviously as what they are
taken to be. This is considered a
fact-stratum concerning things and j
constitutes the naive sense of givenness.
After phenomenological deconstruction,
givenness is loosened so that the
empirical order shows itself to be a
result of an unstated context of beliefs
. . . . Topographical possibilities
replace the initial fact-stratum with an
essence-stratum. All occurrences that
exemplify facts take their place as
variations upon an essential insight (Ihde
Exper imenta1 109).
Phenomenological seeing not only reveals the essential
qualities of an object, but also illuminates the nature
of the social constructs which cause us to see an object
in a particular way or as a particular thing. As Ihde
observes, once "givenness" is loosened, the unstated
context of beliefs that led to the previous naive seeing
becomes visible and statable. Though Husserl's
phenomenology can be seen as an attempt at a
foundational philosophy, especially in its early forms,
the phenomenological method is not entirely at odds with
social constructivist thought. Indeed, Husserl thought
that these constructs were so important to perception
I
j that he designed his philosophical method specifically
to push them aside, in order to see if there was
anything beyond them. Husserl says, for example, of the
(
problem of different cultures,
When we are thrown into an alien social
sphere, that of the Negroes in the Congo,
Chinese peasants, etc., we discover that
i their truths, the facts that for them are
fixed, generally verified or verifiable,
are by no means the same as ours (Crisis
139) .
Husserl goes on to say, however, that in spite of all j
these relative differences, the life—world does have a I
/
I
general structure. The nature of this structure is
i
defined by Husserl as a "habit" or "style." |
i
The things of the intuited surrounding
world (always taken as they are
intuitively there for us in everyday life
and count as actual) have, so to speak,
their "habits"— they behave similarly
under typically similar circumstances. If
we take the intuitable world as a whole,
in the flowing present in which it is
straightforwardly there for us, it has
even as a whole its "habit," i.e., that of
continuing habitually as it has up to now. 1
Thus our empirically intuited surrounding
world has an empirical over-all style
(Crisis 31).
I
Thus it is not, as Richard Rorty would have it, simply a I
question of whether we are shoved about by physical
reality, or shoved about by values (Rorty Philosophy and
the Mirror of Nature 375). We are moved by both. |
!
Gurwitsch places great importance on the ability of |
phenomenology to reveal cultural and social constructs,
!
saying "One cannot avoid the question of whether the j
main and, perhaps, only task of philosophy consists in I
understanding and accounting for the various cultural
worlds which have made their appearance in history" 1
(Gurwitsch 24-25). Husserl argues that philosophy as a
rigorous science can achieve conclusions valid for
37
everyone through transcendental or constitutive
phenomenology, which is a universal theory of
consciousness.
Ihde's experiments also demonstrate the importance
of context in phenomenological seeing. No figure can be
seen apart from its ground, and in fact, variation of
the context is part of the way to see the object as it
is. Husserl says, "For consciousness the individual
thing is not alone; the perception of a thing is
perception of it within a perceptual field" (Crisis
162). The ground or field is never bracketed, because
it is part of the nature of seeing. Rather, we must use
the variational method to see the object in all possible j
i
contexts. Even if we bracket language, social
knowledge, theoretical concepts, and reduce our
experience to pure perception, we still see the
perceptual object in a context with other objects. Ihde
calls the phenomenological attitude an "access to
context relativity" (Ihde Consequences 190). In one
sense, this is not unlike the insight of the Burkean
ratios. Burke asks, "What does the scene tell us about
the actor?" The difference is that we must imagine the
actor in all possible scenes, or at least a j
representative number of them. J
I
i
38 i
Phenomenology in the Social Realm
Husserl's investigations deal predominantly with
perceptual objects, and the basic concepts of
phenomenology are rooted in the perception of physical
things. The composing process is a much more complex
object of study. It is first, a process. Second, it is
a social activity. Third, it results in the production
of a text, a cultural object. How might
phenomenological thinking be applied to such a process?
Husserl himself says
There is a phenomenology of man, his
personality, his personal properties and
his (human) flow of consciousness;
furthermore, a phenomenology of the social
mind, of social formations, cultural
products, etc. Everything transcendant,
in as much as it becomes given in
consciousness, is an object for
phenomenological investigation not only
with respect to the consciousness of it—
e.g., the different modes of consciousness
in which it becomes given as the same— but
also ... as what is given and accepted
in the modes of givenness (Ideas 172).
Although Husserl moves in this direction himself in the
Crisis, for the most part it was left to others to
develop these life-world phenomenologies.
Don Idhe moves from a Husserlian theory of I
i
perception to a theory of text in the following manner, j
arguing that perceptual objects and texts share the
I
following characteristics: 1) All perceptual objects ]
i
are located in and relative to a perceptual field. They !
39
are contexted. 2) They are multi-stable. because they
are in a context and contexts change (though there are
invariant properties). 3) They are multidimensional.
Just as an object has a hidden side, a text has an
absent dimension (Ihde Consequences 70-71). This is a
move toward a hermeneutic phenomenology. Another
possible move is outlined by George Psathas, who argues
in "Ethnomethodology as a Phenomenological Approach in
the Social Sciences" that a modification of the natural
attitude and a particular epoche make it possible to
work as a social scientist from a phenomenological
perspective.
The ethnomethodologist suspends belief in
; society as an objective reality, except as
j it appears and is "accomplished" in and
through the ordinary everyday activities
j of members themselves. That is, he does
j not suspend belief in members' beliefs or
I in their practices as being themselves in
! the world of everyday life (Psathas
"Ethnomethodology as a Phenomenological
Approach" 78).
Psathas argues that Ethnomethodology is not
phenomenology in a new guise because ethnomethodology
thematizes members' practices and activities in the
everyday world, which phenomenology would bracket.
Ethnomethodology utilizes a special and more limited
reduction "which takes without question that there is a
world of everyday life known bv members in and through
the practices they themselves use for revealing.
40
constructing and describing it. Its topics are these
practices and only these practices." (Psathas 87). A
phenomenological investigation of the composing process
could utilize a similar reduction. Thus, while it is
true that Husserl's phenomenology brackets the cultural
world, special reductions can be designed for specific
purposes in that world without abandoning the
phenomenological method.
Psathas refers to the work of Harold Garfinkle, the
founder of this approach, and defines the reduction as
used in his studies of social groups, such as juries.
General theories, rules, or standards,
derived from the study of other settings
are not to be used to "make sense" or
"explain" the activities. However, where
members themselves use such theories,
rules or standards to make sense of their
own activities, then such practices (i.e.
members' uses) become phenomena deserving
of study in their own right (Psathas 76).
The intention of this reduction, like Husserl's epoche.
is clearly to prevent the use of presuppositions in
observing or recording the data. However, the
presuppositions of the group under study are clearly of
interest; they are part of the object of study. A
phenomenologically-based analysis of think-aloud
protocols would be required to adhere to principles such
as these.
Psathas cites Garfinkle's assertion that all
41
members accounts are "indexical" and essentially
reflexive. This means that the expressions cannot be
interpreted without reference to the speaker and his or
her situation. Social scientists' accounts are
"accounts of accounts," and are reflexive in that they
refer to systems of relevance which apply within social
science. Psathas argues, however,
Once it is recognized that such
i reflexivity is an essential property of j
accounts, direct attention can be given to !
the method by which trans-situational
knowledge and invariant features may be
discovered. The method basically consists ;
of the examination of an occasion as a
particular instance through which i
invariant (i.e., read general or j
essential) properties of members' methods
may be discovered and is a form of eidetic
analysis (Psathas 80-82).
Psathas also cites work in conversation analysis done by
Harvey Sacks and Emanuel Schegloff, two of Garfinkel's
!
j colleagues, as being carried out under a similar j
! ♦
phenomenological reduction. Protocols are no different 1
from member's accounts in this sense, and thus the
variational method should be productive in their
analysis. The orientation of most composition j
researchers toward an objective determination of what J
writers are doing, as opposed to what they think they
are doing and how they think about it, has left gaps in
our understanding.
The most important figure in the application of
42
! Husserl1s phenomenology to the social sciences is Alfred
I
Schutz. A good part of the methodology noted above has
been inspired by Schutz's work. Peter L. Berger and
Thomas Luckmann acknowledge that the fundamental insight
behind their influential book, The Social Construction
of Reality, derives from Schutz (p. 16), and in fact,
Luckmann was co-author and editor of Schutz's last work,
the two volumes of The Structures of the Life World.
Schutz got his phenomenology from Husserl himself, who
asked him at one time to be his research assistant.
Though Schutz declined for personal reasons, a
correspondence ensued which lasted until Husserl's
death, in 1938. Husserl viewed Schutz's work with warm
approval, though Schutz worked predominantly on a
phenomenology of the life-world rather than in the
i
transcendental realm. Maurice Roche writes:
From his nearest approximation to pure
phenomenology, that is from his
non-transcendental phenomenological
reduction to, and analysis of, the
non-transcendental mundane ego, Schutz
turns completely away from Husserl's path
. . . . In expanding his account of the
socially situated self, Schutz preoccupies
himself with the descriptive analysis of
the natural attitude, or Lebenswelt.
something which the pure phenomenologist
would hardly bother himself with at all.
To the latter, the natural attitude is
something to be reduced or thought-away to
facilitate the analysis of consciousness
in the transcendental mode; Schutz is
interested, on the other hand, precisely i
in the mundane and this-worldly mode of
i
43
consciousness, and thus makes the natural
attitude his thematic object (Roche 275).
Schutz, of course, was not alone in making this move.
Even Merleau-Ponty, who is closest to Husserl in his
perceptualist orientation, thought that the epoche
required the bracketing of the knowledge of natural
science, but not of the pre-scientific knowledge of the
life-world. Merleau-Ponty argued that theories,
deductions and ideal types were kinds of essences, and
that phenomenology's purpose is to investigate the
life-world and the natural attitude. "The epoche,
rather than diverting us to the transcendental realm, as
it is supposed to do, helps us to see the ordinary as
strange and in need of some explanation" (Roche 26-27).
The existential turn in phenomenology has meant that it
now deals explicitly with man and the social world.
Though Schutz appears also to take this "existentialist
turn," he remains respectful of Husserl's methods and
achievements, and proceeds in a Husserlian fashion
wherever possible.
Writing is a social action, and as such requires
the communication of thoughts and feelings, a message,
from one consciousness to another. Unlike the face-to-
face situation which it resembles, the writing process
has elements of planning and rehearsing which are
displaced from the time and place of the receiver's
44
attempt to comprehend the message. In order to do a
phenomenological investigation of this process, we need
to explore 1) the way in which a subject plans and
accomplishes an act, including the production of a text,
2) the nature and possibility of intersubjective
experience, and 3) the use and function of language and
other semiotic systems in consciousness. Alfred Schutz,
largely following Husserl's lead, takes up each of these
investigations.
In discussing the planning of an act, what Schutz
calls a "project," he distinguishes between "action,"
human conduct based on a preconceived project, and
"act," which is the accomplished action, the outcome.
All projecting consists in anticipation of
future conduct by way of fantasying, yet
it is not the ongoing process of action
but the fantasied act as having been
accomplished which is the starting point
of all projecting .... I have to place
myself in my fantasy at a future time,
when this action will already have been
accomplished. Only then may I reconstruct
in fantasy the single steps which will
have brought forth this future act (Schutz
Problem of Social Reality 20).
Paul Ricoeur also invokes this concept. I will discuss
his analysis in a later chapter.
If each of the elements of the project must be
rehearsed in fantasy, and if the deliberation is to be
rational, the actor must know the following elements (I
am paraphrasing Schutz):
a) The particular state of affairs within
which the project has to start, including
his biographical situation in the physical
l and socio-cultural environment (We might
!
! call this the "scene.").
I b) The state of affairs to be brought
about by his projected action, including
interactions with other projects and
secondary results.
c) The various means necessary for
attaining the established end, the
possibility, and compatibility of those
means (Schutz Problem of Social Reality
| 3 0—31).
So far, there is nothing in this analysis
i
j unfamiliar to any rhetorician, except for the fact that
| most of us are not used to the idea that the fantasized
]
future goal of the project is the point of origin, and
that the steps that lead to the projected goal are
reconstructed from that point. Compositionists have
found this process to be both recursive and
unpredictable in that the finished product often is a
surprise or a discovery, something much different from
what was imagined. Schutz deals with this aspect of
projecting too, but I will take up this part of the
analysis a bit later.
The temporal structure of the project is important
in determining what kind of information about the
i
process is accessible as the project unfolds. Schutz
distinguishes between "in-order-to" motives, which
involve ends and goals, and "because" motives, which are
explained in terms of the actor's background,
environment or psychic disposition. The former are ]
dominated by the future tense, while the latter are !
I
dominated by the past. Maurice Natanson explains: j
As I project my action now, I am aware of !
my in-order-to motives; indeed, it is I
precisely these motives which spur my
action. But the because motives which
could explain certain aspects of my I
projecting, their causal conditions, i
remain obscure and marginal to my
awareness. These temporal differences ,
I lead to a larger differentiation:
in-order-to motives form a subjective I
category; because motives an objective
category. The actor caught up in his
| action, understood as part of the on-going
process of projecting, defines and i
interprets the meaning of his action in I
terms of in-order-to motives (Schutz j
Problem of Social Reality XXXIX). j
i
"In-order-to" motives are accessible to phenomenological !
inspection and description, while "because" motives are j
i
causal, and subject to natural scientific inquiry. If j
this analysis is accurate, it means that information j
about goals and purposes will appear in a talk-aloud j
I
protocol, but biographical, psychological, cultural or !
environmental causes will not. I will investigate this
claim in a later chapter. Of course the "because"
motives are accessible to the subject, but only in |
reflection.
I
The actor who lives in his ongoing process !
of acting has merely the in-order-to t
motive of his ongoing action in view, that
is, the projected state of affairs to be
brought about. Only by turning back to
his accomplished act or to the past
initial phases of his still ongoing action j
or to the once established project which ,
anticipated the act modo futuri exacti can !
the actor grasp retrospectively the ;
because-motive that determined him to do j
what he did or what he projected to do. j
But then the actor is not acting anymore; j
he is an observer of himself (Schutz j
Problem of Social Reality 22). 1
i
Schutz argues that things present themselves as suitable j
j
to certain purposes, because they appear in the light of |
schemes of apperception and apprehension which belong to
what Schutz calls the "stock of knowledge at hand."
These are socially derived rules and maxims of behavior
in typical situations, recipes for handling things to
obtain certain results. This move brings scripts,
schemas, frames, and all sorts of cognitive devices into
play. Only a small part of this knowledge originates in
the experience of the individual; the bulk of it is
transmitted from parents, teachers, and all kinds of
associates. This is accumulated, probably sedimented
knowledge from many sources in the life-world, the kind
48
of knowledge that Husserl would bracket. Gurwitsch
says, "The schemes of apperception and apprehension play
a determining role in and for perception; they
contribute essentially toward making the things
encountered such as they appear in perceptual
experience" (Gurwitsch 19-20).
All projects of my forthcoming acts are
based upon my knowledge at hand at the
time of projecting. To this knowledge
belongs my experience of previously j
performed acts which are typically similar I
to the projected one. Consequently all J
projecting involves a particular j
idealization of 'I-can-do-it-again,1 i.e., |
the assumption that I may under typically
similar circumstances act in a way
typically similar to that in which I acted
before in order to bring about a typically
similar state of affairs (Schutz Problem j
of Social Reality 20-21).
|
Again, however, the temporal structure is important, for |
I
although the projecting is based on knowledge at hand, I
the stock of knowledge is modified during the j
accomplishment of the project. Because the stock of j
knowledge grows and changes constantly, the repeated j
action is more than a re-performance. The pedagogy of I
i
most compositionists reflects an assumption of a
t
learning process similar to this. !
Schutz points out that the elements of the stock of j
i
knowledge are not always consistent, rational, or
i
logical and may in fact be in conflict from a scientific j
i
or theoretical perspective.
49
Elements of knowledge are sedimented in
various situations and are related to the
mastery of heterogeneous situations.
Specific elements are not "always on
hand," like knowledge of the basic
elements of the situation and habitual
knowledge in the narrow sense. Rather,
they are simply "at hand"; they come into
use from case to case according to their
relevance for the mastery of a current
situation. In the natural attitude, there
is no motivation to keep all elements of
knowledge in fundamental agreement. Even
if elements of knowledge are
"theoretically" in contradiction with one
another, that is if they contradict one
another within a closed, formal, logically
ordered system of knowledge, they do not
need to conflict in the natural attitude
(Schutz and Luckraann Structures of the
Life World I 154).
This analysis reveals, I think, why apparently illogical
assertions are common in student texts. Students
inhabit the natural attitude. Though part of our
intention as teachers is to enable them to occasionally
operate in a reduction that Schutz calls the
"theoretical attitude," they do not stay in it for long
and may jump back and forth, connecting thereby logical
and illogical arguments. Schutz argues that serious
crises of daily life, such as war, enslavement,
unemployment, or illness, can cause a transformation of
the natural attitude, because death appears unnatural.
Under such circumstances a person approaches, and may
even place himself in, the theoretical attitude, which
brackets out everyday relevances.
i
50
But even in pre- and semi-theoretical i
transformations of the natural attitude !
(in which the accent of reality is not j
completely removed from the province of i
daily life and its taken-for-grantedness
is not unreservedly drawn into doubt), a
person begins to ask himself and the world
questions that he would never think of in
the natural attitude. He is not outside
himself; but since he is distanced, he is
in a certain sense outside this reality,
which had been so familiar to him and now
shows an alien threatening face (Schutz
and Luckmann II 129).
! The theoretical attitude is appropriate to certain of J
! . . . . !
j what Schutz calls "finite provinces of meaning," or what
compositionists tend to call "discourse communities."
Schutz develops William James' concept of sub-universes,
i
each of which, while attended to, may be considered a j
reality after its own fashion. James lists the world of J
sense or physical things, the world of science, the !
I
world of ideal relations, the world of "idols of the !
i
tribe," the various supernatural worlds of mythology and
religions, the various worlds of individual opinion, the
i
i
worlds of sheer madness and vagary. To this we could
add the world of Academia. Each of these has its own
cognitive style, and an internal consistency. We
experience a "leap" or "shock" when we shift from one to
another (Schutz Problem of Social Reality 2 30-32). A
crisis can cause this leap, or a partial leap. This j
i
# # i
analysis is helpful m illuminating exactly what we are J
asking our student to do when we assign a "project" ;
51
which involves a leap out of the natural attitude.
i
Of course, for our purposes we want to know how the
concept of the project relates specifically to writing,
to the composing process. Schutz analyzes writing as a
goal of a project, and discusses some of the steps
involved, and he provides thorough analysis of the j
I
operation of language on the semiotic level, but j
i
doesn't, to my knowledge, ever analyze the composing j
i
process on the level we are interested in. In one |
passage, Schutz discusses the project of writing a !
letter to a friend. Schutz sees this project as a :
series of component actions which are subordinate ends j
toward a higher goal.
I must write down specific marks; I cannot i
merely imagine the letter. I have the |
choice of only a few possibilities, which .
I know about through my previous
experience: pen, pencil, typewriter, each
of which has in turn a horizon of meaning ;
which has already been explicated, such as |
impersonability, carelessness, etc. These j
possibilities will for their part, each !
according to my particular interests, my \
relations to my friend and to the 1
limitation of the situation (I write more i
of less on the line and only have a pencil !
handy), force upon me decisions within a j
hierarchy of plans. If I choose the pen, I
I cannot on the other hand write the same j
letter with the pencil. If I "ask" my j
friend for certain information, then I
cannot "beg" for it, etc. If my time is
very limited, then I can only write my .
friend X, but not also Y and Z (Schutz and j
Luckmann 12 0).
i
The considerations Schutz raises here range from the :
j
52 j
j practical to the rhetorical. Some are choices which
J
j dictate other choices (if I ask, I cannot beg), in the
j
| pattern of a decision tree. Schutz argues that the
1 project is carried out as a series of choices within a
I
{ series of limitations and incompatibilities, the
| consequences of which are anticipated through
j typifications.
Within the natural attitude I do not act
only within a biographically determined
hierarchy of plans. Rather, I also see
typical consequences of my acts which are
apprehended as typical, and I insert
myself into a structure of
incompatibilities that is lived through as
being obvious. They are partially
ontological in character (I cannot write
letters with my eyes), partially
historical (it would never "have occurred
to me," in the fifteenth century, to write
other than with a pen), and partially
biographical (I have never learned to
write legibly; I have to write with a
typewriter). Thus the purely conceivable
hierarchies of plans confront specific and
partially unalterable spheres of
incompatibilities; the result is a system
of motivations for practicable goals
(Schutz and Luckmann I 20).
These "incompatibilities" have not been sufficiently
taken into account by compositionists, because they are
lived through as being obvious, but still we want to say
that it is more complicated than this analysis would
have it. What happens when the projected steps are more
difficult than they seemed when they were planned?
Schutz argues that under certain circumstances, the
53
stages of an act "'unreel' quasi-automatically and the
actor takes it for granted that they are oriented by the
t
j project." In such a case, as long as the steps remain
<
j unproblematic and meet no unexpected resistance, or are
i completely routine, the actor is not necessarily
reflexively cognizant of the individual steps.
However, if the steps themselves contain component acts
f
that become more or less problematic, then these steps |
are thematized. The action then does not "unreel" but j
t
becomes a chain in which each link must be forged to the j
next (Schutz and Luckmann II 16). The concept of
routine and non-routine action is important in an
analysis of the composing process, especially in a
pedagogical situation. Composing is non-routine at its j
best, but without routine components it is burdensome |
and perhaps impossible. Students undoubtedly want the
process to be more routine than it can or should be.
The problem is determining what aspects should be
routine and what not.
The concept of the routine is related to what
Schutz calls "typification." The stock of knowledge at
hand, the sedimentation of past experience, includes "a
network of typifications of men in general, their
typically humans motivations, patterns of action,
hierarchies of plans, schemata of expression and
interpretation, knowledge of Objective sign systems,
especially language," as well as specific detailed
knowledge of particular people and situations. The
i
j stock of knowledge is tested and modified in the course
of new experiences (Schutz and Luckmann I 66-67). The
vocabulary and syntax of everyday language itself is a
key to typification and sedimented knowledge structures.
The vernacular of everyday life is
primarily a language of named things and
events, and any name includes a
typification and generalization referring
to the relevance system prevailing in the
linguistic in-group which found the named
thing significant enough to provide a
separate term for it. The pre-scientific
vernacular can be interpreted as a
treasure house of ready made
pre-constituted types and characteristics,
all socially derived and carrying along an
open horizon of unexplored content (Schutz
Problem of Social Reality 14).
An important part of composition teaching is to get
students to see as untypical a task that they want to
see as typical, because typification derails planning.
If a label for the act exists
"objectively" in this way, if action of
this type has often been done and the
steps of action leading to the goal are
"subjectively" unproblematic, then the
project planning is, in a given situation,
simply called off. In this case the
planning is little more than the singling
out of a fully constructed theme in the
flow of experience (Schutz and Luckmann II
15) .
On the other hand the composition teacher must keep in
55_j
mind that typification is the natural mode of
consciousness in the life world. All action involves
i
I tension between the typical and the new. Education and
!
I
socialization are activities of sedimentation and
typification, which build up shared relevance systems
which become the very basis of our ability to
communicate. In these processes, the thought objects of
private experience are superseded by typifying
constructs of public objects of thought. "Typification
is indeed that form of abstraction which leads to the
more or less standardized, yet more or less vague,
conceptualization of common-sense thinking and to the
necessary ambiguity of the terms of the ordinary
vernacular" (Schutz Problem of Social Reality 323).
Although Schutz never specifically takes up the
composing process as a compositionist would, he is aware
of the backtracking and recursiveness that
compositionists have discovered by other means. Schutz
points out that while engaged in a project it may become
apparent that the planned steps are not feasible, or
that the results will not be as anticipated. When this
is the case, as it almost always is in writing,
. . . there is often an associated
reversal of meaning. Looking back at the
act, I read the result of the act into the
original act's project. I say that it was
implied or hidden in the original goal of
the act. In hindsight I then grasp the
56
event not as overlaying the original goal, ;
but rather as a discovery (Schutz and |
Luckmann I 13 0).
I
Schutz goes on to give examples of things discovered by
i
accident, like the discovery of America by Columbus, but )
i
, # t
finishes off with the example of writing a letter, ;
during which "new ideas, new nuances so to speak, come
I
to me out of my pen while I am writing down ideas which
i
have already been clearly planned. The result seems i
l
!
other than what was projected" (Schutz and Luckmann I
130). This analysis nicely accounts for the exploratory :
and creative aspects of writing, as well as planning 1
operations and goal orientations. In his analysis, the '
act of writing is both planned and discovered.
I
Difficulties in the steps of a project can be due
i
to deficiencies or contradictions in the stock of
knowledge, or in problems with the perception of the
i
object. In the latter case, the problem is, "What j
I
i
knowledge, from what system of relevance, should be '
applied?" Schutz explains the concept of thematic j
relevance in terms of perception, but the analysis could i
be legitimately applied to a text. Schutz imagines a j
man entering a room to sleep who asks himself, "Is that '
a snake or a coil of rope in the corner?" When a theme !
is constituted for consciousness, relevant elements of
i
knowledge are brought to bear. If the room is in India, 1
in a place where cobras are common, the project of
discovering what really is in the corner takes on a life j
or death importance, and the likelihood that the correct j
{
judgment is "snake" is increased. However, if the J
!
"room" is a cabin on a boat, "rope" is more likely, and
!
I
the inhabitant of the room may go to sleep without I
troubling himself too much. This, Schutz calls J
"interpretational relevance." If the knowledge is
sufficient to master the situation, interpretation is
i
routine. If not, the theme becomes a problem and the
routine flow of experience falters. When interpretation
is problematic, the observer is motivated to apply a
more or less explicit, step-by-step "judging"
explication. "In this case a 'motivated'
interpretational relevance is involved . . . the theme I
I
and the elements of knowledge are compared, until a j
founded judgment of similarity, sameness, etc. can be I
i
passed" (Schutz and Luckmann I 198).
Schutz argues that there are four main forms of
: "imposed" thematic relevance:
! The unfamiliar draws attention to itself
within the surroundings of the familiar;
one meets new themes in the "leap" from
one province of reality with finite
meaning-structure to another; changes in
the tension of consciousness within the
same province of reality can lead to
j "unmotivated" changes of theme; or,
i attention can be forced socially (Schutz
and Luckmann I 187) .
58
i The first form is similar to Burke's "perspective by
I
j incongruity." The leap from one province of reality to
I
another has been discussed above. By "tension of
consciousness" Schutz means a variation in
attentiveness, such as between full awakeness and sleep.
The pedagogical situation is a good example of the
fourth form.
In composition, however, the situation of the ,
j
project is also complicated by the fact that its i
accomplishment or success depends on the reaction of
i
another, a rhetorical or communicative effect that ;
1
cannot be determined entirely by rules of thumb, or
through the actor1s own stock of knowledge. The
application of shared "relevance systems" is essential
to the communicative process, and in overcoming the
boundary between the self and the other.
Projecting rationally such a kind of
j action involves sufficiently clear and
distinct knowledge of the situation of
departure not only as defined by me but j
also as defined by the Other. Moreover, i
there has to be sufficient likelihood that j
the Other will be tuned in upon me and '
consider my action as relevant enough to j
be motivated in the way of because by my j
in-order-to motive. If this is the case, i
then there has to be a sufficient chance
that the Other will understand me, and
this means in the case of a rational
interrelationship that he will interpret
my action rationally as being a rational
one and that he will react in a rational
way. To assume that the Other will do so
implies, however, on the one hand, that he i
!
______________________ 59__ j
will have sufficiently clear and distinct
knowledge of my project and of its place
in the hierarchy of my plans (at least as
my overt actions makes them manifest to
him) and of my system of relevances
attached thereto; and, on the other hand,
; that the structure and scope of his stock
j of knowledge at hand will be in its !
I relevant portion substantially similar to
| mine and that his and my systems of
i relevances will, if not overlap, be at
least partially congruent (Schutz Problem
of Social Reality 31).
i
j I quote at length because this seems to me to be a very j
j clear analysis of the rhetorical situation in Schutz's i
i i
phenomenological terms. Rhetoric naturally engages the 1
i
problem of the Other. In these terms one could define
teaching as an attempt to overcome the problem of the
Other by achieving a partial congruence in certain j
systems of relevance and certain areas of the stock of j
knowledge at hand. j
For Schutz, the face-to-face relationship is the j
basis upon which all other types of intersubjective |
!
relationships are founded. In the face-to-face
encounter, two humans share a sector of the life-world's
space and of world time in common in perceptual !
l
immediacy. Under these conditions the living body of ' ■
the Other is "a perceivable and explicable field of j
expression" which makes his conscious life accessible. j
In this situation it is possible for the two streams of i
consciousness to flow in true simultaneity. In Schutz's
phrase, the two "grow older together." The face-to-face j
i
encounter is the only social situation characterized by
temporal and spatial immediacy. Schutz calls this j
relationship, an essential turning-toward the Other, the !
thou-orientation (Schutz and Luckmann I . 62). 1
The thou-orientation can be either unilateral or !
i
reciprocal, but Schutz designates the reciprocal J
thou-orientation as a we-relation. The we-relation is
i
actualized, as is the thou-orientation, in various ;
stages of concrete apprehension and typification of the
Other, or in other words is a matter of degrees (Schutz
and Luckmann I 62-63). The concrete we-relation is the
only kind of encounter which allows one man to share in
the conscious life of another. When one speaks to the
other, the objective significance of the words can be
made explicit in reference to language, a "highly
anonymous system of signs." ;
I participate in the step-by-step
constitution of your speaking in the I
genuine simultaneity of the we-relation.
As a conseguence, I can (more or less
adequately or inadequately) grasp the
subjective meaning-configuration that your i
speaking and your words have for you. The i
processes of explication through which I
grasp my consociate's subjective
meaning-configuration do not, precisely
considered, belong to the we-relation. My
fellow-man's words are above all signs in
an Objective context of significance.
Further, they are also indications
("symptoms") of the subjective meaning
that all his experiences, including his
actually present speaking, have for him.
61
It is I, though, who explicate the signs
in Objective and eventually subjective
meaning-contexts. The process of
explication consequently does not belong
to the we-relation, though it presupposes
it (Schutz and Luckmann I 63).
Due to the nature of intentionality, the fact that the
mental act focusses on its object and everything else
becomes background, in the we-relation the Other is in a
i
certain sense presented as more "alive" and more ;
i
"immediate" than the ego is to itself.
Naturally I "know" myself much better than
him: my biography is recallable by me in
an infinitely more detailed fashion than
it is by someone else. But this is
knowledge about me, memory of my past, and
demands a reflective attitude. Because,
however, I unreflectively live and merge
in the actual experience, my fellow-man is
before me in his relation to me with a
j greater abundance of symptoms than I am to ;
; myself— as long as we remain just in the
1 temporal and spatial community of the
we-relation (Schutz and Luckmann I 66).
Just as the biographical information that underlies the j
"because" motives of the project are unavailable while j
i
the project is actively under way, that same information I
i
is unthematic or "backgrounded" when the intentional j
object is another person. Of course, it is possible to j
"step out" of the we-relation and engage in reflection, j
but the more this is done, "the less I live in the j
common experience and the more distant and mediate is my
consociate." The Other experienced in reflection J
becomes an object of thought, no longer experienced
without mediation. In this situation the Other can be
typified, or transformed into a typical consociate
(Schutz and Luckmann I 64). But generally, in the
we-relation experiences of consociates are not only
coordinated with one another, but are also reciprocally
determined and related. As Schutz says, "I experience
myself through my consociate, and he experiences himself
through me" (Schutz and Luckmann I 67). This
I
intentional focussing will be important in our later
analysis of talk-aloud protocols.
What Schutz calls the thev-relation. modeled on the
relationships described above but differing in spatial
and temporal immediacy, is of greater importance for the :
rhetorical situation faced by the writer. Rather than j
"consociates," Schutz calls the participants in a they- !
relation "contemporaries." "Contemporaries are not i
bodily present; therefore they are not given to me in
prepredictive experience as this particular unique
person" (Schutz and Luckmann I 73).
The "pure" thou-orientation arises from
the immediate attention directed to the j
mere existence of the fellow-man, and that j
the grasping of the being-thus-and-so of a I
fellow-man is founded on this orientation. j
This does not prove true for the
they-orientation. Basically, the latter j
consists in the fact that one imagines ;
certain typical properties .... The
typical property is anonymous with i
reference to each individual person
(Schutz and Luckmann I 80). I
*
I
__________________________63 |
The typification of the other is important in defining a
writer's self and audience. Schutz points that even in
the face-to-face thou orientation we can choose to
typify the other, giving him a partial self as the
performer of certain roles and functions, but in doing
so we typify ourselves as well. Only certain layers of
personality become involved in such a relationship. "In
I
| typifying the Other's behavior I am typifying my own,
j which is interrelated with his, transforming myself into
I
; a passenger, consumer, taxpayer, reader, bystander,
etc." (Schutz Problem of Social Reality 19). The role
of writer and audience is another such typification, but
a more complex one, because while most social roles can
be acquired naturally in day to day interaction, the
writer/reader typifications exist at a higher level of
social abstraction.
Signs and the Intersubjective Boundary
An important phenomenological question is whether
indications, marks, signs, and symbols, allow us to
cross the boundaries between ourselves and the Other, or
merely overcome them by sending back, as Schutz says,
"news from the beyond." His answer is
First, indications, marks, signs, and
symbols all convey news from beyond the
boundaries of immediate experience by
co-presenting in the experience everything
64
that is thematically, interpretively, and
motivationally relevant to the actually
present experience but in some way or
other transcends the kernel of the
experience. Second, signs, even though
essentially "newsbearers," also help us,
in reciprocal communication with other
people, to cross the boundaries to them;
symbols, although essentially embodiments
of a different reality in the everyday
world, can, in combination with certain
(namely ritualized) acts, be of assistance
in crossing the boundaries to other
realities, including the last boundary
[death] (Schutz and Luckmann II 131).
So, do indications and signs allow us to cross the
intersubjective boundary? There is a bit of hedging
here. "Help” and "assist" are very weak claims in
philosophical discourse. The problem has to do with
Husserl’s theory of "appresentation" and "apperception."
According to Husserl, the other is given as both a
physical body and a subject with psychological life.
The body, like all other material objects, is given to
my perception. His psychological life, however, "is not
given to me in originary presence but only in
copresence; it is not presented, but appresented.” This
is to say that my knowledge of the psychological life of
the other is mediated by his physical body as a field of
expression, and that I know of it only indirectly.
Facial expressions, gestures, bodily movements and
postures, all are "indications" of the psychological
I
I life of the other. For Husserl, and for Schutz
65
following Husserl, this system of appresentations, of
I well ordered indications of psychological life and
t
! experiences, is "the origin of the various forms of the
systems of signs, or expressions, and finally of
language” (Schutz Problem of Social Reality 314). |
I
Husserl's theory of signs is based on his j
j
i
understanding of perception. When we view an object in j
the world, we see it from a perspective. We see only
the side which is facing us. But this perception of the
visible part of the object "involves an apperception by
analogy of the unseen backside, an apperception which,
to be sure, is a more or less empty anticipation of what j
i
we might perceive if we turned the object around or if j
we walked around the object," based on our past J
i experiences of normal objects of this kind. Schutz goes I
I
on to point out that Husserl argues that "all j
i
significative relations are special cases of this form
of analogical apperception or appresentation," and that
I
"by appresentation, we experience intuitively something j
as indicating or depicting significantly something else"
(Schutz Problem of Social Reality 295) . 1
t
Thus for Schutz (and Husserl) all semiotic
I
J
relationships are based on apperception, a mediated
seeing, or, from the other side of the intentional act,
"appresentation." Appresentation is the phenomenon of
i |
I
i
66 :
pairing or coupling, "a form of passive synthesis which
is commonly called association." In the most primitive j
case this relation "is characterized by the fact that j
two or more data are intuitively given in the unity of !
1
consciousness, which, by this very reason, constitutes j
I two distinct phenomena as a unity, regardless of whether
i
or not they are attended to" (Schutz Problem of Social
Reality 295). It is a passive synthesis because it is
not a conscious act.
Husserl has shown that a passive synthesis
of pairing is also possible between an
actual perception and a recollection,
between a perception and a fantasm
(fictum), and thus between actual and
potential experiences, between the
apprehension of facts and possibilities.
The result of the passive synthesis of
association here involved is that the
apprehension of a present element of a
previously constituted pair "wakens" or
"calls forth" the appresented element, it j
being immaterial whether one or the other
is a perception, a recollection, a j
fantasm, or a fictum. All this happens, j
in principle, in pure passivity without j
any active interference of the mind j
. . . . These appresentational relations I
may occur on various levels: an ■
appresented object may in turn appresent
another one, there are signs of signs, and
symbols of symbols, etc. (Schutz Problem
of Social Reality 296-97).
This is much more satisfying than Saussure, because the
analysis deals not only with the structure of the sign,
but also with the way it operates in the mind.
Husserl's system calls to mind Peirce's tripartite sign,
which also allows for layers of signification and signs
that refer to other signs. However, this theory of the
apperceptive sign is also the crux of Derrida's critique
! of Husserl, one of his first "deconstructions."
I
I
j As I have noted above, one of the goals of
i
i
! Husserl's phenomenology was to achieve a
"presuppositionless" philosophy. The obvious way to
j attack such a claim is to catch Husserl making an
j unexamined assumption. This is the strategy Derrida's
| takes up. Derrida argues that Husserl assumes that
presence is the foundation and norm for language but
i
that there is no reason not to turn it the other way
j around.
j Phenomenology seems to us tormented, if
; not contested from within, by its own
1 descriptions of the movement of
temporalization and of the constitution of
intersubjectivity. At the heart of what
ties together these two decisive moments
| of description we recognize an irreducible
I nonpresence as having a constituting
I value, and with it a nonlife, a
nonpresence or nonself-belonging of the
living present, an ineradicable
nonprimordiality (Derrida Speech and
Phenomena 7).
Derrida first scrutinizes the assumptions inherent in
the terms "representation" and "appresentation," which
he says "render more palpable the resistance to the form
of presence .... What in the two cases is called a
modification of presentation (re-presentation,
I
I
I
j
| 68
ap-presentation) ... is not something that happens to
presentation but rather conditions it by bifurcating it
a priori1 * (Derrida Speech and Phenomena 7).
Derrida1s argument is derived from Husserl's
analysis of internal time consciousness. Derrida
regards Husserl’s analysis of internal time to be
admirable, and notes that Heidegger called it the first
in the history of philosophy to break with the concept
of time inherited from Aristotle's Physics. Still,
| Derrida uses this analysis to show that Husserl himself
includes the absent and the non-perceived in the
primordial now, in the form of retentions and
protentions (Derrida Speech and Phenomena 61-64).
Derrida asks,
What if we were to show, finally, that the
theme or import of "pure presentation,"
pure and primordial perception, full and
simple presence, etc., makes of
phenomenology an accomplice of classical
psychology— indeed constitutes their
common metaphysical presupposition? In
affirming that perception does not exist
or that what is called perception is not
primordial, that somehow everything
"begins" by "re-presentation" (a
proposition which can only be maintained
by the elimination of these last two
concepts: it means that there is no
"beginning" and that the "representation"
we were talking about is not the
modification of a "re-" that has befallen
a primordial presentation) and by
reintroducing the difference involved in
"signs" at the core of what is
"primordial," we do not retreat from the
level of transcendental phenomenology
69
toward either an "empiricism" or a
"Kantian" critique of the claim or having
primordial intuition; we are here
indicating the prime intention— and the
ultimate scope- of the present essay
[Footnote] (Derrida Speech and Phenomena
45-46).
Derrida claims that Husserl's theory of signs is
{ designed to keep semiosis out of primordial experience,
because signs involve a play of presence and absence,
and Husserl wants only presence to be involved in the
foundational level of experience. However, because in
Husserl's analysis of internal time consciousness the
experiencing consciousness contains retentions of the
immediate past and protentions of the immediate future,
as well as the experience of the unfolding present,
Derrida claims that primordial experience is already a
structure of absence.
Husserl makes a distinction between indicative and
expressive signs. Schutz uses the example of watching
the movement of tall grass— we can learn to distinguish
between the movement caused by wind and the movement
caused by a stalking carnivore, but both movements are
"indications" of something unseen. Similarly, facial
expressions, gestures, and bodily movements are
indications of psychological life. Is language
essentially different? Derrida argues that for Husserl
it is not.
70
Everything in my speech which is destined
to manifest an experience to another must
pass by the mediation of its physical
side; this irreducible mediation involves
every expression in an indicative
operation. The manifesting function
(kundgebende Funktion) is an indicative
function. Here we find the core of
indication: indication takes place
whenever the sense-giving act, the
animating intention, the living
spirituality of the meaning-intention, is
not fully present (Derrida Speech and
Phenomena 38}.
Husserl reduces almost all signification to indication.
Every sign that crosses the boundary of the flesh is
indicative. The only expressive sign is that which
occurs within consciousness, and that is not in
communication. Derrida tries to refute Husserl's whole
project with this contradiction.
Having its "origin" in the phenomena of
association, and always connecting
empirical existents in the world,
indicative signification in language will
cover everything that falls subject to the
"reductions": factuality, worldly
existence, essential non-necessity,
nonevidence, etc. Would we not be already
justified in saying that the whole future
problem of the reduction and all the
conceptual differences in which it is
articulated (fact/essence,
worldliness/transcendentality, and all the
oppositions systematically involved with
it) are opened up in a divergence between
two kinds of signs? And would we not be
right again in saying that this system is
set up at the same time as this
divergence, if not in it and as a result
of it? Does not the concept of
parallelism, which defines the relations
between the purely psychic— which is in
the world— and the purely
71
transcendental— which is not— and which
thus sums up the whole enigma of Husserl's
phenomenology, already present itself here !
in the form of a relation between two !
modes of signification? (Derrida Speech i
and Phenomena 30). I
------------------- / j
t
This is why Schutz hedges on the question of whether j
signs can cross the intersubjective frontier. Husserl j
says that they cannot, but for a life-world j
phenomenology, as is practiced in a social science, the
question can be put aside.
Such critiques as Derrida's are important to the
understanding of Husserl's thought. Husserl's project
is so daring and audacious that it is not difficult to
find problems. One must be careful not to make to much
out of small matters, however. For example, Frank
Lentricchia, discussing this critique, has suggested
that if we credit Derrida, there is no such thing as
phenomenology, and Husserl is not a phenomenologist.
After Derrida we cannot avoid asking
whether there is such a thing as
phenomenology: whether the attack, in the
name of intentionality, on the
dualistically isolated, self-sufficient,
and enclosed substance of Descartes's
subject is anything but the most
self-deluded of antimetaphysical moves in
the history of philosophy which succeeds
only, as Derrida shows in the case of
Husserl, in reinstituting a Cartesian
minimum under the guise of the
"phenomenological voice." If Derrida is
anywhere near the truth on Husserl, then
Poulet's explicit rejection of
phenomenology in the letter to Miller sets
him off from Heidegger, Sartre, and
72
Merleau-Ponty only in that the latter
three are less aware of their true
inclinations (Lentricchia After the New
Criticism 79).
Derrida's questions are fundamental and important,
and he is undoubtedly near a truth about Husserl, but
his critique does not invalidate Husserl's methods or
his insights, especially the fundamental insight of
intentionality. What Derrida has shown (if we credit
Derrida) is that Husserl has not succeeded in
demonstrating that a pure unmediated perception of the
object, uncontaminated by any significations or
representations of absence, is possible. But Husserl's
method has taken us very close indeed to achieving this
ideal, and in the process has shown us much about the
j nature of signification. Lentrichhia appears to think
I
j that doing philosophy is like solving an equation, and
that if you make a figuring error at the beginning of
the process the final answer must necessarily be wrong,
and therefore useless. But the practice of philosophy
is not an endeavor of right or wrong in this sense.
Most who have followed Husserl have had to deviate from
him, and most would agree that he did not succeed in
achieving all of his goals. But Derrida himself would
not be Derrida without Husserl, or without Saussure,
whom he also critiqued rather savagely in Of
Grammatoloav.
73
Don Ihde argues that Derrida' s critique of the j
!
j metaphysics of presence is founded on a metaphysics of
| f
I text which is equally subject to critique. He points i
1 . i
out that the metaphysics of language of analytic i
philosophy is to reduce expression to univocal meaning, j
I
while continental philosophies, those based on Husserl ■
I
but which have become existential or hermeneutic, |
privilege metaphor, a paradigm of a contextual, j
multi-stable, multidimensional expression. However, j
continental thought also privileges a certain kind of
text, the literary or sacred text, and then uses this !
I
text as a metaphor for reality, a metaphysical move Idhe
considers equally suspect (Ihde Consequences 74-75).
Ihde also argues that some aspects of this debate i
] emerge out of an attempt to achieve a premature closure.
One can juxtapose, for example,
Merleau-Ponty and Derrida. The one argues
for the primacy of speech incarnate in
spoken sound: the other for the primacy of
writing. But taken in the more radical
sense this turns out not to be a debate,
but a set of exercises on the
phenomenologies of the verbal embodiment
of Language varied with the inscriptional
embodiment of Language. Only when the
recalcitrant and vestigal metaphysical
interpretation tempts one to premature
foundational closure does the "debate"
emerge (Ihde Consequences 76).
One can thus argue that Derrida's metaphysics of text
and absence is a proper counterpart to Husserl's
74 !
metaphysics of perception and presence. And in fact a
dialectic of this sort is good preparation for Paul
Ricoeur's phenomenological practice.
In this chapter I have introduced the basic
concepts of phenomenology, and explored some possible j
I
applications to composition. I have brought up some
questions along the way which will be investigated in
detail in later chapters. These questions include:
1. What is the connection between talk-aloud
|
i
protocols and actual psychic events and acts?
2. What does it mean to study the composing
!
process "objectively"? j
i
3. Is the discipline of composition a social j
science? Can it be a phenomenological science? i
4. What ecoche or reduction is appropriate to the
study of the composing process? ;
5. Phenomenology makes a sharp distinction between 1
l
acting and reflecting. Are protocols reflexive
accounts? Or are they non-thematic peripheral
echoes of ongoing thought processes? Or do they
contain elements of both, as the subject alternates
between acting and reflecting?
6. Do protocols contain information about both
!
motives and causes? If they are not reflexive, j
i
causes should be excluded if Schutz's analysis is !
75
correct. i
I
7. How might the variational method be applied to ;
the investigation of the composing process? ^
8. What effects do conflicts in the "stock of
knowledge at hand" have on the composing process?
I
\
9. Is composition teaching at some level a way of :
i
socially forcing a theoretical or critical attitude
that is appropriate to a new "finite province of
meaning" or discourse community? !
i
10. How does the interplay of the typical and the
immediate affect the composing process? '
i
11. How does a text call forth schemes of ;
i
"interpretational relevance”? Does the writer j
!
imagine a typified reader and apply typified j
schemes of relevance? Is composing already a j
i
hermeneutic activity? j
12. What does it mean to investigate the composing |
process as a problem in intersubjectivity? j
!
In the next chapter, I will take up the phenomenology of j
Paul Ricoeur, and explore these questions and others in i
more detail. !
76 I
CHAPTER TWO
JANET EMIG AND PAUL RICOEUR:
TOWARD A PHENOMENOLOGICAL UNDERSTANDING
OF THE COMPOSING PROCESS
i
I
| In this chapter I will provide a relatively brief
i
' introduction to Ricoeur's thought, especially his
i
phenomenology of the will. I will contrast his
perspective to that of Schutz wherever it is appropriate
and useful to do so. I will outline Ricoeur's concepts
of project, motive, decision, and habit, and discuss the
problem of reflective consciousness as it relates to
these. And I will apply these concepts to a critique of
Janet Emig's Composing Processes of Twelfth Graders.
Ricoeur's phenomenology of the will is presented in
three books: Freedom and Nature, which is a reflexive
description and analysis of the problems of the
voluntary and the involuntary carried out under a number
of different phenomenological reductions; Fallible Man.
a sequel to the first book which discusses some of the
same problems, but with certain brackets removed; and
Symbolism of Evil, which begins Ricoeur's detour into
77
j hermeneutics. For the purposes of this chapter the
| first two books are the most important. Ricoeur's
phenomenological analyses are thorough and painstaking,
and often difficult to summarize. Brackets, once
installed, are carefully respected. When the brackets
are finally removed, the hard won insights are carefully
applied to the new perspective. In this chapter, I will
not be engaging in phenomenological description, but in
applying the insights gained by Ricoeur through his
practice of Husserl's difficult procedure. At its
heart, phenomenology is an attitude. Ricoeur says,
"phenomenology becomes astonished at what goes without
saying and submits matters to clarification that
initially appear as clear as day" (Husserl 118). My
purpose here is to problematize assumptions which have
been made in using protocols to investigate the
composing process. In doing this at times I will be
synthesizing insights gained under different reductions,
and thus to a certain extent doing violence to Ricoeur's
careful procedure. We do not live under a reduction,
however, and the reductions have already served their
j purpose.
In a sense this critique is also unfair to Janet
Emig. The Composing Processes of Twelfth Graders was,
as Earl W. Buxton says for the NOTE Committee on
78
Research, "an audacious venture into relatively
unexplored territory" (introduction Composing Processes
J
' . . .
I v). This book, however, revolutionary as it is,
i
I necessarily reflects the scientistic biases of the time.
| Emig herself has said, "A mature inquiry paradigm . . .
requires an appropriate methodology . . . Dissonance ,
between methodology and intent or methodology and
content, is often easy to discern" (Web of Meaning
166-67). I will use the phenomenological perspectives I
I
of both Schutz and Ricoeur to show that the methodology j
and conceptual framework of Emig's first book do not
represent a mature inquiry paradigm, a statement with
which I am sure Emig would agree. Researchers following |
! Emig, however, have perpetuated and even exacerbated the
i
I problems inherent in this early study, and that is why a
critique of this book is important to the field of j
<
composition at this time.
The Phenomenology of the Will
Ricoeur's phenomenology is not a life-world
i
phenomenology like that of Alfred Schutz, nor is it j
identical to Husserl's, but it partakes of both :
orientations in a careful, rigorous way. Ricoeur has 1
said, "My departure from Husserlian phenomenology was j
largely due to my disagreement with its theory of a 1
t
I
79 I
r 1 —
| controlling transcendental cogito." Ricoeur argues that
i
the "subject is not a centralizing master but rather a
disciple or auditor of a language larger that itself"
(Kearney Dialogues 27) .
Ricoeur believes that the desire to posit a
universal constituting ego is part of the nature of the
I
reflecting consciousness itself. The problem of i
i
reflection as an essentially different mode of
consciousness from the intentionality of perception or
action reappears over and over again in Ricoeur's !
!
1
analysis. His phenomenological investigation of the J
I
I
will in dialogue with the problem of the involuntary, j
the unavoidable limitations which derive from the j
character of the individual, the nature of the body, and
from circumstances in the life-world, clarifies the
nature of reflecting consciousness. These obstacles to
action he calls "the necessity."
It is precisely in the face of the
necessity both outside and inside myself
that my consciousness tends to recoil upon
itself, to make a circle with itself, in I
order to expel outside, into an empirical I
subject which it is supposed to '
constitute, these limitations of I
character, unconsciousness, and <
life-situation. By this act of expulsion, t
reflection tends to posit itself as a
universal constituting ego, which is I
supposed to transcend the limitations of ’
the empirical subject (Ricoeur Philosophy j
of Paul Ricoeur 9). !
80 i
: Ricoeur speaks of an implicit refusal of the human
i
j condition which emerges in reflection. Reflecting
consciousness has a threefold wish: "the wish to be
total, that is, without the finite perspective
associated with a particular character; the wish to be
transparent in the perfect correspondence of
i
self-consciousness with intentional consciousness; the j
wish to be self-sufficient, without the necessity for
being dependent on the nutritive and healing wisdom of
the body, which always precedes the will." For Ricoeur, |
"self-sufficiency is the supreme desire of reflecting
i
consciousness" (Ricoeur Philosophy of Paul Ricoeur 9) . ;
I
Thus the prejudices of intellectualism, and the
Cartesian split itself, arrive as a result of the nature
of reflective consciousness. Any investigative method
i
which involves reflection or introspection, such as
Emig's analysis of interviews and protocols, must not I
only allow for this distortion of the human condition, '
but also recognize the movement from an acting
intentional consciousness to a reflecting one.
I
As we have seen, Alfred Schutz and those who
followed him designed a life-world phenomenology that j
practices Husserl's method more by analogy than through !
any legitimate extension of transcendental reflection. '
J
Ricoeur remains truer to Husserl's practice, at least in j
this early work. He says that phenomenology, unlike
objectivist approaches, "wagers that the lived can be
understood and said," and that this search for the
meaning of the lived is what justifies the title of
phenomenology, the logos of phainomena. Husserlian
phenomenology presumes that the correlated structure of
| intentionality reigns in every experience. Ricoeur
| argues therefore that the principle of intentionality
i
j can legitimately be extended to the realm of action•
i
To tell the truth, Husserl has shown this
only in the order of perception; but, if
it is true that intentionality
characterizes the whole psychological
order in contrast to the physical order, a
phenomenology of volition and action
should likewise take as a guide the
noetico-noematic analysis (Ricoeur
Philosophy of Paul Ricoeur 66-67).
And in fact Husserl's metaphor of intention for the
perceptual experience derives from the sphere of action.
Ricoeur is simply reapplying the concept to its original
field. Ricoeur argues that "it is volition which is
intention par excellence."
Every intention, in the strong sense of
the term, is attention and every attention
reveals an "I can" at the heart of the "I
think." Thus, far from it being the case
that the intentional analysis of volition
is simply a transposition of that of
perception . . . it can be said that the
analysis of volition places us at the very
heart of the intentional function of
consciousness (Ricoeur Philosophy of Paul
Ricoeur 67).
For Ricoeur, the bracketing of the life-world is
both the strength and weakness of the method of
transcendental reflection. Ricoeur's procedure is to
bracket the life-world in a Husserlian fashion in order
to achieve such insights as can be had in that way, and
then to gradually remove the brackets and restore the
j entirety of human experience. For Ricoeur,
j transcendental reflection is a first step, but a step of
i
j fundamental importance.
Ricoeur believes that the method of transcendental
reflection provides rigor, but lacks depth. He agrees
that something is lost with a phenomenological
reduction. However, "only a method of abstraction
enables us to understand the fundamental meanings
entailed by life, and when the brackets can be removed
the meanings won by arid description will serve to
clarify as much as possible the obscure gushing of
freedom" (Ricoeur FN 38).
On a general level, Ricoeur's practice is unique in
that it is thoroughly dialectical. Ricoeur's dialectic
employs a weighted focus, a favored method, which is
always some version of phenomenology (in Freedom and
Nature and Fallible Man it is a Husserlian intentional
description), against another method, often an
objectivist or naturalistic one, such as a behaviorist
83
psychology. Ricoeur's dialectic of opposing values
moving toward a third term is similar to Hegel's in that j
< ■
| the notion that the "truth" of one moment is revealed ,
: i
only in the moment that follows is important to Ricoeur
(Ihde Hermeneutic Phenomenology 15). This dialectical j
practice allows the discoveries of objective methods to !
be incorporated into a phenomenological view, and j
remains true to the idea that phenomenology is not the J
I
way toward knowledge, but & way. Phenomenology thus |
becomes a powerful corrective to the distortions !
inherent in objective methods. j
Ricoeur has little tolerance for reductionism. For j
I
example, he argues that empirical psychology is j
i
characterized not only by a preference for external j
(
I
knowledge, but also by a reduction of acts to facts. !
I
Introspection is degraded to a knowledge of facts by the
omission of the mental as intentional act with reference ^
to an ego (FN 10). Ricoeur says, "A radical critique of
|
psychological determinism consists entirely of the i
rediscovery of intentionality of conscious acts. |
Consciousness is not a natural phenomenon" (Ricoeur FN
69). Ricoeur, however, occasionally finds such
science useful, albeit with reservations.
Behavioristic psychology, indeed, is
condemned to work with metaphorical !
notions, such as equilibrium,
maladjustment, tension, resolution of
tension. These metaphors are useful: they
bring about well-conducted experiments,
regroup empirical results into a coherent
view; in short, they are good working
concepts. But what is the meaning of
these metaphors borrowed from another
"region" of reality, that of the physical
thing? What are they metaphors of?
(Ricoeur FM 86).
i As we shall see, Janet Emig's work in Composing
I
l Processes is infected with the terminology of
behaviorist psychology, and this will be one opening for
my critique.
Ricoeur will sometimes oppose facts gleaned from
psychological experiments to phenomenological
descriptions and show that they complement one another,
though the phenomenological method remains privileged.
He calls this use of objective science a "diagnostic."
It is not as simple, however, as practicing two methods
and combining the results. Ricoeur states emphatically,
"We have been totally opposed to the dualism of method,
the dualism which opposes a reflective consciousness,
caught up in a circular relation with itself, to an
objectified involuntary life, leveled down to the status
of a thing (Ricoeur Philosophy of Paul Ricoeur 16). He
explains:
The reconciliation of the voluntary and
the involuntary presupposes that they
l confront each other in the same universe
j of discourse. I cannot therefore treat
i the will in a style of subjectivity and
85
- — j
the involuntary in a style of empirical |
objectivity (Ricoeur Philosophy of Paul j
Ricoeur 10).
This is also to say that both subjective and objective
methods must be done well and respect one another. They
j cannot remain split.
j Thus, in one sense, the dialectic between
I subjective and objective sciences is unavoidable, and
t
, this is all the more reason to remain vigilant regarding [
i I
1 . i
| the conceptual influences of one domain upon the other. ;
: i
As Ricoeur points out: |
I
The realm of subjectivity and the realm of !
objectivity infect each other. In this
way physics becomes loaded with !
anthropomorphisms and forces of nature are ;
conceived as types of human energy. At
the same time psychology becomes loaded
with physics: corporeal force of willing
is conceived as a cause whose effect is
movement. Thus the continuous lived
relation of idea to movement and the
external objective relation of cause to
effect face into each other (Ricoeur FN
208) .
The Composing Processes exhibits this problem at a very
t
basic level, in the confusion of motive and cause. I
will analyze this confusion in detail below.
Ricoeur argues that psychology needs phenomenology,
that the scientist has no means for resolving "the
enigma which inner experience poses" on his own grounds.
There is naturally a temptation to find an
86
objective indication of the power of will,
whether a negative indication in some gap
in determination, or a positive sign in a
force which is superior to organic life
and at the same time belongs to its level.
But if we explain these attempts clearly,
they inevitably become stranded:
determinism is always right on its own
grounds, the grounds of "empirical facts."
This obstacle, which misleads the
positivist, calls for a change of
attitude, for a passage from the "natural"
viewpoint to the "phenomenological"
attitude: only here does my body acquire
meaning by its submission or resistance to
my willing (Ricoeur FN 221).
Determinism is always right on its own grounds because
it excludes or reduces all phenomena which do not fit.
We shall see that Janet Emig is caught between a desire
to discover objective causal factors at work in the
composing proces, and an unwillingness to reduce or
exclude subjective mental phenomena. Rather than
structuring a dialectic, however, in Emig's work these
opposed purposes often create confusion.
Ricoeur has engaged a number of powerful reductions
of human experience in his philosophical career:
Behaviorism, Freudianism, Marxism, structuralism among
others. He has treated each as if it is not so much
wrong as one-sided (Walter J. Lowe, introduction FM ix).
In each of these theoretical approaches, experience is
reduced, or bracketed, but the brackets are never
removed, and the narrowed field is treated as if it were
the entire truth of the matter. For example, Richard
87
Kearney asks Ricoeur, "Is narrative language primarily
! an intentionality of subjective consciousness, as
J
j phenomenology argued; or is it an objective and J
impersonal structure which pre-determines the subjective j
operations of consciousness, as structuralism
maintained?" (Kearney Dialogues 19). Ricoeur answers j
I
that it is both at once. He finds structuralism's j
I
contribution to be an exact scientific description of j
the codes and paradigms of language. However, he says
i
that creative language use is governed by objective |
I
language codes which it continually brings to their j
i
limit in order to invent something new. Thus Ricoeur j
engages in a dialectic between structural methods and J
1
phenomenology. Similarly, when Kearney asks about the i
i
relation of Ricoeur's philosophy of language and that of !
I
Analytic Philosophy, Ricoeur says that he shares with 1
i
Analytic Philosophy a concern with ordinary language in j
contradistinction to the scientific language of j
documentation and verification. He criticizes Analytic
Philosophy, however, in that it does not take into :
account the fact that language itself is a place of •
I
prejudice and bias. ;
Therefore, we need a third dimension of
language, a critical and creative
dimension, which is directed towards
neither scientific verification nor
ordinary communication but towards the i
disclosure of possible worlds. This third
88
dimension of language I call the poetic.
The adequate self-understanding of man is
dependent on this third dimension of
language as a disclosure of possibility
(Kearney Dialogues 44-45).
It is typical of Ricoeur's thought to use the dialectic
to arrive at a third term which resolves the apparent
contradictions.
Ricoeur is able to establish a dialogue between
phenomenology and each of these approaches because he
treats each as if it were a phenomenology, recognizes
the brackets and removes them at the appropriate time.
This is the primary source of the power and usefulness
of Ricoeur's thought. My ambition in this chapter is
not to deconstruct Emig's work in a negative Derridean
fashion, but to recognize and remove the appropriate
brackets, opening up the work to broader implications
for understanding and research.
The Composing Processes of Twelfth Graders:
Background of the Study
The intellectual climate in which Emig wrote this
study was partly the literature department scientism
discussed in the previous chapter, and partly the
objectivist orientation of education and psychology.
Chomsky's transformational grammar was the latest thing
in both linguistics and education, as is evident in Earl
89
; W. Buxton's preface.
J
1 An analysis of the product, even when it !
I involves using some of the
j transformational instruments designed by
j recent investigators, does not provide
answers to a question that has loomed
large in most discussions of contemporary j
education. This question is concerned
with what happens to the student's self as
a result of the educational process
(Composing Processes v).
i
Thus, Emig’s discussions of style and syntax derive from j
i
a transformational vocabulary and perspective. This |
I
view is not entirely incompatible with her project, as
| Chomsky is interested in both linguistic competence and
i
l 1
■ production, and therefore in mental phenomena, although
1 j
i l
I the method is mechanistic. Composing Processes. j
however, does not attempt to address Buxton's question
about the influence of the educational process on the
self, except in a few hypothetical speculations.
Emig's aspirations for her project are overtly
scientistic. She says that her "hope and ambition for
this study is that it may provide one rung of a ladder
up from alchemy, so the learning and teaching of
composition may someday attain the status of a science
j as well as an art" (Composing Processes 5). She !
] believes in the progress and accumulation of data I
typical of an empirical science. At one point, for I
example, Emig hypothesizes that longer pauses precede ;
more complex grammatical transformations, and that !
i
i
__________________________________________________ 9 0 |
hesitation pauses precede "a sudden increase of ;
i
<
information as represented by grammatical ;
transformations of varying complexity." She did not
confirm these hypotheses in her study but hopes that
I
future studies with "finer calibration" and "more
careful recording techniques" will discover some sort of
I
transformational hierarchy (Composing Processes 67). In (
Husserl's analysis of the development of geometry as a ;
I
science, he points out that it was advances in the art 1
of measurement that made it possible for this abstract
mathematical knowledge to be applied to the physical j
I
world. Here we see Emig, speaking as an empirical j
i :
scientist, blaming what could well be problems of
method, on problems of measurement. While this attitude
causes problems in both the methodology and analysis of
her study, Emig's faith in scientism is not so firmly
held that she is unable to overcome it when it denies
access to the composing processes of her subjects.
Emig1s interest in the composing process is evident
long before she began to work on The Composing Processes
i
I
of Twelfth Graders. In an earlier paper, Emig indicates
her astonishment at the mechanical simplicity of the (
I
composing process as it is depicted in handbooks. She j
quotes the version that appears in Warriner's Handbook. j
j and says, j
If one were to believe this inaccuracy,
the student-writer uncomplexly sits down,
contemplates briefly what is left
carefully unspecified, completely
formulates this what in his head before
j writing a word, and then— observing a
j series of discrete locksteps in the
i left-to-right progression from planning to
i writing to revising, with no
I backsliding— builds a competent theme like
I a house of dominos. He never, for
I example, writes, then plans (Web of
j Meaning 47).
!
In the Composing Processes. Emig describes a preliminary
study she did in order to test some of the
recommendations made by popular rhetorics and handbooks.
This pilot study was designed to find out if formal or
informal outlines led to better organization and grades.
After three judges evaluated 20 themes and
submitted their grades, the grades were
coded according to whether that theme was
accompanied by a formal outline, by an
informal outline, or by no outline at all.
These data, analyzed by a program of
covariant analysis, revealed no
correlation between the presence or
absence of any outline and the grade a
student receives evaluating how well
organized that theme is (Composing
Processes 27).
This study utilizes a more conventional methodology.
The later movement to the case study methodology shows
that Emig was not only dissatisfied with the traditional
advice about the composing process, but also the
conventional methods of investigation.
Janet Emig introduced not only protocol analysis
92
into composition research, but also the case study
method itself. Richard Braddock, Richard Lloyd-Jones,
and Lowell Schoer had recommended case study in Research
I in Written Composition, published in 1963, but none of
the research covered in that book had used the method
i
’ (Composing Processes 2). Emig says, with some
I
j defensiveness,
Case study has long been a mode of inquiry
j within the physical, biological and social
i sciences: Powerful modern instances can be
i cited, for example, from psychiatry and
: from education. From his earliest papers
Freud built his theory of personality upon
the foundation of case study. In
contemporary psychiatry, Bruno Bettleheim
uses a combined method of introspection !
and case study within such accounts of
normal and deviant behavior as Love Is Not
Enough. The Informed Heart, and The Empty
Fortress (Composing Processes 2) .
Along with the think-aloud protocols Emig used extensive
. . . 1
interviews, biographical, and educational information.
The problem with this sort of research, from a
scientistic point of view, is that it generates large
volumes of data that are almost impossible to quantify
or to categorize in any sort of objective or unbiased
manner. (A speaker at the National Reading Conference
once said that most ethnographic researchers would not j
I
be too unhappy if their data just burned up.)
Researchers in this mode sometimes talk about the t
i
t
"richness" of the data, which is another way of saying
93
that there is more complexity than they can explain. As
we shall see, Emig's solution to this dilemma is to
provide both an outline, a schematic representation of
her categories, and a narrative. The narrative version,
while organized according to the outline, contains
speculations, opinions, judgments, and asides which are
interpretations of the data, but which also go well j
beyond it. In addition, the system of categories
itself, which includes such things as "context" and ;
"contemplation," goes beyond objective concerns to
include both the lifeworld and the self. The conflicts
between Emig's interests, her terminology, and her j
methods are not resolved in this book. [
Though it is clear throughout the book that Emig !
aspires to the notion of composition as a science, the i
i
notion of composition as art is never far from her
thought either. Evidence of this can be seen in her
survey of the writing processes of literary artists,
gleaned from interviews such as those from the Paris
Review, and in her correspondence about the writing
process with numerous professional writers, including
novelists and poets. She notes that the writing
processes described by professional writers in the Paris
Review interview series do not reflect the advice given
by Warriner's. and she discusses Hemingway and his fear
94
that he might have to give up writing because what Emig j
calls the manualitv of the task is so crucial to him J
(Web of Meaning 50). Clearly she is searching for an j
!
understanding that transcends scientistic explanations. I
Thus Emig's concept of the writing process is
based, in part, on psychological studies of the creative
process. Emig appropriates a four stage theory of the
creative process from The Art of Thought (1926) by
Graham Wallas. The stages are: preparation, an
investigation in all directions; incubation, a period of
I
no conscious thinking about the problem; illumination,
the appearance of the "happy idea"; and verification, in
l
which the validity of the idea is tested and its J
substance perhaps modified (Composing Processes 17). I
She says:
Although, as the brief review of rhetoric
and composition texts revealed, these data I
do not provide generative |
category-systems, several of the theories
of creativity do. Examples include the i
attenuation of the poetic or creative j
precess suggested by R.N. Wilson, i
particularly the notions that the i
"selective perception of the environment" j
and the contemplation of a product and its 1
"meaning" represent components to describe I
if they apply to the composing process of 1
students. The four stage description of
the process delineated by Helmholtz,
Wallas, and Cowley will serve as the j
center of the delineation of the writing |
process in this study (Composing Processes 1
28) .
"Selective perception of the environment" is a clue to
95
the operation of intentionality, a phrase that resonates
with phenomenological possibilities. But Emig states
clearly that she is going to impose categories on the
composing process data that come from another field and
other data, a move that is suspect from a
phenomenological perspective. Emig has some awareness
of the problem, however. As we have seen, she presents
her delineation of the writing process in two forms: as
an outline and as a narrative. She says, "The use of an
outline, which is of course linear and single layered,
to describe a process, which is laminated and recursive,
may seem a paradoxical procedure; but its purpose is to
give a category system against which the eight case
studies can be examined." She sees the narrative
portion as an attempt to convey the actual density and
"blendedness" of the process (Composing Processes 33).
And she says:
Although this category system is set forth
before the analysis of the data, it was
derived from an extensive analysis of the
eight case studies. The procedure for
analyzing the data was inductive; the
presentation is deductive" (Composing
Processes 33).
This statement, however, contradicts the previous
statement that Wallas's four stage model of creativity
"will serve as the center" to a delineation of the
writing process. Emig's outline has ten categories:
96
context, stimulus, prewriting, planning, starting,
composing, reformulation, stopping, contemplation, and
"seeming teacher influence" which might also be called
"intervention." These categories will be investigated
in detail below. Emig does not describe in this book {
I
the process by which the categories of the creativity
model evolved into the categories of the writing process
model, but we may imagine that the confusion described
above represents the operation of a kind of hermeneutic
circle in which it is not clear what came first. In The
Web of Meaning. Dixie Goswami asks Emig, "You spent
almost two years revising The Composing Processes.
Why?" Emig answers,
Initially all I had was a description of a
process that was based on a few
individuals. The hard work was trying to
do a useful taxonomy. And part of the
difficulty with the writing was that a) I
had no models that I was aware of, and b)
the view of writing at that time was that
it was unanalyzable because it was a
skill; that it was not worthy of inquiry
(62) .
It is clear that Emig was looking for a model for her
t
taxonomy. She goes on to discuss her insecurity in the j
i
face of the view that writing was both unanalyzable and
not worthy of study. To proceed with her work she
needed both to find a method of analysis and to prove
that the endeavor was worthy. The trappings of science
fulfilled both needs, but not without contradictions.
97
Emig's presentation of her taxonomy both as an
outline and as a narrative demonstrates that in her view
neither mode can accurately or completely exhaust the
reality of the process. In part this is simply a
recognition of the need for interpretation; even hard
quantitative studies become qualitative when the numbers
are interpreted, when meanings are assigned to the
results. However, she also states that the outline does
not reflect the recursive, laminated, blended character
of the actual process. In other words, the outline is a
reduction of the phenomena of the composing process.
The question is, does the outline represent the
essential qualities of the phenomena? Are the
complexities presented in the narrative version, and in
the analysis of individual cases, simply contextual
variations which have led to the apprehension of these
essences? Or is it the case that the composing process
is essentially recursive and blended, and that the
outline is a distortion of these essential qualities?
Emig's evidence, and the evidence of later researchers,
argues for the latter position. This is an important
point because the categories of the outline are in one
sense presented as the final product, the final
discovery of this research.
Ricoeur's sharp distinction between understanding
98
and explanation may help us untangle the contradictions
between these two modes of exposition. For Ricoeur, an
explanation reduces a phenomenon to the causal factors
to which it can be attributed. Explanation is the goal
of science and the scientific method. To understand, on
the other hand, is nonreductive; it means to
appropriate, to grasp in its full significance (Ricoeur
FN xxxv). Thus we might speculate that the outline is
the basis for an explanation, a reduction to causal
factors such as stimuli and interventions, and that the
narrative is an attempt to understand, to grasp the full
complexity of the process. These characterizations are
not satisfactory, however, for Emig's narrative
discussion engages in more explanatory speculation, in
terms of psychological, environmental and linguistic
factors, than the outline does. Both presentations are
thus explanations in the scientific mode. Emig
conceives of her problem as one of presentation, but it
is actually more fundamental. She has discovered that
the composing process is actually a web of buzzing
complexity, comprising continuous and interactive
subprocesses, sequences and simultaneities, multiple
modes of perception, internal and external influences,
future and past orientations, and who knows what else.
Analyzing this process is like counting fish in a tank,
99
only more difficult; nothing will stop moving. It is
easier to understand this process than to explain it,
and it is far easier to simply do it. The mind can
grasp it; it is discourse that is linear and incapable.
Even to say that the process is recursive is to
think in a sequence and may be a distortion. Process
itself might even be a deceptive, sedimented concept in
this regard. How is it possible to discuss this sort of
complexity? Ricoeur's solution is the judicious use of
phenomenological bracketing.
Ricoeur's intent in Freedom and Nature is to
explore the will as a reciprocity of the voluntary and
the involuntary and to argue that it is the voluntary
which gives meaning to the involuntary. In this he is
opposed to behavioristic and atomistic psychologies
which attempt to derive the will out of the involuntary.
In order to explore this relationship, he brackets the
emotions and the passions, the concept of Evil, and the
concept of God. These brackets are not removed in this
book. This is not to say that these concepts are
unimportant; the brackets are a methodological procedure
which make the philosophical task possible. In the
first chapter, time and duration are also bracketed,
giving us an "instantaneous segment" or a synchronic
perspective. These brackets are removed at the end of
100
the chapter, and the resulting discoveries are put into
a temporal context. Ricoeur describes this moment:
The eidetic of decision held the actual
birth of choice in suspension: the triple
relation to the project ("I want this1 1 ) ,
to itself ("I make up mv mind to do
this"), and to motive ("I decide because")
includes no reference to the history from
which choice arises. Choice is the
advance, the maturing, the growth which
brings about the relation to project, to
itself, and to the motive. The fact is
that pure description could suppress
existing history only by concentrating on
an instantaneous segment and fixating an
already performed choice in a non-temporal
state. Within such a moment a simplified
consciousness is guided by one single
project, while at the same time it becomes
determined as one and invokes, in an
eternally petrified gesture, an invariable
constellation of motives (FN 135).
With time stopped, we can see certain
interrelationships. In the flux of time, we see others.
The paragraph above describes the reduction of
experience that the bracketing achieves, the limitations
and in a sense the disadvantages of the bracketed view,
*
but this is a necessary step. Ricoeur says, "Process j
can only be elucidated in light of a non-temporal
structure established by pure description."
The purpose of bracketing is to reduce the
complexity of the phenomena to the point that they can
be dealt with in discourse. This is roughly equivalent
to the practice in objective science of eliminating or
i
controlling variables in order to isolate the phenomenon
101
under investigation. Thus the most productive way to
analyze a process is to establish a non-temporal
structure through which to view the ongoing maturation
and functioning of temporal events. This is the purpose
of Emig's outline. However, as we shall see, the
outline obscures this distinction because it is
structured in terms of temporal categories.
"Reflexive" and "Extensive" writing
The categorical problems in this study are perhaps
nowhere more apparent than in Emig's preference for the
so-called "reflexive" mode of writing. Emig suggests
the terms reflexive and extensive to replace Britton's
terms poetic and communicative. and this becomes one of
the basic dichotomies of her work. She defines
reflexive writing as taking a basically contemplative
role: "What does this experience mean?" Extensive
writing takes a basically active role: "How, because of
this experience, do I interact with my environment?"
She finds that school writing is mostly extensive, while
the writing that students do for their own purposes is
likely to be reflexive (Composing Processes 37). She
says:
The crucial distinction in this study is
between what are presented here as the two
dominant modes of composing among older
secondary school students: the reflexive
102
and the extensive. The reflexive mode
defined here as the mode that focuses upon
the writer's thoughts and feelings
concerning his experiences; the chief
audience is the writer himself; the domain
explored is often the affective; the style
is tentative, personal and exploratory.
The extensive mode is defined here as the
mode that focuses upon the writer's
conveying a message or a communication to
another; the domain explored is usually
the cognitive; the style is assured,
impersonal and often reportorial
(Composing Processes 4) .
Emig insists that teachers should teach both types of
writing, though her own students resisted writing in the
reflexive mode. If the chief audience for such writing
is the writer, why should they do such writing in
school? Surely the students feel that they already know j
how to write for themselves. She is asking them to j
write for her as if they are writing for themselves— an i
I
artificial double rhetoric. Yet she makes repeated j
attempts to get them to do so.
Although directions given the subject
concerning the first piece of writing j
encouraged personal reflexive writing,
every subject produced for the first
session a piece in the extensive mode— a
choice that might be regarded as some kind
of comment upon the mode with which the
subject was most comfortable and easy
and/or the one which he had been asked
more regularly to produce in school. In
the second session, the investigator again
attempted, through the assigned stimulus,
to elicit a piece of writing in the
reflexive mode. The responses, however,
with one exception, were not engaged, nor
even very personalized (Composing !
Processes 31).
103
It is clear that Emig values "reflexive" writing more
than "extensive," though her students don't wish to
j
produce it for her, and in spite of the fact that what
she calls the "extensive" is the writing of action and
interaction. She seems to have a non-rhetorical
literary bias in spite of all the empirical psychology
in her background. Part of the problem is the way in
which this "extensive" writing has been taught. Emig
says of two of the three boys in her study,
Their responses are not only all written
within the same mode; they are all written
within the same species of that mode. The
species is that curious one indigenous to
the American secondary school
classroom— the three-, five-, or
seven-paragraph theme. These themes, no
matter their length in paragraphs, all
have three segments: an introduction of
one or two sentences; one to three
paragraphs of development with flat,
unequivocal "topic sentences" and
development by one of three
techniques— expansion, comparison and
contrast, or example; and a concluding
paragraph which is, with a possible change
of a word or phrase, identical with the
introduction. It is as if the two
boys— and Bradford with them when he
writes prose— are so thoroughly programmed
to a single species of extensive writing
that they can readily and comfortably
compose no other. The sensation is not
unlike asking two unsophisticated
computers to compose, then discuss, three
themes each (Composing Processes 81-82).
Here is the essay reduced to a formula. But is there
not a contradiction in railing against the reduction of J
writing to the mechanical and formulaic, while at the
104
same time attempting to scientifically reduce the
writing process to a schematized outline? Why is
reflexive writing so important? Does she really want to
read writing that is not meant for herself or any
audience to read?
One way of thinking about this dichotomy is as
another set of brackets. If we bracket the relation to
the Other, both as audience and as intervener, we can
perhaps isolate certain important internal components of
the writing process. An operation of this kind would
also clarify the nature of the intersubjective
relations, when those factors were unbracketed. A
discussion of writing as a problem in intersubjectivity
might then lead to the beginnings of a phenomenological
rhetoric. It is clear that Emig makes this distinction
as part of a taxonomy of writing, the first major
division in her analytical categories, and not to
isolate certain operations in the composing process.
Let us see, however, what these brackets around the
rhetorical might achieve.
In discussing Derrida's critique of Husserl I noted
that Husserl makes a distinction between expressive
signs and indicative signs. This distinction can be
made to roughly parallel Emig's reflexive and extensive
categories, if we take reflexive to mean writing that is
105
addressed to the self, or non-addressed, and extensive
to mean writing that is addressed and has a social
purpose. For Husserl, all communicative signs are
indications, because they must pass through the
mediation of the body. Derrida characterizes Husserl's
position in this way:
The relation with the other as nonpresence
is thus impure expression. To reduce
indication in language and reach pure
expression at last, the relation with the
other must perforce be suspended. It will
no longer then have to pass through the
mediation of the physical side, or any
appresentation whatever (Speech and
Phenomena 40).
The expressive sign is internal. It is not directed
toward an other. Derrida points out that Husserl admits
that expression is "originally framed" to serve the
function of communication. "And yet expression itself
is never purely expression as long as it fulfills this
original function; only when communication is suspended
can pure expression appear" fSpeech and Phenomena 38).
As with so many Derridean arguments the problem is which
one is prior, communication or expression, and in this
case, surprisingly, Derrida is taking a common-sense
position against Husserl's somewhat forced argument.
Derrida is unwilling to accept the idea that there
is a difference between immediate and mediated
experience.
106
Husserl thinks I may have a primordial
intuition, that is, an immediate
perception of what is exposed of the other
in the world: the visibility of his body,
his gestures, what may be understood of
the sounds he utters. But the subjective
side of his experience, his consciousness,
in particular the acts by which he gives
sense to his signs, are not immediately
and primordially present to me as they are
for him and mine are for me. Here there
is an irreducible and definitive limit.
The lived experience of another is made
known to me only insofar as it is
mediately indicated by signs involving a
physical side (Speech and Phenomena 38-
39) .
Most of us would agree with Husserl here. The history
of the sign, how it came to be, is another issue. The
basis for intersubjective understanding is the body. I
am conscious; I experience my body. I see the body of
the other and that it is like mine; by analogy I believe
that he is conscious, but I do not experience this
directly. I use signs which have meaning for me. I
sense the signs of the other, and by analogy with my
own, interpret their meaning. I use signs to influence,
to persuade. Thus the practice of rhetoric begins with
the self; the understanding of the other, and of the
social world, begins with an understanding of the self.
The boundary is the body.
To bracket the relation to the Other is to isolate
discourse before it has crossed the physical barrier of
the body. It is to delay the consideration of the
107
rhetorical goal of the project and to put aside the
questions of where the ideas came from and of the social
origins of the signs. It is to focus on the internal
operations of knowing and understanding, divorced from
the operations of representing ideas for some purpose,
for some audience. It is to focus on the movement of
what Ricoeur calls "the body of my thought."
In some respects motivation is a type of
action and even of voluntary motion. It
is not only my body which is the terminus
of my action, but also all my thoughts
which constitute, so to speak, the body of
my thought. This is rather difficult to
understand correctly: all that I have
learned, all that I call my experience and
that I bear with me even if I am not
actively aware of it at the moment,
briefly, all that we can call knowledge in
the widest sense of the term, has to be
moved like my body. These are the
methods, the tools, the organs of thought
which I utilize in order to form new
thoughts. When I call up my knowledge,
not only to think of the same object anew,
but also in order to form a new thought
with the help of old thoughts of which I
am not explicitly aware, I control what I
know just as I control my body . . .
Insofar as effort applies to knowing. we
can find in it the characteristics of
recalcitrance applicable to the body
(Ricoeur FN 2 04).
Thus knowing and thinking, making an effort to
understand, are types of actions and involve both effort
and the recalcitrance of the involuntary. This aspect
of the composing process does not appear much in the
protocols, because the writer is focused on the
108
rhetorical project. This activity is undoubtedly a
major component of the prewriting process, but it
happens at a level that is to a large extent below
consciousness, and thus remains obscure. This is a
realm of habitual structures, schemes of relevance and
interpretation, as Schutz speaks of.
Thus, what we might call the "reflexive" reduction
isolates an important and interesting domain of the
composing process, but it is an artificial and abstract
realm. If there is such a thing as a reflexive mode of
writing, its purpose would be the same as that of the
protocol— to externalize and preserve a process of
thought. If it is the case that such externalization
helps us in the activity of thought, then truly
reflexive writing might be a productive part of
prewriting, but as a mode of writing in itself such non-
rhetorical writing seems artificial and of little use.
In distinguishing reflexive writing, in which the
audience is the self, from extensive writing, and
privileging the former, Emig is reducing the
communicative function of writing itself. This
distinction could in fact be quite damaging, because she
is also separating the intersubjective understanding
necessary to the rhetorical orientation of extensive
writing from its root in reflection, in the self. The
109
reflexive is not a type of writing, but a part of the
process.
The Composing Process and the Project
It should be clear by this point that the composing
process is a species of project, a type of action in the
world. (From this perspective, it does not seem that
composition can define itself as a discipline which
studies one particular type of project and no other. If
composition is to separate itself from Rhetoric and
Literature, and all the other disciplines it draws on,
as some argue it should, the object of its study must be
defined in a different way.)
As we have seen, Alfred Schutz defines a project as
the anticipation of future conduct through fantasizing.
The elements of the project are rehearsed in fantasy,
with consideration of the state of affairs in which the
project begins, the state of affairs to be brought about
by the project, and the available means. From this
perspective, the composing process is clearly oriented
toward action, toward altering or creating a certain
situation or state of affairs. Emig's privileging of
reflexive writing obscures this aspect of writing, for
we can ask "What kind of action is this?"
For Ricoeur the most important trait of the project
110
is its reference to the future. The project is an
intended future, not restricted to the continuous and
irreversible order of lived time.
I leap from project to project across dead
time, return to earlier moments, sketch
the most interesting lines of future
action, compress empty spaces, posit ends
prior to the means which precede them,
insert secondary projects into primary
projects by gradual emendation or
interpolation, and so on. This is the
pattern of human purposiveness which
organizes time ahead of the present
.... A proposed order is not the
experienced order, or, so to speak, the
acted-out order (Ricoeur FN 49).
It is important to keep in mind this jumbled complexity,
and the conflict between proposed and enacted orders of
time, as we work our way through non-temporal
descriptions of the relationships of different elements
of the project. Eventually this complexity must be
restored.
Ricoeur's concept of the project is in a broad
sense similar to Schutz's, but Ricoeur's project is
dominated by the interaction of the voluntary and the
involuntary, and his descriptions are more painstakingly
precise.
1. The willed is first of all that on
which I decide, the project I form: it
contains the direction of action to be
done by me in accord with my abilities.
2. Now a project is unreal, or rather a
type of the unreal. Its inscription into
the real by action constitutes the second
structure of the will: the voluntary
111
movement. Here the great difficulty is to
identify the intentional structure of
consciousness insofar as it is an
effective action, an action carried out. .
•
3. But there remains a residuum. The will
does not resolve into an empty project and
its practical execution in action. It
consists also of acquiescence to the
necessity which it can neither propose nor
change (Ricoeur FN 7).
Ricoeur's phrase "inscription into the real" is
especially apropos to our purposes, although he is
discussing action in general, not composing
specifically. Ricoeur's point here is that it is too
simple to think only of the project and its execution.
The "acquiescence to the necessity" is the dialogue of
the voluntary and the involuntary, the interaction of
the will, the project, and the obstacles which stand in
the way. Composition has not paid much attention to the
involuntary, as such. In part, these are the
limitations which Schutz has characterized as
ontological, historical, and biographical, problems of
the nature of the body and the mind, of available
technology, and of the state of learning and the stock
of knowledge at hand.
Ricoeur calls voluntary actions controlled actions
and notes that the project may be so implicit that it is
lost in the action itself. His example is rolling a
cigarette while speaking. "To the extent to which an
112
automatic action is even minimally observed— in a sense
out of the corner of my eye— and an explicit will could
recognize it after the fact and go back over it," it is
a voluntary action. The truly involuntary is "an
explosive, impulsive action in which the subject cannot j
recognize himself and of which he says that it escaped
him" (FN 39). Thus, the self can be recognized in any
voluntary action. Ricoeur is here distinguishing the
truly involuntary from the automatic or habitual action,
I
a topic I will discuss toward the end of this chapter. j
For experienced writers many parts of the "project" are
as unconscious as lighting a cigarette, but not so for
basic writers. I will argue later that the absence of
useful automatic habits is a distinguishing feature of
weak writers.
Following Ricoeur's analysis, if we wish to study
the composing process as a project, we have three phases
to study: the act of projecting, the fulfillment of the
project through action, and the interplay of the
voluntary and the involuntary. Note that these phases
do not form a temporal sequence, and in fact Ricoeur
brackets time in the early part of this analysis. The
involuntary is of central importance to educators; part
of our interest is in empowering and enabling the will,
I
I
in removing aspects of the involuntary, and in creating j
i
I
____________ 113 |
useful habits. To enact a project, one must have the
i
capability. Ricoeur says:
When I have decided to carry out a
delicate proceeding I feel in a sense
charged with it, in the way a battery is
charged— the act is within my power, I am
capable of it. Such power, such
capability, already belongs to the order
of action: it potentially fulfills the
project on whose contemplated execution it
is intent (Ricoeur FN 39).
He continues, saying that this impression of power can
be contradicted in the course of events. For many
students, this feeling of capability is never present.
They are asked to embark on a project that they have not
themselves willed, forced to take action they are not
capable of fulfilling, or at least that they think they
are not. Here we encounter authority structures which
are outside the purview of Freedom and Nature and to a
large extent Composing Processes as well.
Emig is well aware of this sense of capability,
however. Of Lynn, the best of her small group of
subjects, she says, "Her ego strength seems great enough
for her to believe that she can complete any assigned
i
writing task, not only adequately but skillfully enough j
to please herself and her evaluator, whoever the
evaluator may be" (Composing Processes 64). Emig finds
evidence of this in the protocols:
Lynn is not often self-congratulatory; but
occasionally her ingenuity in solving an
114
immediate writing problem delights her:
"Now this gives me a tie-in to relate to
different people's reactions. Hmmm,
that's pretty clever"; and "'Alas'— oh,
this is goodI" Her most frequent comments
indicate self-acceptance and willingness
to try (Composing Processes 63).
Obviously not all students feel this way. Unfortunately
the protocols and interview materials of the other
subjects in the study are not included in the book.
Ricoeur says "A decision can be separated in time
from any corporal execution, yet it is the power or
capability for action (or movement) which makes it an
authentic decision." He goes on to say that execution
is the only way to know the strength of the power.
Thus, "A decision implies that the project of an action
is accompanied by a power or capability of movement
which could realize the project" (Ricoeur FN 40).
Ricoeur then uses this analysis to distinguish between a
decision and a wish or a command, in which the subject
does not have the power to accomplish the action. The
involuntary dimension of the writing assignment itself
infects the whole project with a kind of hopelessness
which can only be resolved through a feeling of power
and capability, and thus the possibility of personal
engagement. But here we are caught in the paradox of
teaching, another version of the hermeneutic circle.
Capability to do arises from having done.
115
Emig's Categories: Context and Stimulus
Emig's treatment of context is fairly minimal. For
her it consists mainly of socio-economic, family, and
educational background. She does discuss the
intervention of the peer group in self-sponsored
writing, and of teachers and family members in writing
which is school-sponsored. By and large context is
treated as a ground in which causal factors are located.
I
This minimalization of context could be justified [
i
as necessary in an intense focus on the composing !
process itself. It is reflected in the structure of the
categories themselves, however. Emig begins her
analysis of the writing process with a discussion of
"the nature of the stimulus that activates the process !
i
I
or keeps it going."
For students, as for any other writers,
stimuli are either self encountered or !
other initiated. Either the student
writes from stimuli with which he has
privately interacted or from stimuli
presented by others— the most common
species of the second being, of course,
the assignment given by the teacher. Both
kinds of stimuli can be nonverbal or
verbal, although it is an extremely rare I
and sophisticated teacher who can give a
nonverbal writing assignment. All areas J
of experience, or fields of discourse, can
provide the stimuli for writing (Composing
Processes 33).
There are two sets of metaphors in use here which are
borrowed from objective science. One is the
116
stimulus/response metaphor borrowed from behaviorist
psychology, and the other is the genus/species
relationship, used in logic, but here probably borrowed
from biology. This objectivist terminology confuses all
the issues. A stimulus is a mechanistic cause.
Translating this vocabulary, we are given two types of
"causes" of writing: self-initiated and other-initiated,
and two modes of being for these causes: semiotic and
experiential. This gives us a four cell matrix, but we j
are told that the other-initiated experiential cause is j
a rarity. Emig further defines the possibilities of the {
stimulus by dividing experience into four categories. !
In his essay on poetic creativity, the
psychologist R.N. Wilson divides i
experiences tapped by writers into four
categories: (1) encounters with the j
natural (non-human) environment; (2) human i
interrelations; (3) symbol systems; and ;
(4) self ("Poetic Creativity" 167). For
the analysis of student writing in this '
inquiry, "symbol systems" becomes
"encounters with induced environments or
artifacts" (Composing Processes 36).
i
These are very odd categories from a phenomenological
i
perspective. A phenomenologist would define the first j
!
category as the thing-world, the view achieved by the ;
j
bracketing of the life-world and scientific knowledge, j
but it is likely that what is meant is something like j
I
the Romantic view of nature. The second category, the I
I
I
social, must also include the semiotic, and all other j
117
constructs of the life-world. "Encounters with induced
environments or artifacts" is a curious replacement for
symbol systems, but perhaps Wilson is thinking of the
elaborate symbol systems of a poet like Yeats, and not
something like a Saussurian semiotics, and Emig doesn't
quite know how to apply this. The fourth category, the
self, is reflection.
Emig uses these categories to define "fields of
discourse."
I
I
She tells Lynn, the subject who is most thoroughly J
analyzed in the book, "Write about a person, idea, or j
I
event that especially intrigues you." Lynn comes up j
with four possible subjects. ,
Note that two of the possible four topics j
involve Lynn: her grandmother and the two
boys; the fields of discourse in these ;
topics would be the two categories of self !
and human relationships; and that the mode I
if she wrote about either might well be
the reflexive. The other two subjects of
the old ladies on the bus and Snoopy could
be handled at greater distance, with less
self-involvement. Lynn chooses Snoopy,
the cardboard dog. Why? (Composing
Processes 48).
However, the other two topics also involve "self" and
"human relationships," though it is not clear in Emig's
system exactly how Snoopy might be categorized. These
categories do not seem to provide any useful
distinctions. Notice also that Emig is still rooting
for the reflexive mode without success.
118
Even on its own terms, this is not a very
satisfying or useful matrix of categories. But from
Ricoeur's point of view, the founding of this system on
the concept of the stimulus incorporates a fundamental
and fatal confusion. It confuses the order of
consciousness with the order of objects. It confuses
motivation with causation. This happens for two
reasons. First, in her attempt to be scientific. Emig
is attracted to the objectifying language of j
I
behaviorism. Second, Emig does not understand, or at
least is not thinking about, the fact that the project
is oriented toward the future; it is the goal of the
project, the completed act, which motivates the
decision, and the actions which lead toward it. Of
course, when one looks back at a completed act motives
look like causes, it looks like a certain action was
taken because of a particular reason. This perspective,
however, is exactly what Emig is trying to escape— the
orientation toward product. And even from the
perspective of the completed act, there is still a
i
I
distinction. As Ricoeur says, I
j
It is the nature of a cause to be knowable j
and understood prior to its effects. A !
set of phenomena can be intelligible
without reference to another set of
phenomena which result from it. The cause
confers its meaning on the effect:
understanding proceeds irreversibly from
cause to effect. On the other hand, it is |
i
!
1 1 9 i
the essence of a motive not to have a
complete meaning apart from the decision
which refers to it . . . The will, in a
single movement, determines both itself
and the definitive form of its affective
as well as its rational arguments (FN 67).
Ricoeur also distinguishes a pretext, but says that we
can only understand a pretext in contrast with its
genuine covert motive.
Here the absurdity of an objective, scientific
study of the composing process is readily apparent. If
the problem of the composing process is one that will
yield to proper measurement and proper schematic
categories, then a certain stimulus will invariably
lead, under the same conditions, to the same response.
Our problem then is simply defining with sufficient
rigor the stimulus and the conditions, as Emig has set
out to do. It is quite clear, however, that this
project will never succeed. As Ricoeur says, "motive
inclines without compelling" (FN 71). A motive is not a
cause, but is in the realm of the voluntary. Thus, "the
motive is precisely that which inclines the will by
depicting for it a good (at least a good for me, here
and now); it is therefore purely reciprocal to a will
which receives it, invokes it and bases itself upon it"
(Ricoeur Philosophy of Paul Ricoeur 6). Even if we
could somehow isolate and identify the particular
120
motives or motivational structures which led to a
particular decision, we could never be sure that the
same motives under similar conditions would lead to the
same decision. At best we can identify only tendencies
and inclinations.
Science and Method
This essential contradiction between the desire to
be "scientific" and the need to investigate human mental \
I
processes has had paradoxical results in composition j
research and practice. One interpreter, Ralph F. Voss,
says that Emig's work brought "science consciousness" to
composition, but he criticizes the procedure of taping
and observing, saying that this is not empirical |
i
science. He says that the general prestige of science
in our society was not earned from such highly
inferential procedures as case-study research, but by
laboratory science. In Voss's view, the prestige of
laboratory science brought Emig1s research more acclaim
than it deserved, and inspired imitators who were even
less careful (Voss "Emig: A Reassessment" 279). By and
large he is correct in his analysis. However, the
composing process cannot be investigated in a
laboratory. If Emig had been more scientific and
objective, she would have been even more frustrated in
I
121 i
her attempt to gain access to the composing process,
because it is the nature of objectivity to study
external manifestations rather than inner subjective I
processes.
Emig does not claim that her report is "a
definitive, exhaustive, nor
psychometrically-sophisticated account." She points to
two main problems. The first is that the sample of
students, and the sample of writing they produced, is
"far too small and skewed." The sample is small, eight
students of whom only one is analyzed in detail, but for
a case study approach this is not unusual. Sampling
error is a problem when the researcher is using a j
statistical methodology, which is not the case here.
One gets the impression that Emig feels that it would
have been better if she hadn't had to resort to case
study, but had been able to develop some clever and
accurate method to quantify the composing process data.
Her use of the word "psychometric" in this context
reinforces this impression. Her second qualification is
a more important one. She says, "even the most mature
and introspective students in the sample found composing
aloud, the chief means the study employed for
externalizing behavior, an understandably difficult,
artificial, and at times distracting procedure." If
I
I
122 |
composing aloud is difficult, artificial and
distracting, we have ample reason already to doubt the
validity of any conjectures based on this kind of data.
Emig is caught in a dilemma. She can not think of a way
to investigate the composing process indirectly, through
objective, quantifiable data, nor can she discover a
satisfactory way to investigate it directly, short of
telepathy. She chooses what seems to her the best
available means. After acknowledging these
difficulties, Emig characterizes her study as, "an
experiment in capturing a process in process," and
asserts quite correctly that "the study provides certain
kinds of data— humanistic data— that other kinds of
inquiries into composition have not yet elicited"
(Composing Processes 4-5) .
As I have noted before, Emig assumes that a
protocol "somehow reflects, if not parallels," the
writer's actual inner process (Composing Processes
40-41). There are a number of important questions here
for phenomenological analysis. Does the oral
representation actually parallel psychic processes? Or
does it reflect those processes in a changed or altered
fashion? And if either of the above is true, how does
this happen? What is the process of this representation
of the process? Emig describes a protocol as "a
writer's effort to externalize his process of
composing." If "effort" is involved, toward what
project is it applied? What is the goal of this project
for the writer? Is the protocol a sort of wiretap on
consciousness, or is it rhetorically addressed to the
investigator? And of course, writing itself is an
externalization of consciousness. The most important
difference between these externalizations is not their
embodiment in oral or written language, but their
rhetorical purposes. What is the effect on psychic
processes of this interactive double externalization?
Emig does not ask these sorts of questions. She is
interested in whether or not composing aloud will reveal
recurring behaviors, if these behaviors can be
profitably categorized, and what hypotheses can be
generated. Composing aloud is thus both the most
startling innovation and the major conceptual weakness
of this study.
From a phenomenological point of view the most
serious question about the nature of protocol data
concerns the relationship of action and reflection.
Because of the structure of intentionality, we cannot
act and reflect simultaneously. As Schutz has pointed
out, because of intentionality, in the midst of a
project our in-order-to motives, our goals, are present
124
to us, but causal factors, which are part of history and
biography and in the past, are not. Thus, if a protocol
actually parallels psychic processes, what we would
expect to find is a strong focus on the ongoing task
with minimal episodes of reflection. The protocol
discourse should be non-thematic and peripheral to the
writing task, in a sense automatic and transparent. It
should exhibit the future orientation of the project as
it unfolds. The interview data, on the other hand,
should contain more evidence of reflection, and exhibit
an orientation to the past.
Emig, however, does not make a sharp distinction
between the interview data and the protocols. In fact,
the protocol transcript for Lynn's "Terpsichordean
Greetings," (Appendix B, pages 129-136 in the book)
begins and ends with questions from the interviewer.
The session begins when the interviewer asks, "Could you
tell me a bit about what kind of contemplation you had
on the topic?" After a little more than a page of
ruminations, the protocol itself begins when Lynn asks,
"Now, did you want me to start writing?" I assume that
the researcher nods her approval at this point, as there
is no recorded response and Lynn begins immediately to
compose. Emig has given the subject the writing topic
the week before, and the purpose of the initial question
125
is to elicit data concerning the prewriting period.
Emig undoubtedly regards this question as a necessary
stimulus, and then attempts to remove herself from the
process. But the result of this procedure is that
portions of the protocol are clearly addressed to the
researcher. For example, Emig reports that at one point
in a protocol, "Lynn seems to need to confess to the
investigator that she did not do well on a verbal test."
Lynn also adds an aside at another point which explains
to the investigator the identity of a character from
Arsenic and Old Lace. This explanation does not enter
the text she is writing (Composing Processes 65). The
interviewer and the subject are clearly involved, as
Schutz would put it, in a "we-relation," and this
relation occasionally directs the writer's attention
iaway from the task to the protocol discourse, making it
!thematic. This is part of the substance of Voss's
critique of Emig's investigative process as a scientific
procedure.
The rhetorical nature of the protocols is not a
problem for Emig, because, as I noted above, in practice
she makes little or no distinction between interview
data and protocol data, except to note that they come
from different segments of the process. This treatment
would seem, however, to seriously compromise any claim
126
that the protocols parallel the actual psychic processes
involved in the production of a text in any pure way.
An examination of the protocol data in terms of the
phenomenological categories of intentionality and
reflection should reveal whether or not these
difficulties have actually compromised the validity of
results. There is, as we shall see, a noticeable
difference between the interview portions and the
composing portions of the data, at least in the one
protocol we are given in the book. Here is the
transition from interview to protocol discussed above:
This business about writing about my
grandmother, it would have been better if
I had come last week because it was sort
of fresh on my mind to do it then. But, I
don't know, this Snoopy might be
interesting if I could think of enough
examples. Yeah, I think I'll write about
that. Now, did you want me to start
writing?
Now the problem is how to start it. I
could say that I walked into the house one
evening after work and there it was
sitting in the middle of the living room.
Or else, I could say something like, it's
going to sound like a third grade
introduction, but something about, "Can
you imagine our surprise when we received
a three foot by three foot cardboard
packing thing in the mail?" That's also
not too good a start. I have to think of
how to begin the thing (Composincr
Processes 131).
Before the decision to begin composing has been made the
discourse is clearly rooted in the past. After the
127
I composing has begun, Lynn's discourse is indeed oriented
j
|toward the future, toward how this "is going to sound."
I
jThe focus of her intention remains tight, but the beam
[
i
I (to use a metaphor) illuminates first the rhetorical
i
Jgoal, then the present syntactical construction, and
then back to check on the goal. As Emig says, "There
are clear signs of efficiently divided attention, as
they focus upon the here-and-now while at the same time
considering where the future element will eventually,
and best, appear" (Composing Processes 41). The entire
composing section of the transcript exhibits this
pattern. Even the asides directed toward the researcher
are tightly bound to the rhetorical goal. This may, in
fact, be one of the necessary abilities of a fluent
writer: the ability to maintain a divided attention.
It is here that the problem of composing aloud may
cause the greatest difficulty as a research method. If
ordinary composing requires a divided and active
attention, composing aloud, to the extent that it is not
a transparent activity but a rhetorical project in
itself, necessitates a third focus for attention. It is
likely that only fluent writers, such as Lynn, will be
able to participate in this process.
Emig is aware of Lynn's focus on the rhetorical
! purpose of her writing, and marks it as unusual among
128
jthe subjects she is studying.
! Lynn proceeds as if what she writes will
I find an audience— and one wider than a
i single teacher or investigator. Note her
j use of everyone in the following sentence:
"I don't know if everyone who reads this
would get the implication." She also
seems to consider it part of her writing
task to pique and maintain the interest of
her readers. She chooses not to write a
story about her current social
difficulties with one of two young men:
"that wouldn't be too interesting to
anyone else" (Composing Processes 64).
The future rhetorical effect on an audience is one pole
of the writer's attention during the composing process.
The other is pole is rooted in the lexical and
syntactical possibilities of the current inscriptional
project. Of course, the protocol discourse makes
reference to the past as well, because the written
discourse is about past events. But the protocol
discourse is about the project, and thus is oriented
toward the future.
Prewriting, Planning, Starting, Stopping
The section of protocol quoted above also captures
the moment of decision, the acceptance of the project.
This is an important moment in the composing process.
Emig argues that there are two possible preludes to the
act of writing, prewritina and planning. She says,
Prewriting is that part of the composing
process that extends from the time a
129
writer begins to perceive selectively
( certain features of his inner and/or outer
i environment with a view to writing about
' them— usually at the instigation of a
I stimulus— to the time when he first puts
j words or phrases on paper elucidating that
i perception. Planning refers to any oral
and written establishment of elements and
parameters before and during a discursive
formulation. Prewriting occurs but once
in a writing process; planning can occur
many times (Composing Processes 39).
jThe above description would seem to say that the "act"
l
of writing is inscription, and that prewriting ends when
i
inscription begins. It is important to be clear about
what kind of distinction is being made here. If we are
trying to establish a stage model of the composing
process that runs in a linear sequence, the moment of
inscription seems to be a logical and reasonable
jsignpost marking a border between two stages. Emig
jdefines prewriting, however, in terms of a selective
perception of the environment, a narrowing of the
intentional focus in conformity with the goals of the
project. This focusing does not end with the
commencement of inscription. If anything, this focus
becomes tighter at that point. Thus it is an odd claim
to say that prewriting, or any other interactive mental
process, "happens but once." It is more likely that
prewriting processes combine with other more specific
processes at this point.
Ricoeur's more general term for what Emig calls
130
"prewriting" would be "projecting," which signifies both
the future orientation of the project and the activity
of its definition. Ricoeur then makes a distinction
between decision and voluntary motion, or action. He
cautions, however, "It was a defect of eclectic
psychology that it presented an artificial image of
reality by distinguishing different phases within the
voluntary process: deliberation, decision, and
execution” (Ricoeur FN 38). Ricoeur argues that it is a
mistake to define these categories in terms of temporal
sequence or duration. He points out that there need not
be a temporal distinction between deliberation and
decision, but a distinction between decision and action
lean be made because the interval that separates them is
!
jnot temporal but one of meaning. "It is one thing to
jindicate an action in a project, another thing to act
jbodily in conformity with the project" (Ricoeur FN 38).
Is Emig making this mistake, a mere temporal
distinction, arbitrarily cutting off the prewriting
process at the moment of inscription? It would be
better to bracket the temporal and define prewriting,
and the other categories in this outline, according to
non-temporal criteria. Emig's category after
"Prewriting" and "Planning" is "Starting," another
obviously temporal concept. Subcategories under
131
"Starting” include "Seeming Ease or Difficulty of
Decision" and "Element Treated First Discursively,"
under which she indicates that she is interested in the
reasons for the selection of that element and its
eventual placement in the completed piece. This scheme
would seem to preserve Ricoeur*s distinction between
decision and action, in that elements are clearly
developed after the moment of decision and before they
are inscribed, although elsewhere Emig identifies the
moment of "starting" with "the first placement of words
on a page" (Composing Processes 43). In the protocol
, quoted above the crucial change in the focus of the
verbalized discourse comes at the moment of decision,
the acceptance of a particular project, rather than at
the moment inscription begins, which is 16 lines later.
Of course, when inscription begins, consideration of
lexical and syntactic choices is more specific and more
constrained, but such consideration is introduced into
the activity of the project with the decision, and not
with inscription. The point is that the effort of
i
thought oriented toward a specific project is already in
the realm of action. Here is more of the protocol,
beginning where I left off above.
You could say that it * s not the ordinary
thing to have in your living room and
people don't expect to see it. Uh, you
could say something about "one of the last
132
things that people would expect to find in
our," yeah, I could say, "people in a,"
no, "in a living room such as ours with
the oak," what is it, I think it's walnut-
paneled "walnut-paneled entranceway and
shelves and book," yeah, "and book shelves
lining the walls, almost the last thing
one would expect to find would be a two-
foot-high cardboard cut-out of Snoopy, the
Peanuts dog. However, in the A_____' s
house just about anything is possible."
(laughing) I think I can start like that.
Let's see, does this pen write, yeah (4
secs, pause) One of the last things (16
secs, pause [writing]) someone would
expect to find in a now wait, to say, to
get "in our living room," I'm now
describing the living room. I either put
"a," I could say, "our living room with
its walnut-paneled entranceway and book-
lined walls," but that's kind of a, I
don't like the construction. I use it too
much, I think (Composing Processes 131).
In the above, planning, defined by Emig as "the oral or
written establishment of elements and parameters,"
clearly begins before inscription. Prewriting and
planning, defined in terms of inscription, overlap. It
is likely that the concern with the moment of
inscription in this sense is a residual objectivism in
Emig's thought, a residual concern for the product, and
is a red herring in this analysis. The more important
moment takes place in consciousness, the moment of
decision.
Emig also distinguishes between anticipating and
! planning.
Planning involves the projection of a
133
total piece of discourse; anticipating the
projection of a portion of the discourse.
Planning does not occur in the language of
the piece; anticipating often employs the
exact lexicon and syntax that will appear
in the finished piece of discourse
(Composing Processes 41).
She then quotes Jerome Bruner: "The speaker or writer
rides ahead of rather than behind the edge of his
utterance." Planning is global; anticipation is at the
level of the sentence. But is this not simply another
description of the intentional movement from the
rhetorical effect to its local manifestation, and not a
description of two separate processes or even two levels
of the same process? And there is another problem
lurking here. Both planning and anticipating seem to be
defined as categories of form. In one sense the
categories of prewriting and planning as defined by Emig
seem to represent a content and form distinction. In
fact, it is only when Emig is discussing prewriting that
she refers to the selection of what to write about, the
"selective perception of the environment." Emig may
have been led in this direction by Lynn's practice. She
says, "In fact, it is accurate to suggest that for Lynn
ialmost all major elements that fuse to form both essays
seem present, from a very early moment in the composing
process, within the foreconsciousness of the writer."
The one exception was the final line of one essay, which
Emig considers Lynn's "finest act of rhetoric"
(Composing Processes 58). Emig also notes that Lynn
always begins at the beginning, and that composing for
her is a left to right thrust without much rearrangement
of elements. Emig says, "Once into the piece, certain
recursive movements, certain pendulum actions, occur;
but Lynn enters the material once and once only, from a
given vantage; and she does not go outside again to
consider another route in" (Composing Processes 56).
Emig calls this the "marching through Georgia" effect.
Here we must be careful about assuming too much from one
case. This sort of composing process may be typical of
twelfth graders, but it is not typical of more
experienced writers, as Emig knows well. If prewriting
l
is the only mode which gathers material to write about,
it cannot be the case that it happens only once. We
need a concept of prewriting that is useful in
describing the practices of all writers.
Emig also discusses both prewriting and planning in
terms of the intervention of teachers. Prewriting is a
matter of time, and thus the due date cuts it off.
Planning is influenced by the teacher's specification of
registers, length, purpose, and audience (Composing
Processes 39). Unfortunately, there is little here to
indicate whether these distinctions are predominately
ones of content and form or of different stages or
levels of the process. In one sense, these concerns are
aspects of the involuntary, fixed obstacles in the
possibility of the project. In another sense, they are
aspects of the relation to the Other, and negotiable
within a structure of authority. Analysis within
appropriate sets of brackets could bring out important
aspects of these relationships.
We still have not quite oriented these categories.
Is "Starting1 1 synonymous with "decision" or "action"?
And when does prewriting end? Or, perhaps a better
question, what is prewriting?
In Lynn's project, the movement from decision to
!
action occurs without measurable or meaningful delay.
As Emig notes:
For Lynn, starting to write presents a
paradox. Her decision to begin is a
swift, and seemingly painless, one. Her
enactment of a first sentence, however, is
an arduous, even a torturous, matter; and
the actual time expended upon its
formulation with both prose pieces is as
long as that spent on any sentence (Emig
Composing Processes 54).
Emig's category, "starting," is a temporal concept which
blurs the distinction between decision and action, but
here her analysis is not far from Ricoeur's. The use of
"enactment" for the formulation of the first sentence
shows that she understands that action has begun.
136
Planning as defined by Emig is part of action, and
although inscription is part of the realization of the
project, the evidence of the protocol indicates that
i
1
I action on this project begins before inscription. It is
I
i not paradoxical that the formulation of the first
i
!sentence is arduous; the projecting of the whole piece
jof discourse is going on at the same time.
Curiously, Emig's description of "stopping" is more
!broadly defined than "starting."
I
Stopping represents a specifiable
moment— rather, moments— in the writing
process because, of course, a writer stops
more than once although the final
stopping, like the first starting— the
first placement of words on a page— has
special, or exaggerated, characteristics.
One stops at the ends of drafts or
j versions of a piece of writing; he stops
when he thinks the piece is finished— when
he feels he has worked through or worked
out the possibilities, contentive and
formal, that interest him in the piece; he
also stops for the purpose of presenting a
piece in a given state for the
reading— and, usually, evaluation— of one
or more others (Composing Processes 44).
These definitions are clearly tied 1) to the pedagogical
situation, and 2) to a theory of writing as expression
of the self. There is no consideration of the writing
in terms of accomplishing an action or serving any
purpose other than expression. The definition or
determination of the stopping point in terms of
"interest" reveals this. In Emig's view, the writing is
137
finished when it has exhausted the interest of the
writer, not when it is ready to accomplish its purpose.
The writer's interest, however, is related to motivation
I
and thus to the intended purpose of the writing. All of
the above definitions can be simplified into one— the
moment of stopping is the moment when the action becomes
a potential act, or, in other words, when the writing
has reached a state where its potential for achieving
the goal of the project can be considered. Emig calls
i
Ithis consideration "contemplation," which may be
misleading; contemplation implies reflection on the
object as a finished creation. To project its
effectiveness, however, the writer must become a reader
and imagine or model the future impact the text will
have on another reader, an Other.
Inscription and Reformulation
There is another distinction which has been blurred
here. Anyone who writes knows that there is an
essential difference between notes and jottings which
have been inscribed as part of the planning of a
discourse, and inscribed words which are intended to
actually end up as part of the discourse. An element
that has been written down as part of the discourse has
a certain inertia, a certain weight. The writer has
138
made a commitment that resists alteration, and the
element exists in a web of syntactical, lexical and
rhetorical relationships that will all be altered if
that one element is altered. In my own experience, this
resistance is strong even when working on a word
processor, where the physical reality of the text is
ephemeral. The mark of an experienced and able writer
is the ability to recognize when the requirements of the
project dictate that the inertia of inscribed text be
overcome, so that the project may be improved by the
abandonment or alteration of that text. Emig's
distinction between planning and anticipating involves
aspects of the distinction I am making here, but seen
from a different perspective. She says, as I have noted
above, "Planning does not occur in the language of the
piece; anticipating often employs the exact lexicon and
syntax that will appear in the finished piece of
discourse” (Composing Processes 41). Once possibly
finished language is under consideration, very specific
possibilities and relationships appear. It is
inscription which objectifies the discourse, which makes
it in a sense external, alien, and resistant. Emig
makes this distinction merely to label a sub-process and
place it in an operational hierarchy, not to describe a
phenomenological reality. In the analysis of the
139
categories above, especially in clarifying Emig's notion
of prewriting, a distinction between planning
inscription and writing inscription, inscription1 and
inscription2 if you will, would have been helpful.
The inertia of inscription2 is also relevant to the
problem of reformulation. Emig says that composing
aloud captures the behaviors of planning and of writing,
but that "partly because of the very definition of
reformulation, and partly because of the attitudes of
the twelfth graders toward this portion of the composing
process, it does not capture reformulating" I Composing
Processes 41-42). Emig defines three kinds of
reformulation: correcting, revising, and rewriting,
which involve increasingly larger segments of the
discourse. In Emig's sessions with her subjects, there
was neither time nor incentive to reformulate, but she
notes that Lynn generally uses the term "rewriting" but
means by it what Emig means by "correcting." These
distinctions are fairly standard in composition
research, but they are merely formal, and the relation
to psychic realities is obscure. All of these
operations generally refer to the reformulation of
inscribed discourse. In discussing the processes of
anticipation, however, Emig defines the nature of the
reformulation activity in terms of its temporal
140
immediacy.
There are other strategies a writer
follows in dealing with the elements or
components of discourse: he can accept,
and immediately employ, an element; he can
accept, then immediately abandon or delete
his choice (if too much time intervenes,
the action becomes reformulation or
revision); or he can combine the element
in some way with other elements in the
discourse ( Composing Processes 41).
In a footnote to the above Emig remarks, "The kind of
self-censoring that eliminates an option before it is
uttered is outside the purview of this inquiry"
(Composing Processes 41). Here she seems to be thinking
of reformulation in terms of uninscribed possibilities,
which would seem from the perspective I have been
developing here to be an entirely different matter.
Again, we are being confused by merely temporal
distinctions. It is not duration which distinguishes
choice from revision, but inscription. Inscribed
discourse has become objectified and resistant, and
student writers need powerful reasons in order to
overcome it.
The project is oriented toward the future act it
will accomplish. Thus reformulation must also be
defined in terms of that act. Writing is a complex act
in this aspect; it accomplishes its aim, or fails in it,
when it is read, or each time it is read. Thus the act
can be re-projected and entirely or partially re-enacted
141
before, or even after, it has found an audience. The
re-projected act can be changed; the means to achieve
the former act can be changed; the semiotic embodiment
of the means can be corrected, enhanced, or tuned. Even
the term "reformulation" implies that only the form of
the act is changed, but form is perhaps the most trivial
aspect of re-projecting. To some degree, however
slightly, re-projecting involves a return to the
prewriting or projecting mode of consciousness.
Decision, Motive, and Value
Emig is aware of the difficulties with her
descriptions of "starting," but sees the problem as one
of access to data rather than as a categorical
obscurity. She says "Starting is a specifiable moment
in the process of writing— and the one perhaps most
resistant to logical characterization and analysis."
She invokes the Freudian tri-partite self, not without
some doubt, and says that starting can be defined in
Freudian terms, as R.N. Wilson does, as the moment when
the id, or the unconscious breaks through the controls
usually exerted by the ego and super-ego. "Because of
the clearly profound, and opaque, nature of this moment,
the kinds of elements that can be accurately specified,
that exhibit themselves in behavior, are
142
contextual— and, usually, trivial" (Composing Processes
40). Emig's examples here include such things as the
physical location of the writer when he begins, habits
jand rituals he observes, what element is first
I inscribed, and where that element ends up in the
i
1 finished piece. The paradox which Emig defines here,
!
[the profound nature of the moment and the trivial nature
I
of the objective indications of that moment, is an
1
I indication that she is not bringing the proper tools to
)
bear on the problem.
The Freudian elements give a sense of the beginning
of a project as a throwing off of social and personal
inhibitions, as if the constraints of the ego and super
ego forbid an irruption of the new. This is a
thoroughly Romantic view of literary creativity, more
appropriate perhaps to a poet who is attempting to
challenge society in some new way than to the practical
concerns of student or professional writing. The
Romantic view is nearly always more successful in
mystifying the creative process than in elucidating it.
And for Ricoeur, if this model is true at all, it is
[
|only half true.
i Ricoeur's description of the moment of decision has
a number of themes. Ricoeur tends to use "decision"
when he has bracketed the temporal, and "choice" when
143
the brackets around time have been removed and he is
considering the sequence of elements. First, we will
examine the non-temporal perspective. Ricoeur does not
see the decision as an irruption of the new.
The project, though it is a novel event,
is not a novel structure which appears
suddenly at the conclusion of an interior
process which makes no allowance for it.
Hesitation is a testing of multiple
projects. Consequently choice is not
j created by an irruption of a projecting
; consciousness but by the simplification of
! a hesitating one (Ricoeur FN 168).
i
There is ample evidence of this sort of process in the
interview that precedes Lynn's protocol. Lynn considers
four different projects and their reasons. One topic is
"Old People on the Bus" which she considers at times
separately and at times in combination with a topic
concerning her own grandmother's advancing age. At one
point in the discussion she characterizes this topic as
"the best of the lot." Another topic she considers
concerns two boyfriends. She decides this "would be
just too trite," and she thinks the audience for it
would be too narrow. Then there is Snoopy the cardboard
dog, the topic she ultimately chooses. Emig asks why
Lynn chooses this topic, the cardboard dog, when the
other topics seem to offer more personal engagement and
more complexity. As we have seen, the answer lies not
j
I in a stimulus, but in a motivational structure.
!
| ________________ 144
Motives, however, are so intimately related to a project
that they cannot be understood separately.
While unable to form a definite project I
do not cease to be a consciousness
i absorbed in a diversity of practical aims
I representing actions which depend on me.
I examine motives by referring to such
tentative projects. In this respect, the
intentional structure of a hesitating
consciousness differs from that of a
deciding consciousness only in the
modality of the projects among which
consciousness is divided (Ricoeur FN 139).
Is prewriting then, a mode of consciousness, a
hesitating consciousness considering tentative projects
through which it examines motives, a mode which resolves
into a deciding consciousness upon choosing a project?
If so, the mark of the termination of prewriting is
clearly the moment of decision and not the beginning of
inscription. Ricoeur develops a concept of "practical
mediation" between motive and project, which are
reciprocal in that "the developing of my choice and the
maturing of my reasons are one and the same thing . . .
For a hesitating consciousness there is a wavering
motivation which accompanies a wavering project, and I
i
i
:fulfill my designs in fulfilling my reasons" (Ricoeur
iPhilosophy of Paul Ricoeur 6).
It is still more complex, however, for there is yet
l another layer of reciprocity behind the dialogue of
I
|project and motivation. Beyond the will and its project
145
is the level of the social and the relation of motive to
I value.
I
i
I encounter values only in a process of
motivation, and because there is no motive
outside of the reciprocal relation with an
emerging decision. On the one side, I
have to say that it is of the essence of a
volition to try to find supporting
reasons; but in its turn this special
receptivity by which any value appears to
me is inseparable from a consciousness
which gropes toward a choice. If it is
thus of the essence of a volition that it
legitimate itself in motives which make
values emerge for me here and now, it is,
reciprocally, of the essence of these
values that they emerge only as possible
motives for a decision which fulfills
their meaning by grounding itself on them
(Ricoeur Philosophy of Paul Ricoeur 6).
Motives, and values as motives, appear to consciousness
only in connection with a project. The following
diagram represents a schematic, non-temporal view of the
structure of Lynn's decision, and summarizes the points
I have made above:
146
TENTATIVE PROJECTS
VALUES
DECISION
ASSIGNMENT
PROJECTED
ACTION OR
GOAL
CAPABILITY/POSSIBILITY
Time, Knowledge, Skill
etc.
MOTIVES
Please Researcher,
Interest Audience,
Interest Self,
Reduce Effort, etc.
Old Ladies on the Bus
Grandmother's Advancing Age
Two Boyfriends
Snoopy the Cardboard Dog
Figure 1: The Intentional Structure of a Decision
I might also call this the intentional structure of
prewriting. The above is not a flow chart; the
relationships depicted above are reciprocal and
]
j interactive. Values and the assignment are depicted as
l
i being somewhat more external than other elements, but
| both are part of the motivational structure. The
j assignment is the "cause" of the writing only in the
1 , , ,
j sense that the writing may not have been done without it
i and the structure of authority it represents. That the
1
i
assignment influences possibilities, motives, tentative
projects and the decision itself is clear, but the
assignment must be integrated with the rest of the
147
motivational structure. If it is not, it may be that no
writing will result; not all assignments become
projects, and not all projects are enacted.
I might also call the above diagram the structure
of hesitation in the composing process. The complexity
of this structure precludes any sort of analysis in the
manner of a mental physics. Ricoeur says,
Intellectualism pretends to believe that
only one series of thoughts unfolds its
implications in a section of the process
under consideration: the truth is that we
must constantly orient ourselves in a maze
of intersections and badly marked roads.
Practical problems are only rarely
amenable to obvious interpretation. The
order of action is the order of the
probable. The heterogeneity of the values
involved and the confusion of aspects of a
situation make the problems of shifting
more important than the problems of
continuity, discontinuous passages more
numerous than logical connections (FN
159) .
Ricoeur argues that "control over process is attention
in motion; choice in a sense is a fixing of attention"
(FN 149). If we were able to follow the flow of the
composing process we would see attention flutter from
goal to this motive and that motive and back again until
hesitation resolves into choice; then the process
continues for yet another choice. Emig's protocols
occasionally reveal this movement; more often it is
hidden behind what Emig calls "hesitation behaviors":
I filler sounds, critical comments, expressions of
1
148
feelings and attitudes, digressions, repetition of
[elements. There is silence filled by inscription
I
activity, and seemingly unfilled silence, in which the
writer "may at these times be engaged in very important
nonexternalized thinking and composing" I Composing
Processes 42) . Of course it is these unfilled silences
which may be most important in understanding the
composing process. Emig is aware of the importance of
I
[attention in this regard. She says of Lynn:
At times she employs morphemes of low
semantic content. There are instances of
the use of "et cetera," "blah, blah,
blah," and "such-and-such" to fill pauses.
Each seems to serve a different function
. . . . The single instance of
"such-and-such" is a substitute for
"Woolworth's" a name Lynn cannot for the
moment remember because her attention
seems focused on a more immediate
composing problem (Composing Processes
62-63).
Emig's interpretation of the motivational structures
behind decisions, and of the processes hidden behind
hesitation behaviors and filler sounds such as those
above, rests on slim evidence, a fact she indicates by
using "seems." Protocols do not provide much access to
motivational structures or the processes of habitual
action. In a later chapter I will discuss other methods
which may provide more access.
I have argued along with Ricoeur that rather than
the will bursting forth from the constraints of the ego
L_.
149
and super-ego, it considers both capabilities and
possibilities, and founds the decision in part on social
values and in part on personal motives, which are
i entangled and in some sense indistinguishable. This
| leads to a double notion of the act: it is both an
assertion of the will and a reception of values.
Ricoeur argues that if we consider the birth and
development of a decision, in other words if we consider
!
a decision in its temporal sequence instead of
considering it abstractly, we see consciousness
oscillate between hesitation and choice.
The occurrence of a choice can always be
read in two different fashions; it is in
one sense attention which fixes upon a
group of motives: I make up my mind
because I yield to this or that reason;
but in another sense it is the upsurge of
a new act: the final meaning of my reasons
is that I so decide. But this double
reading is inscribed in the very structure
of a decision which, on the one hand, is
an invention of the project and, on the
other hand, a reception of values, an
activity and a receptivity (Ricoeur
Philosophy of Paul Ricoeur 18-19).
According to Ricoeur, these two interpretations have led
to two philosophies of freedom, but he maintains that
the two possibilities always remain before us. This
results in what Ricoeur calls "the synthesis of
lawfulness and inventiveness."
For the unity of these two it would be
necessary for choice to satisfy at once
both lawfulness and inventiveness, value
150
and boldness of existing. A fine cleavage
runs through our freedom precisely because
it is active and receptive, because it is
a human freedom and not a creative "fiat."
And that is why I would say that the
synthesis of lawfulness and inventiveness
I remains a limiting idea (Ricoeur
j Philosophy of Paul Ricoeur 18-19).
This is the sense in which the Freudian analysis Emig
gets through R.N. Wilson reveals only half of the
dialogue. Because our project is an act in society and
has social purposes, we are constrained by values. The
act is both inventive and obedient, and is
simultaneously both motivated and justified. The
reciprocal, interactive nature of these relationships
does not lend itself to objective analysis as part of a
process; this is why Emig characterizes this moment as
"opaque," and "resistant to logical characterization."
Part of Emig's difficulty in understanding the
decision, however, comes from the contradictory goals
1
[motivating Emig's project and Lynn's. Emig is a teacher
investigating students; her encouragement of reflexive
writing is motivated by a desire to help a student grow
and a desire to know the life and mind of that student
better. She interprets Lynn's reluctance
psychologically.
The interesting question here is to define
what for Lynn is an "easy" subject and
what is a "hard" one. Clearly, an "easy"
one is a nonpersonal subject, one that
does not demand interacting with her
l __
151
feelings, one that is not reflexive
(Composing Processes 49).
Of course it is possible that Lynn wishes to avoid
discussing her feelings. Emig develops this picture of
Lynn in some detail. None of her subjects wished to
write in this mode, however, and Lynn's given reasons
for her choices are always both rhetorical and
reasonable. Lynn's project is a rhetorical one; her
goal is to produce a piece of writing that will interest
an audience and thus is necessarily directed toward an
Other. Social values discourage self-centered, self-
indulgent expression. Lynn, as Emig observes, is
writing for an audience larger than one evaluator. She
applies the social values as she sees them. To ask her
to do otherwise is to counter her rhetorical intention.
Lynn's deliberation over the more reflexive topics may
have been an essential part of her prewriting activity,
! however, and, as I have argued, it is in the prewriting
mode that reflection and exploration of self is probably
most productive. Teachers who encourage the so-called
reflexive mode of writing over the extensive should take
these contradictions into account.
A Redefinition of Emig's Categories
i
Thus far we have seen that although the category
"Stimulus" reflects behavioristic confusions of motive
152
and cause, Emig is aware of the future orientation of
the composing process, and the fact that the writer must
preserve a divided attention between the future
!
rhetorical effect and the ongoing development of the
language that will achieve that effect. We have also
discovered that although there are certain problems with
the composing aloud method, by and large the one
protocol supplied by Emig exhibits the intentional
characteristics our phenomenological perspective would
predict. We have found an unresolved confusion in the
descriptions of planning and anticipation, and we have
seen that although Emig's category, "Starting," obscures
the more fundamental categories of decision and action,
her analysis basically recognizes this distinction. We
have identified two types of inscription and redefined
reformulation in terms of changes in projected action
and means, rather than merely in terms of form. We have
explored the reciprocity of motive and project, and of
motive and value, in achieving the moment of decision.
j and we have redefined prewriting in terms of this
i
! moment, rather than in relation to the moment of
inscription. We have also found that what Emig calls
"prewriting" might best be characterized as a mode of
hesitating consciousness, weighing motives, values and
possibilities in an intentional relation with tentative
153
composing projects. By and large, in spite of the fact
that a number of problems have become evident, Emig's
early work appears useful and productive from this
phenomenological perspective, especially considering the
powerful influence of scientism prevalent at the time it
was done. The following chart compares Emig’s
categories with my phenomenologically redifined ones:
154
EMIG'S TEMPORAL
CATEGORIES
i
Context.......
I
I
Stimulus......
Prewriting....
Planning
Starting
Composing
Reformulation
Stopping
Contemplation
Teacher
Influence
Figure 2: Temporal and
NON-TEMPORAL CATEGORIES
Context
Projected Goal or Act;
Associated Motives
Projecting of Tentative
Projects; Testing of
Possibilities and Motives
Projecting of a Specific
Project
Decision— Choice of
Specific Project
Action— To Act in
Conformity with the
Project
Change or Clarify Goal;
Change or Clarify Means
Action Becomes Potential
Act
Writer becomes Other;
Project effect of text
Relation to Other as
Audience and Authority
Non-temporal Categories
Habit and the Accessibility
of the Composing Process
There remains one important aspect of the composing
process to investigate, an aspect which the think-aloud
155
protocols leave obscure for the most part. This is the
influence of habit and automated structures.
For the most part, Emig's speculations about habit
concern style, and she discusses style in terms of
"favored transformations." She quotes Richard Ohmann
who gives the following definition: "Style is in part a
characteristic way of deploying the transformational
apparatus of a language" ("Generative Grammars and the
Concept of Literary Style" Word (1964) 431). She goes
on to say:
There is no reason to believe that
nonprofessional writers do not also have
their characteristic ways "of deploying
the transformational apparatus of a
language," although these way may be less
striking, with less reliance on "a very
small amount of grammatical apparatus"
fComposing Processes 42).
Emig asks, "When teachers or critics say that a writer
has 'no style,' is what they mean that the writer in
question has no strongly favored ways of transforming?"
The limitations of this view of style are more obvious
today than they were in 1970 when transformational
grammar seemed to be on the verge of explaining
everything about language and language learning. This
view reduces style to a set of mechanical and habitual
grammatical patterns without consideration of audience
or purpose. Emig then connects this view of style with
a behaviorist view of action.
156
The next question, of course, becomes why
one favors a given cluster of transforms.
One explanation seems to be that a writer
is following some sort of "program" of
style, a series of principles, implied or
explicit, of what constitutes "good"
writing. For example, he might break the
concept "coherence" into a set of
behavioral objectives, such as "Be clear
about referents" and "Repeat necessary
lexical elements" (Composing Processes
42) .
In this context, however, clarity and necessity can only
i
jbe evaluated with respect to an audience, and thus these
are not behavioral objectives but rhetorical strategies.
We are here being led away from a concern with process
toward a renewed concern with product. If an explicit
"program" such as this would effectively produce
"coherent" writing, then there is nothing wrong with the
handbooks and rhetorics which preceded Emig's work, and
which she wishes to overthrow.
Such an analysis reveals only the mechanical
aspects of habit, which is both mechanical and vital.
Habit is the essence of learning; in it lies both the
source of our powers and the dangers of ossification and
automatism. Ricoeur calls habit "the fruit of what has
t
endured" (Philosophy 11). Habit is mysterious in that
it is the voluntary become involuntary, both in
acquisition and use, at least to an extent. Ricoeur
j says:
But while no human habit is acquired
157
entirely apart from the will, the will
does not have a direct power of
constituting habits: it can only activate
or hinder a specific formative function
which we might well call involuntary:
practice is what has this "spontaneous"
power— in itself not willed— of
disengaging from the activity of
perception, movement, imaginative
evocation, judgment, etc. forms which to
the extent to which they are "segregated"
become assimilated to my activity as such,
precede new operations, and ingress in the
living duration which accompanies my
present. Among these are the
"preperceptions" which aid me in orienting
myself among new objects of perception,
the "preconceptions" and "prenotions"
which, by condensing previous thoughts,
become in their turn thinking rather than
thought, the forms of motion which become
detached from learning a particular
movement and facilitate the exercise of
related movements, etc (FN 282).
Ricoeur sees habit as a kind of nature of the will, a
second nature, an internal nature. Habit functions in
an involuntary way like a biological organ such as the
heart or the stomach. Yet it is an acquired nature and
a pliant one. This produces an enigma. Using a habit
is like moving your arm; the motion is voluntary, but
the means— the synapses, nerves, individual muscles—
exist below the level of awareness.
It is a characteristic of habit that it
brings about the same enigma by leaving
behind the intelligence and the will which
might have presided at its construction.
But it is because acts which were never
constructed by knowledge and willing
already establish a bond of the motorium
and the sensorium that habit can in its
turn endow us with powers we find
158
mysterious. It does no more than to
extend that primitive relation to our body
which precedes all knowledge and all
willing relevant to the structure of
movement. I neither know nor will the
structure of what I am able to do in
detail.
Habit thus does not introduce
radically new facts. Thanks to the
passage of time, to repetition, it
indefinitely enlarges body's irreflective
usage. What once was analyzed, thought,
and willed drifts bit by bit into the
realm of that which I have never known or
willed . . . I do not think the movement,
I make use of it (Ricoeur FN 285-86).
I
I
] Because the habitual is preconscious, the use of the
habitual will not appear in a think-aloud protocol. Yet
it is clear that a great part of the composing process,
perhaps by far the greatest part, is habitual. Emig is
aware of the power of habit in composing, but for her it
is negative, something to be avoided or overcome. She
cites the linguist Leon A Jakobovits, who suggests that
"stale art" is algorithmic— produced by "a computational
device that specifies the order and nature of the steps
to be followed in the generation of a sequence"
("Rhetoric and Stylistics: Some Basic Issues in the
Analysis of Discourse," CCC (December 1969), p. 325.).
Emig's example of this is the five paragraph essay,
which she calls "algorithmic, or so mechanical that a
computer could readily be programmed to produce it: when
a student is hurried or anxious, he simply reverts or
I
j regresses to the only program he knows, as if inserting
i
i
I 159
a single card into his brain" (Composing Processes 51).
However, without habit we could not write at all, and it
i
I
is likely that the reliance on the algorithmic five
paragraph essay derives from a lack of useful habits
rather than an overreliance on habit itself. Emig is
closer to the truth about habit when she quotes an
interview with John Ciardi about his process of
composition. The interviewer asks how he goes about
i
composing a poem. He says
You're asking for lies. It's inevitable.
I've been asked to do this over and over
again, and lies come out. Let me put it
this way. The least a poem can be is an
act of skill. An act of skill is one in
which you have to do more things at one
time than you have time to think about.
Riding a bike is an act of skill. If you
stop to think of what you're doing at each
of the balances, you'd fall off the bike.
Then someone would come along and ask you
to rationalize what you thought you were
doing. Well, you write a poem. And
someone comes along and asks you to
rationalize what you thought you were
doing. You pick out a theme and you're
hung with trying to be consistent with the
theme you've chosen. You have to doubt
every explanation .... You have to end
up lying. You know that you had something
in your mind, but you can never get it
straight (John Ciardi, Interview from
Counterpoint. compiled and edited by Roy
Newquist, pp. 122-23. Copyright 1964 by
Rand McNally) (Composing Processes 10).
Emig quotes this to show that retrospective accounts do
not illuminate the composing process very much because
of the problem of "unintentional lies," but she misses
160
what Ciardi is really saying here. Writing depends on
simultaneous, semi-automatic processes that are not
entirely conscious. In this sense protocols have the
same problem. A writer cannot externalize unconscious
or preconscious processes.
Thus habit is the power that allows us to act and
I
to improve. It is the basis of both skill and creation.
And the interaction of habit and intention is both
mysterious and surprising. Ricoeur says:
A habit only grows through tentative
probes in all directions which are not,
strictly speaking, willed. We ''pick up"
manual and mental patterns without
actually knowing how. These are the lucky
breaks which are always disconcerting to
our conscious effort. All the monographs
on acquisitions of habits point to this
curious relation between the intention
which launches out in a specific direction
and the response, arising from the body
and the mind, which always has the air of
an improvisation. This is familiar in the
case of skaters, pianists, and even
aspiring writers (FN 289) .
Talk-aloud protocols will reveal the movement of
attention, the strategies of consciousness. But this is
only the surface of much deeper and more subtle
processes. We may be able to intuit and understand
something of these submerged movements, but it will not
be possible in the terms of objective, physical science.
Janet Emig has made a good beginning. In the next
chapter I will analyze the influential research of Linda
161
Flower and John Hayes to see if they have overcome, or
exacerbated, the difficulties found in Emig's initial
effort.
162
CHAPTER THREE
LINDA FLOWER AND JOHN R. HAYES:
DISCOVERING THE LIMITS OF THE COGNITIVE APPROACH
i
i
| If it is true that Janet Emig introduced "science
[
1
jconsciousness" into composition studies, it is also
l
true, as we have seen, that her work is eclectic in
method and not strictly "scientific." The work of Linda
Flower and John Hayes, based on the assumptions and
experimental practices of cognitive psychology, has been
much more consistent in both conceptualization and
method. Thus they call Emig's work, "valuable, but
rather impressionistic," and disavow Emig as a source
for their own style of protocol analysis ("Response to
[Cooper and Holzman" 96). Flower and Hayes appear to
have employed rigorous scientific procedures in order to
obtain interesting and useful information about the
composing process. This combination has been all too
rare in composition research, and this fact undoubtedly
accounts for the wide-ranging influence their work has
[had in the composition community. In this chapter I
163
will investigate the nature and limits of the cognitive
approach, the degree to which Flower and Hayes conform
to the norms and standards of cognitive science, and
both the validity and the utility of the model of the
writing process they produce.
Flower and Hayes state that they want to study
writing as a "problem-solving, cognitive process," and
that they see writing as a solution to a "self-defined
problem or goal" ("Cognition of Discovery" 23). This is
quite different from Emig's conception, and much closer
to the future-oriented concept of the project that I
have developed in chapter two. Flower and Hayes also
argue against the concept of prewriting and the
resulting "stage model" of the writing process.
The critical assumption of Pre-Writing is
that writing is a stage process; that is,
the act of composition proceeds in a
series of relatively discrete stages. . .
Pre-Writing is the stage, untouched by
pen, paper, or rhetorical purpose, in
which you discover your ideas. There are
various avenues to this discovery.
Depending on your instructor, they can
range from library research, to sensory
experience, to meditation and yoga. The
great social value of the Pre-Writing
notion is that it legitimizes thinking as
a preliminary to immediate paper
production. But as a model of the
composing process, it still leaves the
writer in the same bind; the task of
writing is somehow separated from thinking
but is still unexplained ("The Dynamics of
Composing" 32).
In chapter two I criticized the concept of prewriting
164
i for much the same reason. The "pre" of "pre-writing" is
♦
\
' a temporal concept that implies a sequence of stages.
I
j
jFlower and Hayes also claim that stage models of
j composing are based on a study of the product, because
j inscription produces a product while thinking does not.
I have shown that this is true in Emig's case, where
concern with the moment of inscription led to a neglect
of the moment of decision. Emig, however, while
producing a model which appears to be a stage model,
also calls the writing process "laminated" and
"recursive," and is clearly aware of the limitations of
a single-layered description. One of the goals of the
protocol-based research pursued by Flower and Hayes is
to produce a model that adequately accounts for the
recursiveness observed by Emig and other researchers.
The Flower and Hayes model consists of three major
processes: planning. translating, and reviewing, and
associated sub-processes.
The function of the planning process is to
take information from the task environment
and from long-term memory and to use it to
set goals and to establish a writing plan
to guide the production of a text that
> will meet those goals. The plan may be
drawn in part from long-term memory or may
be formed anew within the planning
process. The translating process acts
under the guidance of the writing plan to
produce language corresponding to
information in the writer's memory. The
function of the reviewing process, which
consists of reading and editing
165 |
subprocesses, is to improve the quality of
the text produced by the translating
process. It does this by detecting and
correcting weaknesses in the text with
respect to language conventions and
accuracy of meaning, and by evaluating the
extent to which the text accomplishes the
writer's goals ("Identifying the
Organization" 12).
In a number of ways this model is clearly superior to
that offered by Emig. It organizes the major operations
of the composing process into interconnected units. It
is not a stage model and avoids confusing temporal
I
| categories. It allows recursive and embedded
operations. It is closely linked to the evidence of the
protocols, and has a kind of intuitive validity as well.
However, while the model carefully specifies where
information comes from, what happens to it, and where it
goes from there, it does not address the question of how
these operations are accomplished. Information is
retrieved and fed into a flow chart box which says
something like "evaluate retrieved element," but how
this evaluation takes place remains mysterious. Is the
cognitive method inappropriate to address questions like
this, or is it simply that more work needs to be done?
In order to answer this question a brief investigation
of the history of cognitive science is necessary.
I
' To a certain extent cognitive psychology is a
reaction to the problems and limitations of behaviorist
166
thought. As Howard Gardner notes, behaviorists
maintained that "those interested in a science of
behavior ought to focus exclusively on behavior:
researchers ought assiduously to eschew such topics as
mind, thinking, or imagination and such concepts as
plans, desires, or intentions" (The Mind's New Science
11). Behaviorism also excluded hypothetical mental
constructs like symbols, ideas, schemas, or other
possible forms of mental representation. These
reductions and exclusions made it difficult for
behaviorists to account for language acquisition or use,
and during the time of their ascendancy, such questions
remained for the most part uninvestigated. According to
Gardner, this situation began to change in September,
1956, when the Symposium on Information Theory was held
at the Massachusetts Institute of Technology. A number
of important firsts were presented at this conference.
Alan Newell and Herbert Simon presented the first
logical proof ever carried out by a computing machine,
Noam Chomsky outlined "Three Models of Language," and
George Miller delivered a paper outlining his claim that
the capacity of human short-term memory was limited to
approximately seven entries (The Mind's New Science 28).
Four years later in 1960, George Miller, with his
colleagues Karl Pribram, a neuroscientist, and Eugene
Galanter, a mathematically oriented psychologist,
published Plans and the Structure of Behavior, a book
that had a tremendous impact on psychology and allied
fields.
In it the authors sounded the death knell
for standard behaviorism with its
discredited reflex arc and, instead,
called for a cybernetic approach to
behavior in terms of actions, feedback
loops, and readjustments of action in the
j light of feedback. To replace the reflex
! arc, they proposed a unit of activity
called a "TOTE unit" (for
"Test-Operate-Test-Exit"): an important
property of a TOTE unit was that it could
itself be embedded within the hierarchical
structure of an encompassing TOTE unit.
As a vehicle for conceptualizing such TOTE
units, the authors selected the computer
with its programs (Gardner The Mind's New
Science 33).
As we shall see, this TOTE unit is an integral part of
sub-process models in Flower and Hayes, though the term
itself is never used. If we compare the above model
with the conclusions or "facts" about writing that
Flower and Hayes published in 1980, there is nearly
perfect correspondence:
I
i 1. Writing is goal directed,
j 2. Writing processes are hierarchically
organized.
3. Some writing processes may interrupt
others over which they have priority.
4. Writing processes may be organized
recursively.
5. Writing goals may be modified as
writing proceeds ("Writing as Problem
Solving" 396).
Point number one is not related to the TOTE unit, but to
168
problem-solving research that came about later in the
same cognitive tradition. However, the hierarchical
organization of processes, the possibility of embedding
and recursiveness, and the modification of goals through
feedback are all part of this early model. Because
these conclusions, except for point number one, are so
similar to a basic cognitive model published twenty
years before, it is fair to ask whether they actually
arise from a study and analysis of composing protocols
or whether they have been imposed upon this data. In
addition, as I mentioned above, there is in the
community of cognitive psychology a substantial body of
protocol-based research in problem-solving. Is the view
that writing is problem-solving a discovery, or simply a
perspective that comes along with the importation of the
cognitive research method? Cooper and Holzman raise
both of these concerns in their critique ("Talking about
Protocols" 284). Flower and Hayes do not answer this
question directly, although they acknowledge that their
I
|method is derived from this tradition. In this
particular article ("Writing as Problem-Solving"), the
evidence for the assertion that writing is goal-directed
is simply a series of quotations from protocols. There
is no attempt to give percentages or rank-ordering of
features, or to indicate how many protocols they looked
at. Ericsson and Simon say of this sort of procedure,
Protocol analysis sometimes consists
largely of extracting examples of concepts
or processes from the verbalization as
evidence for theories that predict them.
As Bayes' Theorem shows us, the
contribution of such examples to the
credibility of a theory depends on what
alternative explanations are available
(Protocol Analysis 287).
To illustrate this idea they discuss experiments
concerning language acquisition in apes, arguing that
[ the significance of Washoe combining the signs for water
i and bird to refer to a duck is dependent on whether the
chimpanzee had also produced water banana, water shoe,
etc.
The appearance of scientific rigor without a
presentation of the necessary details is often a feature
of the published work of Flower and Hayes. This is not
a fatal defect in itself because their methodology is
not exactly experimental, in that there is little or no
attempt to isolate or control variables. However, the
information they present and the speculations they
engage in would be more meaningful if w© knew the size
and nature of the population of subjects and more about
the circumstances under which the protocols were
| gathered. The articles are generally written as if they
I
| are scientific articles for non-scientists. Marilyn
I Cooper and Michael Holzman observe that the published
170
articles of Flower and Hayes are apparently based on
only four transcriptions ("Talking About Protocols"
289). Flower and Hayes respond that they "have
collected about 60 writing protocols for analysis . . .
and another seventy in sessions training researchers"
("Response to Cooper and Holzman" 95). They do not,
however, say that they have transcribed and analyzed
j these protocols. Whatever the truth is, this confusion
j indicates the lack of clarity with which the results
I
j have been presented.
I will take up these difficulties in greater detail
a bit later. The point I would like to raise now is
whether the work done by Flower and Hayes is meant to
stand up as actual research in cognitive science or as a
preliminary exploration of certain composition issues
from a cognitive perspective— also a legitimate
endeavor, but if this is the case we must be clear about
it. To answer this question, we must begin by
establishing what the assumptions and practices of
cognitive science are.
Howard Gardner identifies five features generally
associated with cognitive approaches. These are: 1) The
belief that in talking about human cognitive activities
it is necessary to speak about mental representations
and to posit a level of analysis wholly separate from
171
I the biological and neurological on the one hand and the
I
sociological or cultural on the other. 2) The tendency
to view the computer as the most viable model of how the
human mind functions. 3) The de-emphasizing of
affective, emotional, historical, cultural and
contextual factors. 4) The faith that much is to be
gained from interdisciplinary studies. 5) The somewhat
more controversial idea that cognitive science will
address the main issues and concerns of the western
philosophical tradition (The Mind*s New Science 7).
The first point, the acceptance of a level of
representation in human cognitive activity, is an
important difference from behaviorist theories, and
allows for important connections with the
phenomenological perspective I have been developing in
the first two chapters. In fact, as point 5 indicates,
cognitive science sees itself as a way of performing
empirical investigations of philosophical claims. As
Gardner notes, some cognitivists see their approach as
replacing philosophy. Gardner himself, however,
disagrees.
With science to answer the philosophical
agenda, what is left for philosophers to
do? I shall argue, however, that this
conclusion is wrong. I see at work a
1 dialectic process, where philosophers
propose certain issues, empirical
disciplines arise in an attempt to answer
them, and then philosophers cooperate with
172
empirical scientists in interpreting the
results and in proposing new lines for
work (The Mind's New Science 54).
Gardner's second point, concerning the computer as
model of the mind, is another connection, albeit an
acrimonious one, between cognitive science and
philosophy. For cognitive scientists, the computer as
mind is not simply a metaphor borrowed from another
discipline. Early cognitivists, such as Alan Newell and
Herbert Simon, were also early computer programers, and
work on human problem-solving went on in parallel with
work on machine problem-solving. Newell and Simon
devised a computer program called the "General Problem
Solver," which could solve theorems, play chess, and
solve puzzles. The GPS sought to mimic the processes
used by normal human subjects in tackling such problems
rather than solving them in the most efficient way.
J "Thus, a very important part of this research enterprise
| was the collection of protocols that recorded the
I
! introspections and notations of subjects engaged in
i
problem solving" (Gardner The Mind1s New Science 148).
I Think-aloud protocols were first used in a systematic
i
1
; attempt to model human thought processes on a computer,
I and thus are closely associated with the beginnings of
attempts at Artificial Intelligence.
i
Gardner also notes that Hilary Putnam argued that
with the development the so-called "Turing test” for a
successful artificial intelligence (Alan Turing's notion
of a machine that could be mistaken for a human
conversationalist), and the subsequent invention of the
computer, the classical mind-body problem was nearly
!solved— or "dissolved," as Howard Gardner puts it. "It
i
Jwas apparent that different programs, on the same or
different computers, could carry out structurally
identical problem-solving operations" (The Mind's New
Science 31). The argument is that the software could be
described apart from the particular hardware on which it
happened to be implemented. The implication of this
argument is that the mind is simply software. From this
perspective, given electronic hardware of sufficient
power and complexity, the software of the mind could be
made to run on it.
Hubert Dreyfus has been the leader of the
; philosophical attack on Artificial Intelligence. In his
i
i
j book What Computers Can't Do; A Critique of Artificial
Reason, Dreyfus argues that there are fundamental
differences between human and machine intelligence. He
says that AI researchers "assume that man must be a
device which calculates according to rules on data which
take the form of atomic facts."
Such a view is the tidal wave produced by
the confluence of two powerful streams:
first, the Platonic reduction of all
reasoning to explicit rules and the world
I to atomic facts to which alone such rules
! could be applied without the risks of
interpretation; second, the invention of
the digital computer, a general-purpose
information-processing device, which
calculates according to explicit rules and
takes data in terms of atomic elements
logically independent of one another. In
some other culture, the digital computer
would most likely have seemed an
unpromising model for the creation of
artificial reason, but in our tradition
the computer seems to be the very paradigm
of logical intelligence, merely awaiting
the proper program to accede to man's
essential attribute of rationality
(Dreyfus What Computers Can't Do 2 31).
Dreyfus argues that unlike computers, humans have a
fringe consciousness, and a tolerance for ambiguity, and
that this indeterminacy plays a crucial role in human
j perception. He cites Merleau-Ponty, who pointed out
i
! that most of what we experience must remain in the
: background so that something can be perceived in the
I
1 foreground. This is the ground, or "outer horizon" as
|
! Husserl called it.
I
j Similarly, our sense of the overall
i context may organize and direct our
j perception of the details when we
i understand a sentence. For a computer,
! which must take up every bit of
information explicitly or not at all,
there could be no outer horizon. Any
information taken into account would have
to be as determinate as the figure
(Dreyfus What Computers Can't Do 241).
Dreyfus' original philosophical response to the project
of artificial intelligence was that consciousness was
175
inseparable from embodiment, and that these researchers
(predominately Herbert Simon and Alan Newell) ought to
begin by studying Merleau-Ponty. Dreyfus argues, "If
the body turns out to be indispensable for intelligent
behavior, then we shall have to ask whether the body can
be simulated on a heuristically programmed digital
computer. If not, then the project of artificial
iintelligence is doomed from the start" (What Computers
j
Can't Do 235).
Dreyfus point out that "our sense of the whole
situation, outer horizon, and our past experience with
the specific object or pattern in question, inner
horizon, give us a sense of the whole and guide us in
filling in the details." This is to say that data are
interpreted according to context and expectations, and
that this is more than information processing. He
provides a striking phenomenological example.
This process can be best noticed when it
is breaking down. If you reach for a
glass of water and get milk by mistake, on
taking a sip your first reaction is total
disorientation. You don't taste water,
but you don't taste milk either. You have
a mouthful that approaches what Husserl
would call pure sensuous matter or hyletic
data, and naturally you want to spit it
out .... A computer, which must operate
| on completely determinate data according
to strictly defined rules, could at best
be programmed to try out a series of
hypotheses to see which best fit the fixed
data. But this is far from the flexible
interaction of underdetermined data and
176
J underdetermined expectations which seems
| to be characteristic of human recognition
I (What Computers Can*t Do 242).
f
t
Cognitive science tends to see the problems of
embodiment, such as locomotion and the physical
manipulation of objects, and of perception, such as
pattern recognition and representation, as problems of
calculation. Phenomemologists argue that when a blind
man "sees" the world, or at least crucial portions of
I
1
it, through the tapping of a cane, his intentionality
I
!extends to the tip of the cane. Patrick Heelan even
argues that a scientist can use an instrument in a
non-objective way, in other words that the instrument
can be displaced to the subject side of the cut and thus
used in a way analogous to a communications medium
(Heelan "Hermeneutics of Experimental Science" 32). If
we take up this notion, or the fact that in general our
tools tend to become extensions of our bodies, we can
begin to see the complexity involved in seeing
perception or manipulation as calculation. How can a
blind man perceive the world through an inanimate
object? How can a workman pick up tools of different
sizes, shapes and weights and direct them to the task
without thinking about it, or making any conscious
adjustment? Every physical action in the world
|requires astoundingly complex adjustments and reactions.
177
Dreyfus says
Merleau-Ponty admits that this ability
seems "magical" from the point of view of
science, so we should not be surprised to
find that rather than have no explanation
of what people are able to do, the
computer scientist embraces the assumption
that people are unconsciously running with
incredible speed through the enormous
calculation which would be involved in
programming a computer to perform a
similar task. However implausible, this
view gains persuasiveness from the absence
of an alternative account (Dreyfus What
Computers Can*t Do 253).
However, there are alternative possibilities. In
Mind Over Machine, a later book, Dreyfus considers the
possibility of holistic pattern recognition. He states
that the question is whether the mind/brain is strictly
an information processing mechanism. He claims that
there are physical systems that can detect similarity
without using any features or rules at all. Dreyfus
then discusses the properties of holograms and argues
that they exhibit mind-like properties, and that neurons
have been shown to behave in ways that fit the
holographic model,
! For us, the most important property of
holograms is the way they detect
similarity. For example, one can make a
hologram of this page and then make a
hologram of one of the letters on the
page, say the letter F. If one then
superimposes the light beamed through the
two resulting holograms, the astonishing
result is a black field in which a bright
spot appears at the location of each
letter F on the page. Moreover, the
178
brightness of the spot is proportional to
the similarity between the particular
letter F picked out and the letter F used
as the model. Dimmer spots appear where
there are imperfect or slightly rotated
versions of the reference letter. Thus
holograms can act as virtually
instantaneous similarity recognizers
(Dreyfus and Dreyfus Mind Over Machine
59-60).
Dreyfus (along with his brother Stuart, a professor of
industrial engineering) notes that this process makes no
use of features, as an information processing approach
would do. Dreyfus does not argue that this is how the
mind recognizes similarity. What he does want to say is
that the information-processing model may be entirely
and utterly wrong-headed.
What Computers Can't Do was based on a report that
Dreyfus had written as a consultant for the RAND
corporation. He says
When I had finished my report and
submitted it to RAND to be distributed, I
got my first taste of the unscientific
character of the field. Workers in
artificial intelligence, unlike most
scientists, almost never acknowledge their
difficulties and are highly sensitive to
criticism. RAND usually published the
conclusions of the consultants it hired,
but because I was criticizing RAND's
cognitive simulation research, Simon and
Newell insisted that my paper was nonsense
and that RAND should in no way appear to
condone it. That led to a year-long
j struggle within RAND as to whether the
I paper should be published or suppressed
(Mind Over Machine 8)♦
l
! When the book was finally published, it provoked much
179
hostility and rancor in the AI community. The second
edition, published in 1979, softened the harsh judgment
of the earlier edition, in that Dreyfus admitted that
certain trends in AI, for example, the adoption of
I organizing schemas or frames, make the approach more
|human-like (The Mind»s New Science 163). However, a
]number of AI researchers are still not speaking to
1
j Hubert Dreyfus.
I
j Jerome Bruner also has doubts about the present
direction of cognitive science. Of the cognitive
revolution, Bruner says in a recent book,
Very early on . . . emphasis began
shifting from "meaning” to "information,"
from the construction of meaning to the
processing of information. These are
profoundly different matters. The key
factor in the shift was the introduction
of computation as the ruling metaphor and
of computability as a necessary criterion
of a good theoretical model. Information
is indifferent with respect to meaning.
In computational terms, information
i comprises an already precoded message in
J the system. Meaning is preassigned to
j messages. It is not an outcome of
; computation nor is it relevant to
j computation save in the arbitrary sense of
assignment (Acts of Meaning 4).
Bruner laments this change. He says that in the 1950*s
he and his colleagues thought the cognitive revolution
was "an all-out effort to establish meaning as the
central concept of psychology— not stimuli and
responses, not overtly observable behavior, not
180
biological drives and their transformation, but meaning"
(Acts of Meaning 2).
Jerry A. Fodor is also acutely aware of the
problems at the heart of the conflict between philosophy
and cognitive science, and the problem of meaning versus
data, though his position is quite different from
Bruner's. As Gardner notes, Fodor is pessimistic about
the likelihood that we will ever be able to understand
how content is dealt with by our computational systems.
I
!Fodor says
1
j There is room for both a computational
psychology— viewed as a theory of formal
processes defined over mental
representations— and a naturalistic
psychology, viewed as a theory of the
(presumably causal) relations between
representations and the world which fix
the semantic interpretations of the
former. I think that, in principle, this
is the right way to look at things. In
practice, however, I think that it's
misleading. So far as I can see, it's
overwhelmingly likely that computational
psychology is the only one that we are
going to get (Fodor Representations
233-34) .
Fodor worries that cognitive scientists may be
restricted to describing "syntactic" operations, without
ever understanding how these operations refer to the
external world. Though Fodor notes that "to deny that
mental operations have access to semantic properties of
mental representations is not to deny that mental
representations have semantic properties," he also
181
argues that it may be impossible, with present methods
and assumptions, for the cognitivist to do semantics
(Fodor Representations 244). Semantic properties may
simply be inaccessible to scientific investigation. As
Gardner puts it, in Fodor's view of cognitive science,
J
"the machine of our minds does not know what it is
talking about, and does not care about a semantic
relation" (The Mind’s New Science 84).
Cognitive science's answer to the problem of
reference to a world has been the "micro-world." Terry
Winograd's SHRDLU program, which can manipulate and
discuss in limited English a micro-world of geometrical
|shapes, was considered to be a major breakthrough in AI.
j Fodor says of it,
I The interesting point for our purposes,
I however, is that the machine environment
j which is the nominal object of these
actions and conversations actually isn't
| there. What actually happens is that the
j programmer so arranges the memory states
I of the machine that the available data are
whatever they would be if there were
objects for the machine to perceive and
manipulanda for it to operate upon. In
effect, the machine lives in an entirely
notional world; all its beliefs are false.
Of course, it doesn't matter to the
machine that its beliefs are false since
falsity is a semantic property and, qua
computer, the device satisfies the
formality condition; viz., it has access
only to formal (nonsemantic) properties of
the representations that it manipulates.
In effect, the device is in precisely the
situation that Descartes dreads; it's a
mere computer which dreams that it's a
182
robot (Fodor Representations 2 32).
:Fodor says that people, such as Winograd, who do machine
I
simulation often represent their work as investigating
the question of how thought (or language) is related to
the world. He argues, however, that "whatever else
they're doing, they certainly aren't doing that" (Fodor
Representations 232). Fodor argues that the basic
assumptions that define cognitive psychology, that they
[study mental processes as formal operations on
symbols— "guarantees that their studies won't answer the
question how the symbols so manipulated are semantically
interpreted." To illustrate, he says
You can, for example, build a machine that
answers baseball questions in the sense
that (e.g.) if you type in "Who had the
most wins by a National League pitcher
since Dizzy Dean?" it will type out "Robin
Roberts, who won 28." But you delude
yourself if you think that a machine which
in this sense answers baseball questions,
is thereby answering questions about
baseball (or that the machine has somehow
referred to Robin Roberts). If the
programmer chooses to interpret the
machine inscription "Robin Roberts won 28"
as a statement about Robin Roberts (e.g.,
as the statement that he won 28), that's
all well and good, but it is no business
of the machine's. The machine has no
access to that interpretation, and its
computations are in no way affected by it.
The machine doesn't know what it is
talking about, and it doesn't care; about
is a semantic relation (Fodor
Representations 233).
One solution to this problem is the Derridean one.
Derrida argues that "from the moment that there is
meaning there are nothing but signs. We think only in
signs" (Of Grammatoloav 50). Derrida simply dissolves
the problem of semantics and reference with the concept
of play.
i
I One could call play the absence of the
| transcendental signified as limitlessness
i of play, that is to say the destruction of
j ontotheology and the metaphysics of
I presence. It is not surprising that the
| shock, shaping and undermining metaphysics
' since its origin, lets itself be named as
' such in the period when, refusing to bind
i linguistics to semantics (which all
i European linguists, from Saussure to
Hjemslev, still do), expelling the problem
of meaning outside of their researches,
certain American linguists constantly
refer to the model of a game. Here one
must think of writing as a game within
language (Of Grammatologv 50).
In other words, we are all in SHRDLU's situation. We
manipulate signs which merely refer to other signs
within the linguistic system. The language game is a
game of relations, but not of reference to a world, and
if there are representations, they are representations
of other signs. There is no way out.
Fodor believes that relatively rapid and automatic
mental operations (the syntactic parsing of a sentence
!or the detection of forms in the visual world) are the
1
j aspects of mental processing most likely to be
!elucidated by current cognitive scientific methods. He
is pessimistic about the possibility that those
184
capacities involving sustained judgment and reasoning
will prove susceptible to the kind of syntactic or
formal analysis typical of cognitive science (The Mind's
New Science 85). He is, in effect, expelling the
;problem of meaning outside his research.
1 Fodor ends the essay I have been quoting
i
|("Methodological Solipsism") by invoking a
I
!phenomenological concept from Heidegger.
i
| My point, then, is of course not that
solipsism is true; it's just that truth,
reference and the rest of the semantic
notions aren't psychological categories.
What they are is: they're modes of Dasein.
I don't know what Dasein is, but I'm sure
that there's lots of it around, and I’m
sure that you and I and Cincinnati have
all got it. What more do you want? (Fodor
Representations 253).
Thus Fodor admits that cognitive science must be
supplemented by a phenomenological hermeneutics.
Gardner's third point, the fact that cognitivists
tend to de-emphasize affective, emotional, historical,
cultural, and contextual factors, is an important
limitation of the cognitive view. This is the source of
the disagreement between Flower and Hayes on the one
hand, and Cooper and Holzman on the other, on the
question of "ecological validity." This is a conflict
!of perspectives, and Flower and Hayes are right to argue
i
that "the choice of a method isn't absolute. It depends
on our research objectives" ("Response to Cooper and
i
<
i 185
Holzman" 97). The cognitive approach brackets, out of
methodological necessity, matters that Cooper and
Holzman consider to be of central importance. Of
course, at some point, the brackets must be removed and
Jthe questions Cooper and Holzman raise be considered.
I This is the crux of the problem of the cognitive
i
approach for composition. Are theories and methods that
deal with syntax but not semantics, sense but not
reference, and object but not context, sufficient to our
purposes? Ricoeur's later hermeneutic articles and the
book Rule of Metaphor deal with exactly this problem—
the movement from system to reference. Ricoeur's
adversary in these arguments is structuralism, but the
problem is the same. Structuralism assumes that
language is an object that can be investigated
l
scientifically. In order to make such investigation
possible, structuralism distinguishes between a
synchronic science of states of the system (langue), and
the diachronic study of utterances (parole), and focuses
on lanaue as a more properly scientific object of study.
Lanaue is a closed and autonomous semiotic system, with
no absolute terms, but only relations of mutual
dependence (Thompson Critical Hermeneutics 48).
Ricoeur argues that in founding itself on these
assumptions, structuralism excludes the act of speaking,
186
history, process, and the primary purpose of language,
which is to say something about something, to refer. As
we have seen, cognitive approaches have the same set of
limitations. Ricoeur himself says of Levi-Strauss's
argument in The Savage Mind;
I will characterize the method in one
word: it is the choice of syntax over
semantics. To the extent to which it is a
wager made with coherence, this choice is
perfectly legitimate. Unfortunately, it
lacks reflection on its conditions of
j validity, on the price to be paid for this
| type of comprehension, in short, a
I reflection on limits, which did, however,
! appear from time to time in previous works
J Cl 40).
I
j The reflection on limits is an important step in the use
I
I or evaluation of any method, and in the interpretation
1
land application of the results. Ricoeur says
I
I The consciousness of the validity of a
method is never separable from the
consciousness of its limits. It is in
order to give full measure to this method,
and especially to allow myself to be
instructed by it, that I will seize hold
of it in its movement of expansion,
starting with an indisputable core, rather
than taking it at its final stage, past a
certain critical point where, perhaps, it
loses the sense of its limits (Cl 3 0).
It is important also to note, however, that the
recognition of limits does not necessarily decrease the
value or utility of a methodology. The acceptance of a
method, as I have pointed out before, can be seen as a
l
i
187
kind of bracketing. Ricoeur criticizes Levi-Strauss for
going beyond the legitimate limits of a structural
approach in his later work, not for using such an
!approach. Ricoeur says "Structuralism is part of
i
i
(science, and I do not at present see any more rigorous
i
j or more fruitful approach than the structuralist method
at the level of comprehension which is its own." He
goes on to describe the relationship between such a
method and a phenomenological hermeneutics in the
following way:
The interpretation of symbols is worthy of
being called a hermeneutics only insofar
as it is a part of self-understanding and
of the understanding of being; outside
this effort of appropriating meaning, it
is nothing. In this sense hermeneutics is
a philosophical discipline. To the extent
to which the aim of structuralism is to
put at a distance, to objectify, to
separate out from the personal equation of
the investigator the structure of an
institution, a myth, a rite, to the same
extent hermeneutics buries itself in what
could be called "the hermeneutic circle"
of understanding and of believing, which
disqualifies it as science and qualifies
it as meditating thought. There is thus
no reason to juxtapose two ways of
understanding; the question is rather to
link them together as the objective and
the existential (Cl 30).
Thus, perhaps Ricoeur would support the type of
collaboration between science and philosophy advocated
by Gardner. And in this way, as I hope to show in a
later chapter, we may link the discoveries of cognitive
188
and hermeneutic approaches in composition research.
Protocols and the Cognitive Model
When Flower and Hayes first began publishing their
;research in the early 1980's, it is likely that many of
i
• their readers were familiar with the technique of
i
jprotocol analysis, through the work of Janet Emig and
!others. It is also likely, however, that these articles
were, for many compositionists, the first exposure to
the assumptions and methods of cognitive psychology. In
such a situation it is sometimes difficult for a reader
to separate the discoveries of the particular research
project from the more general insights that are obtained
simply through the adoption of a particular methodology,
which is in itself a new way of seeing. In order to
understand the significance of the work done by Flower
and Hayes, it is necessary to clarify the assumptions
i
| which underlie the perspective, the controversies which
I
|surround those assumptions inside the field from which
]
: they have been borrowed, and the standards under which
l
such research is normally conducted,
j The most definitive work on protocol analysis as a
method in cognitive psychology has been done by K.
Anders Ericsson and Herbert A. Simon, both of Carnegie-
Mellon University, as are Linda Flower and John R.
189
jHayes. "Verbal Reports as Data," published in the May,
i
t
j1980 issue of Psychological Review, is a defense and
I
description of the method, with a detailed analysis of
the major issues, a review of the research, and a useful
bibliography. This article was expanded to book length
and published as Protocol Analysis; Verbal Reports as
Data in 1984. These works are essential to anyone who
wishes to use protocol analysis as a research tool.
Although Ericsson and Simon defend protocol analysis as
a method, they do not accept the validity of all
protocols under all research conditions, so
compositionists must read carefully.
One of the most common attacks on protocol analysis
is to label it "mere introspection." Cooper and Holzman
level this charge at Flower and Hayes, equating the
I
j protocol method with introspection as described by
Wilhelm Wundt ("Talking About Protocols" 284). Cooper
and Holzman note that Wundt's original method, which
made use of trained "introspectors," was said to be
limited to static mental states, because processes
introduced too much complexity, and that after the
j discovery of the unconscious, the Wundtians were
: criticized because the transparency of conscious mental
I
processes became questionable ("Talking About Protocols"
j 289) . From a phenomenological view, however, as I have
I argued in chapter 2, think-aloud protocols are not the
result of introspection.
Ericsson and Simon demonstrate that arguments like
Cooper and Holzman's are based on a failure to make a
distinction between concurrent and retrospective
verbalizations. They note that since the triumph of
behaviorism over introspectively oriented competing
viewpoints, such as Wundt's, verbal reports of any kind
have been suspect as data. They point out, however,
that even behavioral approaches accept verbal responses,
such as yes/no answers, as data. What is suspect from
an empirical point of view are reports of processes
leading up to some sort of verbal product.
Verbalizations about processes, whether concurrent or
retrospective, have been considered by behaviorists to
be merely variations of the discredited practice of
| introspection, to which Cooper and Holzman refer.
i
i . . . .
|Researchers influenced by behaviorist views of what
1 constitutes psychological research have had very little
i
I
iuse for introspection; although they will allow that it
<
I may be useful for the discovery of psychological
i
!processes, they consider it to be worthless for
i
i
■verification (Ericsson and Simon "Verbal Reports" 216).
- Cooper and Holzman, therefore, are throwing in with the
J
J
tbehaviorists in making this particular criticism of
191
Flower and Hayes. Their article raises a number of
important questions about the validity of this research,
but the main thrust of their argument is a rejection of
protocol analysis as a methodology, and in this argument
I think they are wrong,
j Ericsson and Simon wish especially to defend
!concurrent verbalizations against the behaviorist
j empirical prejudice. Of studies which have purported to
i
show that verbal reports are unreliable, they say,
Claims that verbalized information is
inconsistent with other behavior are often
made in general and sweeping terms without
providing specific evidence. When
evidence is provided, it is often
anecdotal, resting on the premise that if
one can produce a single case of a clearly
inaccurate or inconsistent verbal report,
i then verbal reports are wholly
inadmissible as data .... In most of
the cases in which inconsistency has been
observed or claimed, the verbal reports
were retrospective ("Verbal Reports" 243).
This is not to say that retrospective reports should be
rejected out of hand. Ericsson and Simon categorize and
analyze the value and characteristics of a number of
jdifferent types of verbal reports. They are empirical
i
I scientists and the assumptions under which they will
I
accept protocol analysis as scientific research are
quite clear. First,
We will conceive of the recorded
verbalizations as data— exactly like
latencies, eye fixations, sequences of
moves, and so on— to be accounted for by a
192
corresponding model, which generates them
literally or on the level of encoded
patterns or information content. This
means that we will not assume that the
verbalized description accurately reflects
the internal structure of processes or of
heeded information, or that it has any
privileged status as a direct observation;
models that can regenerate the
! verbalizations (or encoded aspects of
] them) can be constructed and evaluated
without such assumptions ("Verbal Reports"
217) .
This means that, as in all objective science, the
results are indirect correlations of data with
inaccessible phenomena, and that the model is one which
will explain or even generate the observed data, but is
not necessarily the actual process.
Cooper and Holzman accuse Flower and Hayes of just
such a mistake— confusing a model of a process with the
reality, or at least being fuzzy about the distinction.
They invoke Max Weber and his rules for model building,
and Chomsky on the difference between a transformational
! grammar, which postulates abstract rules for making
i
! grammatical sentences, and the actual process used by
]
the mind in making such sentences ("Talking About
j Protocols" 285). Stephen North makes a similar argument
| about the Flower and Hayes model and their presentation
i
| of it.
i
1 Dressed out in everyday language, the
! model looks very much as though it's
j making empirical statements about observed
; phenomena. To make matters worse, Flower
193
and Hayes' introductory phrasing tends to
foster the empirical illusion. They write
"We Propose that writing consists of three
major processes," rather than something
more strictly accurate, like "In our
■ model, writing consists of three major
processes" (The Making of Knowledge 249).
:There is nothing wrong with model-building as long as it
!remains clear that the result is a model.
Ericsson and Simon's argument for the acceptance of
verbal reports as data contains another premise,
however, that has profound implications for composition
studies. They argue that protocol discourse can be
treated as if it were equivalent to eye blinks and other
external manifestations of behavior, and indeed must be
so treated if the study is to have empirical validity.
This is related to Fodor1s argument that cognitive
science cannot deal with semantic content. Of course,
t
'there are procedures, such as type/token counts, which
|can be used to analyze discourse without dealing with
its meaning. Even the yes/no responses they mention,
however, depend on a rudimentary level of
interpretation. Is it possible to elucidate problem-
i
solving processes, or composing processes, without being
influenced by semantic intentions? And even if it is
i
possible, is it desirable? If we accept this limitation
I
in the name of scientific objectivity, we are not t
J
allowed to understand what the subject is trying to tell j
us. We can only explain the correlation between the
words which have been said and the processes which
allowed them to be produced.
Thus, as Bruner noted, the cognitive model of human
thought is that of "information processing," and the
assumptions behind this model are those of "information
theory." As Howard Gardner puts it, the key notion of
Jinformation theory is "that information can be thought
|of in a way entirely divorced from specific content or
j subject matter as simply a single decision between two
equally plausible alternatives" (The Mind’s New Science
21). This is how computers are programmed— in a series
of ones and zeros which mean "yes" or "no," "on" or
"off." From this perspective, information is simply
information, regardless of how or in what it is
embodied. Although Gardner notes that recently a few
cognitive scientists have begun to wonder whether they
can afford to ignore issues of content (22), Ericsson
and Simon feel that the information processing model is
acceptable to nearly everyone working in their field.
The most general and weakest hypothesis we
require is that human cognition is
information processing: that a cognitive
process can be seen as a sequence of
internal states successively transformed
by a series of information processes. An
important and more specific assumption is
that information is stored in several
memories having different capacities and
accessing characteristics: several sensory
I 195
stores of short duration, a short-term
memory (STM) with limited capacity and/or
intermediate duration, and a long-term
memory (LTM) with large capacity and
relatively permanent storage, but with
; slow fixation and access times compared
j with other memories ("Verbal Reports"
j 223).
! Given these assumptions, the important questions are
about information storage and flow— how it is retrieved,
organized, constructed, associated, and saved— not about
what it means. Thus, a cognitive model is similar to a
wiring schematic for an audio amplifier; there is a lot
J of information about where the signal goes and how it is
j processed, but nothing about what kind of music it will
play or how people will feel about it.
Ericsson and Simon assume that information recently
acquired or attended to is kept in STM and is directly
accessible for verbal reports, whereas information from
j LTM must be retrieved to STM before it can be reported.
i
I Thus, "with the instruction to verbalize, a direct trace
is obtained of the heeded information, and hence, an
indirect one of the internal stages of the cognitive
process" ("Verbal Reports" 22 0). The model employed by
Ericsson and Simon also includes a central processor
"which controls and regulates the nonautomatic cognitive
processes, determines what small part of the information
in sensory stimuli and LTM finds its way into STM"
("Verbal Reports" 224). This is the information that is
196
heeded or attended to. Their explanation of how this
processor works is quite similar to the actual
functioning of a computer. They note that the amount of
information that can reside in STM at one time is
limited to a small number, perhaps four, of familiar
patterns or chunks, and that each chunk is represented
by one symbol or pointer to information in LTM. As new
information is heeded, information previously stored in
STM is displaced.
It is clear that the researcher's access to
knowledge in short-term memory through the verbal report
is more direct than any possible access to the processes
which retrieve it. However, it is questionable whether
the verbal report is actually a "direct trace."
Ericsson and Simon make a careful analysis of the types
of verbalizations which are possible. In what they call
"direct" or Level 1 verbalization, the information is
reproduced in the form in which it was acquired from the
central processor. Level 2 verbalization involves
intermediate recoding into verbal code, because the
internal representation of the information was not
linguistic. This sort of verbalization is not direct.
Level 3 verbalization involves intermediate scanning or
filtering processes, because the task instructions ask
for only a selected type of information. Another type
197
of Level 3 verbalization involves intermediate inference
or generative processes, because the task instructions
ask for aspects of the situation which are not
ordinarily verbalized or attended to. An example of the
|latter would be an investigation of motives and reasons
i
i for action ("Verbal Reports" 219-20). This third type
interferes with normal processing by demanding further
information processing for the protocol itself, and
certainly cannot be considered to be a direct trace. As
we shall see, in most problem-solving experiments the
I
task is simple enough that it is possible to determine
I
fairly well what level of complexity the verbalization
will involve. Writing, however, is a task of such
complexity in itself that it is more difficult to
predict what sort of verbalization the protocols will
require, although the instructions given by such
researchers as Emig, and Flower and Hayes, have
generally avoided asking for information that would
require the kind of verbalization that Ericsson and
Simon label level 3.
Flower and Hayes are of course aware of the above
i
analysis (They summarize it in their response to the
Cooper and Holzman Article (p.95) and in "Uncovering
Cognitive Processes in Writing: an Introduction to
Protocol Analysis," p. 214-215.). However, they still
198
speak of their method as if it is a direct apprehension
of cognitive processes.
Writing research methods can be divided
quite naturally into two groups: The
input-output methods and the
process-tracing methods. The relationship
between the two methods can be understood
as follows. When we use input-output
methods to study writing, we act as if the
writing processes were occurring in a
locked room— a room which we are forbidden
to enter or look into. We put writers,
writing assignments, and reference books
(inputs) into the room, and we receive the
finished text (the output) at the door.
However, we never try to observe directly
the processes that occur in the room.
Rather, by varying the inputs and
observing their effects on the output, we
try to infer what the intervening writing
processes must have been.
In contrast, when we use
process-tracing methods, it is as if, in
addition to the information above, we have
a window that allows us to look into the
locked room and observe some of the
processes by which input leads to output
("Uncovering Cognitive Processes" 210-11).
I am afraid their "locked room" or "black box" metaphor
is a little mixed-up here. If the locked room is
consciousness, it is confusing to say that we put
writers in it. Even allowing for that minor confusion,
j however, there is evidence of a greater confusion. In
I
|
jshowing that input-output methods are indirect, and
j
I stating that when we use such methods "we never try to
observe directly the processes that occur," they seem to
■ be claiming that process-tracing methods are in fact
' direct. When they state that input-output methods must
199
infer the processes that produced the output, they seem
to be claiming that process-tracing methods do not
involve inference, when in fact all they can
!legitimately claim to have access to is a trace of the
i contents of short-term memory, from which they must
I
infer the processes they are attempting to study. They
are arguing that input-output methods involve black box
theories, and that process-tracing methods open a window
into the box. What they see through their window,
i
jhowever, is a number of new impenetrable black boxes.
There are a number of other limitations to protocol
analysis as a research methodology, each producing a
black box, or a 1 1 drop-out" in the information that is
available for a verbal report. In chapter two I argued
that automated or habitual actions would not appear in
protocol discourse. Ericsson and Simon are very clear
about this particular limit of the verbal report
i
i
methodology. Their model "assumes that only information
in focal attention can be verbalized." They go on to
say
In our model, as in most theories of the
structure of the human
information-processing system, a
distinction is made between fast automatic
processes that are not necessarily
conscious (and that are often thought to
I proceed in parallel) and the slow serial
! processes that are executed under
cognitive control .... With increase in
experience with a task, the same process
200
I may move from cognitively controlled to
j automatic status, so that what is
I available for verbalization to the novice
may be unavailable to the expert ("Verbal
Reports" 235).
Ericsson and Simon note that practice leads to
automation, which means that "intermediate steps are
carried out without being interpreted, and without their
i
iinputs and outputs using STM." They see the automation
Iof performance as analogous to "executing a computer
algorithm in compiled instead of interpretive mode"
("Verbal Reports" 225). The consequences of both
automation and compiling are that the process is speeded
iup, typically by an order of magnitude, and that the
jintermediate products are unattended to, unavailable to
j STM, and thus also unavailable for verbal reports.
Ericsson and Simon ignore the issue of whether automatic
|
actions are voluntary ones, which is an important issue
for Ricoeur. Ricoeur's phenomenological analysis of
habit is in general agreement with the operational
Jaspects of this cognitive model, though the larger
I
! questions of will and intention are left untouched by
the scientists.
Ericsson and Simon's observation that what is
;available for the novice to verbalize may not be
i
javailable to the expert would seem to contradict the
i
finding by Flower and Hayes that expert writers actually
201
verbalize more than basic writers in protocols
("Cognition of Discovery1 1 ). However, Michael Carter
argues in a recent article that neither group of writers
were expert in the task given to them in the Flower and
Hayes study, to write an article for Seventeen magazine
("The Idea of Expertise" 277). Thus the more
experienced writers had a greater stock of general
strategies, which they were able to verbalize. If the
study had included some regular contributors to
Seventeen, different results might have been obtained.
I Ericsson and Simon note other limitations of verbal
t
I
{
[reports. They distinguish two different retrieval
.processes, recognition and association.
1 Recognition associates the stimulus, or
i some part of it, with existing patterns in
LTM and stores in STM "pointers" to those
j familiar patterns .... The LTM may be
' pictured as an enormous collection of
inter-related nodes. Nodes can be
| accessed either by recognition, as just
! explained, or by way of links that
associate these nodes to others that have
already been activated. Information
accessed by association is then also
represented by pointers in STM. Thus,
! information can be brought into STM from
sensory stimuli via the recognition
! process, or from LTM via the association
j process. Association processes are much
■ slower than direct recognition processes,
requiring at least several hundred
j milliseconds for each associative step.
! Associative processes may use STM to store
! intermediate steps ("Verbal Reports" 224).
i
An example of associative processes given by Ericsson
202
and Simon is recalling a name that is not immediately
I
j accessible. In such a case, a person may use a sequence
i
j of cues to find an associative path, complete with dead-
■ ends and wrong turns, to the sought-for name. They note
I that such processes may last tens of seconds, or even
I
i minutes, and may leave numerous intermediate symbols in
i
I STM, where they are temporarily available for verbal
I
■ reports. Ericsson and Simon also cite studies which
show that while these associative processes tend to
leave a trace in STM, information that leads to direct
recognition of the appropriate action tends not to be
verbalized ("Verbal Reports" 238). This further reduces
the part of the actual retrieval process which will be
represented in the protocol discourse.
Ericsson and Simon note that Polanyi has argued
: that logical inference and explicit knowledge are not
the only types of knowledge available to people. For
i
i
• Polanyi, perception and recognition of objects are
! primary examples of the use of tacit knowledge. On the
J
basis of psychological evidence, Polanyi argues that
recognition is not a reportable operation because
i subjects are, in general, unable (1) to report what
i
I particular information they relied on to make the
!
I recognition, and (2) to state what the necessary
i perceptual attributes are for identifying instances of
i
■
l
i
| 203
that particular category. In response to these
iobservations, Ericsson and Simon say,
While the evidence Polanyi adduces would
create difficulty for a theory that
assumes all thought processes to be
; applications of the rules of logic, it is
: not in any way incompatible with the model
of thinking we are proposing here. We
1 have assumed that LTM is indexed by a net
j of tests that can be applied to stimuli in
I order to discriminate among and recognize
! them. Nothing in this assumption implies
that the person making a recognition has
! conscious access to these tests, or can
: state what they are— and, indeed, the
empirical evidence shows pretty clearly
that there is no such access (Protocol
Analysis 113).
: It is clear that any process to which there is no
: conscious access becomes a "black box" for this
Imethodology. Flower and Hayes have not clearly
irecognized, or at least articulated, these limits. The
icomparison of input-output methods to process-tracing
methods noted above is especially misleading in the
light of the following analysis by Ericsson and Simon,
who characterize the information processing model as
.describing cognitive processes producing a sequence of
istates, each marked by a small amount of heeded
I
information in STM.
t
r
J Formally, we can represent the time course
j of information in STM as a sequence, each
element of which is the output of the
process that operated on its predecessor
and the input to the process that will
produce its successor. We deliberately
omit further description of the process
204
I
j that creates these input-output relations
j because, from the viewpoint of attention
j and consciousness, this process is opaque
I and its intermediate states are not
j reportable (Protocol Analysis 118-119).
I
As I have argued above, protocol analysis carried out in
^the cognitivist fashion, while it does access
iinformation unavailable to other methods and could be
j
I
said to have penetrated at least the outer wrapping of
i
the larger "black box" of consciousness, produces a
number of smaller black boxes, with inputs and outputs
of their own.
Ericsson and Simon also discuss the effects of
interruptions in processes, both external and internal.
They note that
The flow of attention is diverted, from
time to time, by interruptions through the
higher control mechanism. Intermediate
stages in these interruptions, not being
symbolized in STM, are not reportable.
Sudden movements in peripheral vision,
loud noises, and emotions operating
through the reticular system are important
causes of interruption and shift in
attention .... Even though information
heeded immediately before or after a shift
in attention may sometimes allow subjects
to give a relatively clear account of the
interruption, we would expect such
information to be less complete than
reports of an orderly process that is
induced by the successive contents of STM
itself (e.g., a thought sequence during
which goals in STM are guiding the thought
processes) ("Verbal Reports" 225).
Interruptions therefore produce a "drop out" in
information related to the cognitive processes.
205
| Thus verbal reports, from this cognitive
j
j perspective, will contain information that is heeded or
I
i attended to about processes which have not been
i
!
I extensively practiced or automated. Interruptions of
i
; various kinds will further reduce the content, or the
! accuracy, of the protocol discourse. From the imperfect
j trace of short-term memory which the protocol records, a
!
,model is constructed which should account for both the
.protocol and the output of the process. The question at
;this point is how to analyze the information in the
1 verbal report, or how to move from protocol to model.
i
' Bridge Principles and Models
j As I noted above, Stephen North has criticized
j Flower and Hayes for presenting their model as if it
I
I makes "empirical statements about observed phenomena."
I
| North calls them "Formalists," taking the term from Paul
jDiesing's Patterns of Discovery in the Social Sciences.
|He notes, however, that they don't quite fit Diesing's
i
j description, because although they attempt to develop an
I
| "axiomatic, deductive structure," they don't develop a
formal language.
The object in all these kinds of Formalist
inquiry is model building. An
I Experimentalist sets out to sort through
I competing theories by attempting to
! isolate the features of some portion of
! the empirical world and working to
correlate the relationships among them. A
206
! Formalist, by contrast, investigates by
what amounts to analogy. The basic idea
is to construct, usually in a formal
language, a model whose internal logic in
some specifiable ways resembles that of
the phenomenon under study. A fully
realized Formal theory, argues Diesing,
will have both an explicit, clearly
articulated model or calculus; and one or
more sets of what are variously called
bridge principles or rules of
interpretation outlining its
correspondence to one or more empirical
domains (North The Making of Knowledge
238-39).
North argues that in the matter of interpretive or
bridge principles the Flower and Hayes model is
; deficient. He attributes this to the fact that their
i
! model is only implicitly Formal. He says, "because the
i
i
jline between model and description in implicit Formal
I
i
jtheories is often fuzzy, it is not always clear that the
|investigators recognize the need for rules of
t
■interpretation at all, let alone the need to make them
i
explicit” (The Making of Knowledge 246-47).
North believes that part of the problem is that the
model is expressed in ordinary language, and that when
we see a label like "Writer's Long Term Memory" we cease
to wonder about what characteristics that "black box"
might possess. North rewrites the model with labels
such as "Black Box #1" and "Process P," in order to show
what a true formalism might be like, in other words so
that an abstract account of how a model functions will
207
not be read as an empirical description. Translating
the model into less meaningful terms does not seem to me
to help or even change the model, but it does highlight
i
,a problem we have encountered a number of times in this
i
! analysis— the problem of the rhetorical purpose of the
Iarticles by Flower and Hayes. They are writing
(primarily for the consumer of composition theories, the
i
practicing writing instructor, who would not be well-
jserved by North's rendering of the model into
i
meaningless symbols. On the other hand, North's point
is well taken. The relation of the model to the
i
I empirical domain is obscured by their presentation of
it.
] North makes a strong distinction between
(Experimentalist knowledge which is, according to the
; rules of empirical science, knowledge directly about the
|
! world, and Formalist knowledge, which derives from a
t
imodel which will be in and of itself, perfect, a
|tautological whole, but which has imperfect or
incomplete correspondence to any empirical domain. Of
course, from the Husserlian point of view laid out in
the first chapter, the scientific method employed by the
experimentalist is indirect as well. This problem is
■not relevant to North's argument, however, which is that
jFlower and Hayes are inconsistent in the assumptions of
'their own method.
|
j We might summarize the trade-off this way:
! While Experimental inquiry holds out the
I potential for a full and paradigmatically
I stable explication of the order of things,
| any progress that is made is made very
j slowly. Formalist inquiry gives up that
I paradigmatic fullness and stability in
i exchange for the broader-ranging, more
i quickly accessible explanatory power of
I its models, but the knowledge they can
I generate is far less stable. And in fact,
i it is worth pointing out from the
| beginning that Formal models may well be
valuable less for what they can tell us
! about any system or process than for the
! heuristic power of the discrepancies
! between a model and any empirical domain.
Especially in the study of human
| activities, they highlight for us what we
, do not understand (The Making of Knowledge
; 242).
j North thus sees experimental methods and formal methods
(
:as complementary. However, he accuses Flower and Hayes
j
i of confusing their model-building activity with an
|empirical experimentalist procedure and thereby reducing
'the value, utility, and coherence of their own model.
t
| As we have seen, Flower and Hayes are not very
i
|helpful in illuminating the principles through which
i
i
jtheir model is related to their data. Ericsson and
!Simon are much more explicit about how this is to be
;done. As I noted before, they consider verbal reports
i
!to be valid data— just as valid as such observable and
quantifiable phenomena as eyeblinks and yes/no answers.
They note that it is desirable in science to maintain as
209
clear a separation as possible between data and theory,
and that data are supposed to derive directly from
observation. Theories, they say, are supposed to account
for, explain, and predict these observation-based data.
It is not as simple as that, however, for they make a
i
i
I further distinction between "hard" and "soft" data.
Data are "hard" when there is
intersubjective agreement that they
correspond to the facts of the observed
behavior .... When, however, an analyst
codes a five-second description of a dream
as "oral fixation," many psychologists
would argue that this encoding is not a
datum but a subjective interpretation Of
the data (i.e., of the verbal description
of the dream). Surely, theory-laden
inferences were required to derive the
encoding from the verbal protocol. Data
are regarded as "soft" to the degree that
they incorporate such inferences,
especially when the theoretical premises
and rules of inference are themselves not
completely explicit and objective
(Ericsson and Simon Protocol Analysis 4).
Husserl would certainly agree with most of this. There
are presuppositions involved in such coding which must
be de-sedimented and made explicit. If even "hard" data
requires "intersubjective agreement," however, it is
clear that interpretation of some kind is necessary to
deal with any kind of data. The ideal of a thoroughly
objective, thoroughly scientific manipulation of data is
simply unobtainable.
The goal, however, at least for cognitive
scientists like Ericsson and Simon, is to approach this
210
ideal as closely as possible. With this in mind, they
allow that "in less formal kinds of analysis, the
encoding scheme is not defined formally and a priori,
but the search for interpretations proceeds in parallel
with the search for an appropriate model or theory."
They recognize the value of such an interactive process
in the search for theories in new domains, but limit
itheir own concerns primarily to studies in which "the
theoretical terras are fixed before the actual encoding
begins" (Protocol Analysis 6). Clearly, Janet Emig's
work falls into the category of "less formal" from this
perspective.
Ericsson and Simon note that subjects are inclined
to believe that they have direct access to their own
mental states and processes. Subjects generally have a
strong conviction that they are able to report their own
mental experiences truthfully. Experimental
psychologists, however, have shown that under numerous
circumstances such self-reports are unreliable.
Ericsson and Simon argue, however, that the issue of the
reliability of self-reports should be avoided entirely.
They say
The report "X" need not be used to infer
that X is true, but only that the subject
was able to say "X"— (i.e, had the
information that enabled him to say "X.")
By following this path, we can even show
that there is an inverse relation between
211
how much subjects need to be trusted and
how much information they verbalize. For
the more information conveyed in their
responses, the more difficult it becomes
to construct a model that will produce
precisely those responses
adventitiously— hence the more confidence
we can place in a model that does predict
them (Protocol Analysis 7) .
If this is the case, it is no wonder that
compositionists have found it difficult to construct a
very specific and testable model of the writing
(processes, because think-aloud composing protocols tend
to produce such large amounts of discourse that the
construction of a general model simply from inferences
may well be impossible. The assertion that the
important "fact" is that the subject is able to say "X,"
not that "X" is true, is easily applicable to simple
problem solving situations (if the subject relates a
solution to the problem it shows that his thought
processes- have generated one), but much harder to apply
to a composing situation.
Ericsson and Simon argue that verbal behavior is
just one type of recordable behavior, to be observed and
analyzed like any other behavior. The goal of their
research is "the same kind of 'mechanical' and complete
process description of verbal behavior as of other kinds
of behavior, and we would not accept magical or
privileged processes as explanations for verbalizations"
212
(Protocol Analysis 9). A "mechanical and complete"
description of the writing process is one that could be
programmed to function on a computer, though we must
remember that semantic operations have been excluded.
Whether this A.I. oriented research has produced or will
produce actual machine intelligence is debatable. What
it has undeniably produced is better and more useful
computer programs. The following description of the
process of retrieval from long-term memory perhaps
illustrates how thoroughly computer-based the cognitive
model of the mind is:
The appropriate storage of relevant
information with reasonable cues is not
mysterious within our proposed model,
which views LTM as a richly indexed and
cross-indexed data base. When a subject
is presented with new information, he will
access (by recognition) the relevant LTM
structures and store some subset of the
heeded information. At the same time, new
recognition access routes to the
information will be created (new "index
entries"). This continuing processing and
storage of heeded information will allow
direct access by recognition to many
important pieces of information (Ericsson
and Simon Protocol Analysis 132).
This is just how a real data base program works. We
must remember, however, that the basic goal of this line
of research is machine intelligence, software that can
solve problems, converse, and perhaps write memos or
even poetry. This goal is fundamentally different from
the humanistic concerns of writing teachers, who must
213
deal with acts, intentions, meanings, and living
imperfect human subjects.
S e g m e n t a t i o n , coding and Analysis
Ericsson and Simon carefully describe the accepted
procedures for the analysis of protocol data. They
produce the following seven point guideline which I
quote at length.
1. In choosing data for transcription,
and in transcribing and encoding them, the
theoretical commitments that are made
should be kept as small and weak as
possible.
2. The theoretical assumptions and
presuppositions underlying the data
analysis should be made explicit. This is
easy when the analysis is done
automatically; difficult when it is done
by humans.
3. The data analysis process should be
not only objective and reproducible, but
it should remain constant from one
protocol to another. In this way, the
degrees of freedom that are used by these
procedures are absorbed by a constantly
growing body of data. This can be done by
dividing the protocol material into two
parts, developing the data analysis scheme
on one half, and testing it against the
other half.
4. Convergent procedures should be used
wherever possible. For example, as much
as possible of the encoding language and
concepts should be derived from an
analysis of the task environment,
independent of the problem spaces defined
by the subjects.
5. Constant and invariant data analysis
procedures should be used over large
bodies of data. It is much easier to
build a consistent ad hoc theory to
explain a protocol of 11 segments than one
of 200 segments.
6. Even where different theories are
214
needed to explain behavior in different
tasks, the theories should share as many
common elements as possible, so that a
(weaker) general theory can be abstracted
from the specific ones.
7. Even where different theories are
needed to explain the behavior of
different subjects in a given task, the
theories should also share as many common
elements and components as possible, so
that a theory of the "representative
subject" . . . can be constructed
(Protocol Analysis 286-87}.
These seven points will provide a framework for an
analysis of the procedures followed by Flower and Hayes
in the research described in "Identifying the
Organization of Writing Processes," published in Gregg
and Steinberg's Cognitive Processes in Writing. 1980.
Of the published work of Flower and Hayes, this article
contains the most explicit description of their methods
for data manipulation and interpretation.
Flower and Hayes begin this article with a
discussion of what protocol analysis is, and with a
presentation of their "tentative" model, which they say
has developed through the analysis of "a number of
protocols over the last 2 years" ("Identifying the
Organization" 10). We are not given the method by which
the model was derived from the protocols, but they do
say that "The psychologist's task in analyzing a
protocol is to take the incomplete record that the
protocol provides together with his knowledge of the
215
nature of the task and of human capabilities and to
infer from these a model of the underlying psychological
jprocesses by which the subject performs the task"
("Identifying the Organization" 9). From the previous
analysis presented in this chapter we can see that this
is indeed a very accurate statement. The protocol, for
all the reasons outlined above, provides an incomplete
record, not of processes, but of the contents of short
term memory. This record is the accessible residue of
actual inaccessible cognitive processes, which must be
inferred from it. A model of the functioning of these
processes must then be made. According to Flower and
Hayes, however, these inferences are to be made from the
protocol plus the researcher's knowledge of the nature
of the task, and the researcher's knowledge of human
capabilities. These are the "bridge principles" North
discusses, the connection between data and model. At
best, this is not strictly speaking "knowledge," but
assumptions about human purposes and characteristics.
By the standards of empirical science, these assumptions
are quite broad and vague. This definition of their
task allows for a considerable amount of inference and
subjective interpretation. The fact that these
assumptions are necessary demonstrates that a purely
"scientific" method is not possible in this type of
216
study.
The purpose of the procedures described in this
article is to test the Flower and Hayes model, which, as
we have noted before, consists of three major processes:
planning, translating, and reviewing. Planning is
further broken down into generating, organizing, and
goal-setting subprocesses. The functioning of these
processes is consistent with the cognitive processes
described by Ericsson and Simon. Generating, for
example, is associated with retrieval of information
relevant to the writing task from long term memory.
Flower and Hayes describe the information retrieval
process in typical cognitive fashion.
We assume that this process derives its
first memory probe from the information
about the topic and the audience presented
in the task environment. Because each
retrieved item is used as the new memory
probe, items are retrieved from memory in
associative chains. In order to focus the
search on relevant material, the retrieval
chain is broken whenever an item is
retrieved that is not useful to the
writing task. Search is then restarted
with a new memory probe derived from the
task environment or from useful material
already retrieved ("Identifying the
Organization" 12-13).
Flower and Hayes assert that the most persistent memory
searches they have observed in protocols never extended
more than three retrievals beyond useful material, and
that most stop after one. We are not shown evidence of
L
217
this, however, or told exactly how they can determine
this from the protocols. They say retrieval of an item
may generate a written note; in fact "the form of these
notes will be used later to identify occurrences of the
generating process" ("Identifying the Organization" 14).
The procedure in a nutshell is to evaluate certain
formal characteristics of the inscribed discourse to
identify the processes at work in the associated
protocol discourse.
As noted above, Ericsson and Simon recommend that
the protocol material be divide into two parts, so that
the data analysis scheme can be developed on one half
and verified on the other. However, they have in mind
large bodies of data. Typical studies cited by Ericsson
and Simon are Montgomery (1977) who coded over 1,000
statements into 9 different categories using two raters
who agreed 95% of the time, and Bettman and Park (1979)
who collected protocols about choosing between 15
different microwave ovens from 68 subjects, which were
encoded into 9 categories by two raters with nearly 80%
agreement (Protocol Analysis 299). On the other hand,
Flower and Hayes test their model against one protocol
"in which the writer gave especially clear indications
of ongoing writing processes and of the transitions
between processes." They claim that "this relatively
218
unambiguous protocol provides a rigorous test of the
model's adequacy" ("Identifying the Organization" 20).
The protocol they chose consisted of 14 pages of
verbal transcript, five pages of notes, and a page of
completed essay. They divided the transcript into
segments and labeled them as "metacomments." which they
describe as comments that writers make about the writing
process itself, "task-oriented or 'content' statements."
I which are statements that reflect the application of
i
I writing processes to the current task, and
"interiections." which were not studied ("Identifying
the Organization" 20). They analyzed about one half of
the transcript. Flower and Hayes's coding of one half
of one protocol does not seem like a very rigorous test
compared to the studies cited by Ericsson and Simon. It
is true that composing process protocols are long,
complex, and extremely time-consuming to transcribe,
much more so than those produced in most other problem
solving research, but the question remains— is this
science or not? If this research is to be accepted as
scientific it must do more than mimic accepted
procedures.
Ericsson and Simon argue that protocols must be
t
idivided up so that each segment will constitute one
instance of a general process, and that proper
219
segmentation is particularly important if the subsequent
analysis makes use of frequencies of processes. They
note that the usual procedure is to determine segments
prior to encoding them, although some investigators
segment and encode simultaneously. They say,
Under the assumptions of our information
processing model, the appropriate cues for
segmentation are pauses, intonation,
contours, etc., as well as syntactical
markers for complete phrases and
sentences— the cues for segmentation in
ordinary discourse. These are also the
j cues most often used in protocol studies.
A slightly different segmentation rule,
based on separating ideas . . . places
more emphasis on the actual content of the
verbalizations (Protocol Analysis 205).
Note that interpretation of the content of an utterance
is not absolutely banished here, though the
methodological problems this would entail are not
discussed. The segmentation scheme used by Flower and
Hayes, metacomments about the writing process and task-
oriented comments, clearly involves understanding the
discourse, and goes beyond such observable physical
criteria as pauses and pitch contours.
Flower and Hayes are concerned about this
characteristic of their procedure. They worry about the
objectivity of their interpretations because protocols
are often incomplete or ambiguous and "we frequently
have to make judgments about the writer's meaning." To
220
counteract this they have taken three steps:
1. Whenever objective evidence was
available, we used it. Thus, reading and
writing processes were identified by
matching the verbal protocol word for word
with the writer's notes and text (the
objective evidence).
2. Whenever possible, processes were
identified by using converging lines of
evidence, e.g., the form of the written
material on the one hand, and the writer's
comments about what he is doing on the
other.
3. The most important analyses were
i replicated by independent judges
| ("Identifying the Organization" 21).
1 It is true that Ericsson and Simon recommend
"convergent" procedures, but in this case it is unclear
exactly how the words of the notes and text are more
"objective" than the words of the protocol. All of this
is simply discourse produced by the writer for different
purposes in different modes. Point number three refers
to coding of segments by raters, and is standard
procedure in this type of research. However, the
segmentation scheme and its implementation, and the
coding instructions, all have a significant effect on
the performance of the raters.
| The protocol was divided into three sections on the
J basis of metacomments that explicitly stated that the
i
writer was moving from one mode to another. Thus, the
first section was supposed to be mainly generating. the
second mainly organizing, and the third mainly
i
221
translating. This division seems to contradict the
argument Flower and Hayes have made against simplistic
stage models of the writing process, although they do
allow sub-processes to be embedded in the dominant
process. The first hypothesis they wished to test was
as follows:
The form of the written material should
vary from section to section. Thus, in
the first section, we expect the
generating process to produce many single
words, detached phrases and incomplete
sentences. In the second section, we
expect the organizing process to produce
material that is systematically indented,
alphabetized, or numbered. In the third
section, we expect the translation process
to produce many complete sentences and
some material associated in the verbal
protocol with interrogatives suggesting
search for sentence continuation
("Identifying the Organization" 22).
This hypothesis leads to a coding scheme which asks
three raters to categorize "items" in the inscribed
material, an "item" being defined as "a word, phrase, or
sentence that was identifiable in the verbal protocol as
being written during a single segment or several
contiguous segments" ("Identifying the Organization"
22). This coding was done in terms of the following
questions:
1. Does it have good form, i.e., is it a
complete, grammatical sentence?
2. Is it part of a systematically
indented, alphabetized, or numbered
structure, i.e., does it appear to be part
of an outline or structured plan of some
222
sort?
3. Is it associated in the verbal protocol
with an interrogative suggesting search
for sentence completion? ("Identifying the
Organization" 23)
They report that inter-rater agreement was over 90% for
each question. This is a very odd procedure in a number
of ways. First of all, they are using the written
product to verify inferences about the protocol. rather
than attempting to infer the nature of inaccessible
cognitive processes from the protocol record. It is
this problem that causes Stephen North to remark,
Instead of designing a model that treats
texts as its primary output data, and
whose "black box" operations can be
somewhat illuminated by protocol analysis,
Flower and Hayes have given the protocols
primary status. The model they
posit— inferred, as they put it, from the
protocols and the knowledge, as
investigators, "of the nature of the task
and of human capabilities" (p.9)— is
designed so as to offer a far better
account of how protocols, rather than
texts, are produced. Indeed, by making
the basic textual unit the
segment-dependent "item," the model
effectively denies text any independent
status as output data at all. For
interpretive purposes, if a textual
feature does not correspond to a protocol
segment, it does not exist (The Making of
Knowledge 265).
I
! Second, the hypothesis they are testing, and the coding
■ scheme associated with it, simply reflect the
|
| conventions of form in this culture. It is like saying
i
that when someone says he is making an outline (in a
I
223
"metacomment") we expect to see Roman numerals and
capital letters. Then we check to see if everyone
j agrees that there are Roman numerals and capital letters
|used appropriately in the text. If most observers agree
|that those things are there then we conclude that the
j subject is indeed making an outline. Their reasoning is
circular, which is unfortunate, because up to this point
their method has interesting possibilities.
On the problem of encoding and encoders, Ericsson
and Simon say that because human coders tend to remember
information from previous segments, and because this
influences their coding, the researcher should make the
segments large enough that "all information for making
an encoding decision is contained in a single segment"
(Protocol Analysis 290). They also recommend that
segments be encoded in random order. Flower and Hayes
simply gave the raters "the written material and verbal
protocol" (23). In their procedure an "item" consisted
of a protocol segment tied to a part of the inscribed
discourse, either the notes or the final text, which
seems like a legitimate adaptation of the methodology,
considering the nature of writing as a task. However,
segments in context are not, strictly speaking,
segments. In one part of the study the researchers do
attempt to determine if consistent judgments of process
224
could be made without context:
We selected 41 content statements from the
protocol and typed them on cards. The
cards were then shuffled and presented for
judgment independently to two coders (not
the authors). Coder 1 agreed with one of
us in 67% of judgments and Coder 2, in 77%
of judgments (25).
In this part of the study, the coders are simply
labeling items with terms from the model, i.e,
generating or organizing. It is also clear from this
statement that in parts of this study, the authors have
done their own coding. It does not seem like a good
procedure to have the designers of the model being
tested do the segmentation and the coding of the
protocols that are to test it.
Flower and Hayes carefully qualify the scope of
their results. They say,
First, although the model was derived
through informal analysis of many
protocols, it has been tested formally
with only one. Second, although the model
is quite complex, only a few of its
properties have been tested ("Identifying
the
Organization" 27).
In this analysis we have discovered a number of problems
with this test of their model. First, the test works
from written product to protocol, rather than from
protocol to cognitive process. An associated difficulty
is the assumption that formal characteristics of the
written product correspond to different cognitive
225
processes. Second, as they note, the test data includes
only part of one protocol. Third, both the segmentation
and coding practices do not follow customary procedures
normally recommended in this sort of research. I think
it is clear that the testing of the model is entirely
inadequate. This is not due, however, to inadequacies
inherent in the cognitive method, but rather in the way
in which this particular research project was carried
f
I out.
)
I
i On the model itself I would say that it diagrams a
sequence of fairly high level operations with allowances
for recycling or recursiveness, leaving the more complex
processes underneath uninvestigated. This situation is
due in part to the limitations of the cognitive method.
I will explore holes, gaps, and black boxes of the
Flower and Hayes model in detail in the next chapter.
226
. CHAPTER FOUR
! A PRACTICAL CRITIQUE OF
THE FLOWER AND HAYES MODEL
In the previous chapter I presented both the
necessary limits of the cognitive approach and a number
of arguments supporting the conclusion that a convincing
case for the scientific validity of the Flower and Hayes
model of the composing process has not been made. In
the present chapter I wish to set aside the question of
scientific validity and explore the model itself from a
number of perspectives. I will take Flower and Hayes at
their word and apply, as they say, "knowledge of the
i
nature of the task and of human capabilities .. . to
infer from these a model of the underlying psychological
processes by which the subject performs the task
("Identifying the Organization" 9).
Flower and Hayes ask, quite legitimately, what is
new about their model, when English teachers have been
talking about planning, organizing, and editing for a
long time. They claim that,
Nonetheless, there is a great deal that is
new about the model. First, the model is
rather specific about the nature of the
individual processes .... Second, and
more important, the model specifies the
organization of these processes. In
particular, it specifies an organization
that is goal directed and recursive, that
allows for process interrupts, and that
can account for individual differences
("Identifying the Organization" 29).
In order to analyze these claims, it is first necessary
to describe the model in detail. The following diagram,
which has been reproduced in slightly different forms in
a number of articles, is here taken from "Identifying
the Organization of Writing Processes," in Gregg and
Steinberg, 1980 (p. 11). In reproducing it, I have
altered the format a bit to make it fit, but the
substance of the original rendering is unchanged.
228
TASK ENVIRONMENT
WRITING ASSIGNMENT
Topic
Audience
Motivating Cues
TEXT
PRODUCED
SO FAR
-V
WRITER* S
LONG-TERM
MEMORY
Knowledge
of Topic
Knowledge
of
Audience
Stored
Writing
Plans
->ORGANI ZING
w PLANNING
G
E
N
E
R
>A
T
I
N
G
— >
TRANSLATING
— w —
REVIEWING
READING
\- L /--
EDITING
MONITOR
Figure 3: The Structure of the Flower and Hayes Model
As can be seen from the diagram, the model consists of
three large boxes, one containing long-term memory, a
second containing the three main processes that Flower
and Hayes consider to be the essential components of the
writing process itself, and a third which represents the
"task environment."
According to Kim Sterelny in The Representational
229
Theory of Mind, a number of different cognitive
theorists have argued that there are three theoretical
domains in psychology, and that each individual
psychological mechanism and process exists and can be
described in each of these domains.
These domains are the subject matter of
three levels of psychological theory. The
top level is the level at which we specify
the function a psychological device
performs. It is the level at which we say
what a system does. One down is the level
at which we specify the information
processing means by which the system
performs its function or functions. At
the base level we specify the physical
realization of that information processing
device (Sterelny Representational Theory
43) .
Clearly, the Flower and Hayes model is a description at
the functional level. On occasion, they will speculate
briefly and vaguely about computational processes, but
in the main their focus is on what the mind is doing
rather than how it is doing it. They do not concern
themselves with the physical realization of the
processes they study, a concern which involves the study
of neurology and the structures of the brain. Of
course, those who begin their study of the mind with the
brain are often inclined to doubt the validity of
approaches which begin with functions.
Thus, the Flower and Hayes model is a system of
categories which describe the functions of writing
230
processes in terms of what they do. In the light of
Sterelny's three-level scheme, one would expect such
descriptions to be fairly general, although Flower and
Hayes claim that their descriptions are "rather
specific." In addition to the claims by Flower and
Hayes for the specificity of their description of the
nature and organization of writing processes, and the
claim that the model can account for individual
differences, I will also investigate the utility of the
model for a writing teacher.
The Interface with the Empirical World
The concept of the "task environment" is one of the
most interesting features of this model. In discussing
problem-solving research, Ericsson and Simon oppose the
task environment, which is seen as objective, to the
"problem space," which is a representation defined by
the subject. According to Flower and Hayes,
This [the task environment] includes
everything outside the writer's skin,
including the assignment or exigency for
writing, the audience, and eventually the
written text itself. It is important to
note that no matter what rhetorical
problem is actually presented to the
writer, what finally matters will be the
problem the writers represent to
themselves in the act of composing
("Uncovering Cognitive Processes" 208).
It is interesting to note that Flower and Hayes
231
emphasize that the internal representation of the
rhetorical task, a concept akin to Ericsson and Simon's
subjective "problem space," is more important than the
task as given. The problem of where the model creates,
stores and processes this representation will be
discussed in detail below.
The boundary between the external and the internal
is an important philosophical problem. Descartes'
original reduction made the physical body part of the
objective world, creating the problem of mind/body
duality. Ricoeur says,
In this manner is born, legitimately, a
scientific biology and psychology . . . to
the extent that my bodily and involuntary
life is for me confused, enigmatic, and
sometimes even foreign, there is every
invitation to relegate it thus to the
realm of things. On the other hand,
consciousness, recoiling from its objects,
tends to identify its own life, recognized
in all its concrete thickness, with
self-consciousness and to exile itself
thus in its capacity for reflection. One
must say, however, that this tendency,
too, is inscribed in the very structure of
consciousness; . . . the philosophy of
Descartes shows well enough how these two
movements are interdependent, namely, the
consolidation of an objective world not
tied to any consciousness, and its polar
opposite, the reduction of consciousness
to self-consciousness (The Philosophy of
Paul Ricoeur 4).
Ricoeur, of course, thinks that Descartes put everyone
on the wrong track when he made the subject/object cut
in this way and defined thought in terms of
232
self-consciousness. As we have seen, the analysis
presented in Freedom and Nature emphasizes the
intentional bonds between consciousness and the world,
and not the isolation of a consciousness which retires
into itself (Ricoeur FN 42-43).
It is unclear whether Flower and Hayes are simply
indulging in a figurative rhetorical flourish in marking
this boundary at the writer's skin instead of between
mind and body, or if they actually want to disagree with
Descartes, and side with Husserl and Ricoeur. As
Ricoeur argues above, in order to have an objective
science of psychology, mind and body must be split.
Cognitive psychology is a science. but it is a science
of the mind, and as such it, among all sciences, cannot
afford to be vague about the subject/object cut. This
is why Ericsson and Simon argue that even language must
be treated as data, or, in other words, that it must
remain uninterpreted on the object side of the cut.
Ricoeur, however, as a non-scientist, can afford to
be a bit more ambiguous. Don Ihde notes that while
Ricoeur speaks of two universes of discourse, the
objective and the subjective.
The non-coincidence of the two universes
of discourse is not total. Between them
lies what may be termed an ambivalence
border which allows the movement from
existence to objectivity to occur (and
which allows Ricoeur to speak implicitly
233
of a "subjectivization" of empirical
notions). Both refer to the body
(Hermeneutic Phenomenology 33).
The personal, lived body is the mediator between
consciousness and the empirical world. It is on the
ambivalence border. The body can be seen as object, but
it is experienced in a way different from any other
object. Ricoeur observes that the body is the center of
orientation, the zero origin, the "here from where"
everything seen is seen. Ricoeur argues that we must
consider an intermediate stage between the otherness of
silhouettes and the here of the body.
This stage is that of the free mobility of
my body: by changing my position I can
make the object's aspect change; a certain
behavior of my body commands the passivity
of the perceived being .... Here we
have the ultimate reference: the otherness
that my free mobility brings into play is
an otherness in relation to an initial
position that is always the absolute
"here" (Ricoeur PM 22) .
The body is thus the ultimate source of perspective, or
point of view.
Of course, there is more to perspective than
location. Ricoeur refines his concept of perspective
with the concept of character. He says "After the
detour through the idea of perspective, character, which
elsewhere we called the finite manner of freedom, now
appears to us as the perspectival orientation of our
field of motivation taken as a whole." (Ricoeur FM 60).
234
Ricoeur argues that character is not the opposite of
humanity, but is humanity seen from a perspective. In
other words, it is not the case that an individual has
human characteristics which she shares with all other
humans, plus other characteristics which make her an
individual.
That is why my character is never
perceived in itself, no more than the
origin of perception is itself the object
of perception. I do not aim at "my1 '
personal idea of happiness and honor but
happiness and honor per se: "my" character
is implied in the humanity of my
individual existence as the zero origin of
my field of motivation; nor do I meet it
as an exterior boundary, for it is not a
fate that governs my life from the outside
but the inimitable way in which I exercise
my freedom as a man (Ricoeur FM 61).
Thus Ricoeur argues that the zero origin of perception,
is not yet a true origin, for although one can change
one's position, Ricoeur believes one's character cannot
be changed (FM 62). As is typical in phenomenology, the
analysis moves from the object to consciousness, from
the perceived object to perspective, from the act to
character, and not the reverse, although the nature of
the perception, and of the act, originate in
consciousness. Clearly, a rhetorical act also
originates in these two aspects of perspective. The
Flower and Hayes model does not account for either
aspect of perspective, because the cognitive view does
not countenance ontological conditions. As I noted in
chapter three, a view that considers the mind to be
software that can be run on interchangeable hardware
configurations is not interested in the conditions of
embodiment.
Of course there is another alternative to the
dualist position. This is "physicalism," in which there
is no mind body split because the "mind" is simply the
program run by the brain, and consciousness is a product
of physical, mechanical operations such as the secretion
of hormones and the firing of neurons. If there is
nothing "supernatural" about the mind, if there is no
"ghost in the machine" and the mind is but a program,
how complex does the program have to be before it
becomes a mind, or an "an intentional agent, a being
that acts on its beliefs and goals"? Sterelny says,
That is a matter of some controversy. In
particular there is an ongoing debate
about (1) the degree of rationality
required, and (2) the need for some
perceptionlike flow of information from
the external world to the mind. But, as a
first approximation, let's say that an
intentional system must (a) have
perceptual systems of some kind, so there
is a flow of information from the world
into the system, (b) have a reasonably
rich system of internal representation;
thermostats aren't intentional systems in
part because they represent only
temperature, (c) have cognitive mechanisms
that enable it to use perceptual
information to update and modify its
internal representations of the world, and
236
(d) have mechanisms that translate its
internal representations into behavior
that is adaptive if those representations
fit the world (Sterelny Representational
Theory 12).
It is revealing of the problems of this sort of position
that we have to explain why a thermostat is not an
intentional agent. The physicalist position results in
some truly ingenious mechanistic descriptions. For
example, Fodor argues in Modularity of Mind for a system
in which there are three types of cognitive mechanisms.
At the skin there are transducers, which translate the
stimuli the world provides into neural code, ready for
"information processing." Next come the modular input
systems which process the raw data from the transducers.
Finally, the results are passed on to the central
processor (Sterelny Representational Theory 75). In
this description,
Transducers are the interface between
symbolic processing and the world. They
are on the mind-world boundary. The input
to the transducer is physical energy in
some form; its output is a symbol
(Sterelny Representational Theory 75).
But again, as we saw in chapter three, cognitive science
can deal only with the syntax of this processing. Even
if it were possible, would we be satisfied, as
compositionists, with a printout of the program of the
mind, toward which the Flower and Hayes model is a first
step? And would such a program be meaningful unless it
237
could be oriented to a world to which it referred? This
debate is at the heart of the conflict of methods that
is the subject of this investigation. The problems it
creates arise again and again.
Describing Individual Differences
The perspective of physical location and the
perspective of character define what is unique about a
particular individual's writing process, along with
accidents of history and experience. Flower and Hayes
claim that their model will account for individual
differences, but their analysis of individuals actually
leads to only four different "configurations" of the
monitor as it controls the goal setting module.
Each configuration corresponds to a
characteristically different way of
producing an essay. Configuration 1, for
example, corresponds to a style in which
the writer tries to produce a perfect
first sentence and then to follow the
perfect first sentence with a perfect
second sentence and so on. The work of
planning, translating, and reviewing each
sentence is completed before the writer
proceeds to the next sentence. With
Configuration 2, thoughts are written down
as they occur to the writer and he reviews
them later. With Configuration 3, the
writer tries to generate a perfect first
draft. Configuration 4 yields a
breadth-first composing process. A draft
is planned and then written out in full
before any review takes place
("Identifying the Organization" 19).
What they are describing are simply strategies for
238
generating text, strategies with which most writing
teachers are already familiar. Some of these strategies
are not as successful as others. For example,
configuration 1, in which the writer goes through the
whole writing process for each sentence, is generally
not a very efficient strategy for student writers. I
suppose that a teacher might attempt to re-configure the
monitor of a student who was having difficulty.
However, the chart of alternative configurations for the
monitor which they provide appears merely to shuffle the
order of the goals, and is not particularly helpful for
devising a pedagogical change. In fact, the more
important question is simply whether or not there is any
advantage in discussing these problems in terms of
configurations of an inaccessible monitor. It is also
clear that this model does not really account for
individual differences.
It is possible for two individuals to experience
the same object, but because of the different
perspectives inherent in the nature of bodily existence,
they can not have the same experience. It is this
condition of existence which makes communication between
individuals both necessary and problematic. However,
the body is also the ground upon which inter-subjective
understanding is based. Ricoeur argues that "empathy is
239
precisely the reading of the body of another as
indicating acts which have a subjective aim and origin,"
and that for this reason, subjectivity is both
"internal" and "external." Through communicating with
another, a relation is established to a body "which is
neither included in my perception of my own body, nor
inserted in an empirical acquaintance with the world."
Thus "knowledge of myself is always to some extent the
guide for deciphering the other" (Ricoeur FN 10-11). As
we saw in chapter one, Alfred Schutz calls the living
body of the other "a perceivable and explicable field of
expression" which makes his conscious life accessible
(Schutz and Luckmann I 62). Ricoeur argues that through
a "mutual elaboration of knowledge of self and other" we
arrive at "true concepts of subjectivity, valid for man,
my fellow" (The Philosophy of Paul Ricoeur 10). The
experience of the body of the other is the basis for
language and all inter-subjective, socially-shared, sign
systems.
For all of these reasons, Ricoeur argues, "a good
phenomenology is always implied in a good psychology"
(The Philosophy of Paul Ricoeur 10). In the psychology
of the Flower and Hayes model, however, the relationship
between consciousness and the external world, as well as
the relationship to other minds, are both problematic
240
and unexamined. What this model represents as
straightforward, uni-directional arrows are in actuality
complex interpretive processes in themselves. All the
problems of the subject's intentional relationship with
objects in the world, of semiotic interpretation, and of
intersubjective understanding, are problems of crossing
this ill-defined subject/object boundary.
The Task Environment
The task environment is the source of the
initiating input: the writing assignment and associated
constraints and cues. (One example of a "motivating
cue" given by Flower and Hayes is "the teacher's stern
expression when he presents an assignment" ("Identifying
the Organization" 12)). The elements of the rhetorical
problem present in the task environment must cross the
subject/object boundary, in some fashion, in order to
form the internal representation, which Flower and Hayes
say is "what finally matters." In addition, the written
text that is the output of the process gradually becomes
an object in the environment, establishing a feedback
loop that influences the process through the reviewing
module. The task environment box is thus necessary to
the model as a place where input originates and output
goes. Although the task environment could potentially
241
contain anything "outside the writer's skin," it does
not actually account for "everything" that might be
there. We must remember that cognitive approaches by
nature tend to de-emphasize context. The real purpose
of the task environment box is simply to account for
input and output, and the hook-up of the model to the
real world is actually quite limited.
In order to see the truth of the above assertions
we have to look at what is not in the task environment
box, and to do that we must again shift away from the
cognitive perspective. Patricia Bizzell characterizes
the Flower and Hayes approach as "inner-directed," and
states, "The Flower-Hayes model consistently presents a
description of how the writing process goes on as if it
were capable of answering questions about whv the writer
makes certain choices in certain situations."
(Actually, as we have seen, it is more accurate to say
that Flower and Hayes show us what the process does,
rather than how.) While Bizzell admits that it is
useful to have an overview of the how, she argues that
this information alone, even if it is accurate, will not
allow us to "advise students on difficult questions of
practice." Bizzell locates the answer to the whv
question in the discourse community. She says, "To put
it another way, if we are going to see students as
242
problem-solvers, we must also see them as problem-
solvers situated in discourse communities that guide
problem definition and the range of alternative
solutions" (Bizzell "Cognition, Convention and
Certainty" 222) . In other words, writing is a social
act, and we cannot understand that act outside of a
social context. The motivations for writing lie outside
the writer, except perhaps in the peculiar case of
Emig's reflexive mode.
Marilyn Cooper is another compositionist who
privileges contextual factors over cognitive processes,
but she argues that even the term "discourse community"
is too limited (Cooper and Holzman Writing as Social
Action 203). She argues, instead, for an ecology of
writing, and sees "writing as an activity through which
a person is continually engaged with a variety of
socially constituted systems" (Writing as Social Action
6). These ecological systems are "inherently dynamic;
though their structures and contents can be specified at
a given moment, in real time they are constantly
changing, limited only by parameters that are themselves
subject to change over longer spans of time" (7). She
argues, further, that "writing changes social reality
and not only . . . in response to exigence" (7). On the
other hand, Cooper argues that the ideal writer implied
243
by the cognitive model is isolated from the social
world, is "a solitary author."
The solitary author works alone, within
the privacy of his own mind. He uses
free-writing exercises and heuristics to
find out what he knows about a subject and
to find something he wants to say to
others; he uses his analytic skills to
discover a purpose, to imagine an
audience, to decide on strategies, to
organize content; and he simulates how his
text will be read by reading it over
himself, making the final revisions
necessary to assure its success when he
abandons it to the world of which he is
not a part. The isolation of the solitary
writer from the social world leads him to
see ideas and goals as originating
primarily within himself and directed at
an unknown and largely hostile other
(Writing as Social Action 4).
One can see this solitary writer as an inevitable
consequence of the cognitive, input/output model of the
writing process. However, Cooper has repeatedly
criticized the "ecological validity" of the research
done by Flower and Hayes, and it is also possible to
argue that the social isolation of this ideal writer is
due to the artificial nature of rhetoric practiced in
the classroom. In the social world, the "why" of
writing is almost always clear. In the classroom,
however, most students, if pressed to tell why they were
writing, would say that the teacher made them do it.
Whether this limited view of purpose derives from the
reduction inherent in the cognitive approach, or from
244
the artificial situation of the classroom, authentic
purposes, as we have seen in chapter two, involve both
individual motives and social values, and thus do not
originate solely within the writer. It is interesting
to note that Emig's model, which Flower and Hayes call
"impressionistic," does not exclude social values and
relationships.
In this analysis we have been jumping back and
forth across the subject/object boundary. Both Bizzell
and Cooper are arguing that what is outside the writer
is as important, and as complex, as what is inside. And
so, in contrast to the solitary author implicit in the
cognitive model, Cooper offers the image "of an
infinitely extended group of people who interact through
writing, who are connected by the various systems that
constitute the activity of writing" (12). For Cooper,
in order to understand the act of writing, it is these
systems which must be studied. This is not, however,
necessarily to privilege sciences of the outside over
sciences of the inside.
Levi-Strauss• structuralism, for example, is an
attempt to describe such social relationships in
scientific terms, but as we have seen, it is a choice of
syntax over semantics. Structuralism, like cognitive
science, is objective, empirical, and mechanistic. The
245
difference is that structural approaches operate in the
world, in the task environment if you will, while
cognitive approaches are directed inside. Structuralism
and cognitive science are, in the sense that they are
both syntactical, complementary approaches. By the same
token, phenomenology is complemented by an ethnographic,
participant observer approach. Ricoeur says of the work
of Clifford Geertz, for example,
We cannot approach perception without also
projecting a network of patterns, a
network, Geertz would say, of templates or
blueprints . . . through which we
articulate our experience. We have to
articulate our social experience in the
same way that we have to articulate our
perceptual experience. Just as models in
scientific language allow us to see how
things look, allow us to see things as
this or that, in the same way our social
templates articulate our social roles,
articulate our position in society as this
or that (Lectures on Ideology and Utopia
11) .
Of course, the debate between the advocates of cognitive
approaches to the study of culture on the one hand, and
interpretive or hermeneutic approaches on the other,
obtains in anthropology as well. Geertz discusses
Gilbert Ryle's example of the difference between a
twitch in the eye and a wink. From an empirical point
of view they are identical, yet from an interpretive
point of view the difference is vast (The Interpretation
of Cultures 6). A wink cannot be reduced to the
246
contraction of an eyelid. Geertz speaks of the
"cognitivist fallacy": the belief that culture consists
of mental phenomena which can be analyzed by formal
methods similar to mathematics and logic, and argues
that this belief is as destructive as the behaviorist
and idealist fallacies to which it is a misdrawn
correction ( The Interpretation of Cultures 12). He
quotes Ward Goodenough as a leading proponent of
cognitive anthropology (also known as "ethnoscience," or
"componential analysis"): "A society's culture consists
of whatever it is one has to know or believe in order to
operate in a manner acceptable to its members." Geertz
continues
And from this view of what culture is
follows a view, equally assured, of what
describing it is— the writing out of
systematic rules, an ethnographic
algorithm, which, if followed, would make
it possible so to operate, to pass
(physical appearance aside) for a native.
In such a way, extreme subjectivism is
married to extreme formalism, with the
expected result: an explosion of debate as
to whether particular analyses (which come
in the form of taxonomies, paradigms,
tables, trees, and other ingenuities)
reflect what the natives "really" think or
are merely clever simulations, logically
equivalent but substantively different, of
what they think (The Interpretation of
Cultures 11).
This is the familiar argument about the relation of the
model to the thing modeled. From Geertz's point of
view, the logical terminus of Flower and Hayes' project
247
would be an algorithm that could "pass" as a writer, a
standard that is a version of the Turing test.
Geertz's preferred methodology is what he calls
"thick description."
Ethnography is thick description. What
the ethnographer is in fact faced with . .
. is a multiplicity of complex conceptual
structures, many of them superimposed upon
or knotted into one another, which are at
once strange, irregular, and inexplicit,
and which he must contrive somehow first
to grasp and then to render . . . . Doing
ethnography is like trying to read (in the
sense of "construct a reading of") a
manuscript— foreign, faded, full of
ellipses, incoherencies, suspicious
emendations, and tendentious commentaries,
but written not in conventionalized graphs
of sound but in transient examples of
shaped behavior (The Interpretation of
Cultures 10).
Thus we have a strange paradox. On the one hand we have
cognitivists who disallow the interpretation of
language, and on the other we have ethnographers who
read culture like a text. If we are to make sense of
composition research it is necessary to know, at the
very least, which side the researcher is on.
Other Elements of the Model
So far our discussion has revolved around the task
environment box. It is time to turn to the other two
boxes in the model.
Stephen North notices that the box surrounding the
248
three main writing processes has no name.
What they never do, though, is explain the
characteristics of the figurative "box”
where all this takes Place. I attribute
that omission to the absence of formalism.
My guess is that Flower and Hayes, and
their readers, automatically fill in
something like "the writer's mind" in that
third box, attributing to it whatever
properties allow them to make sense of
these various processes. But, once again,
there is an obvious and dangerous
circularity in such thinking. The object
of building the model is to contrast it
with some empirical system. Any time an
element of the model slips from its status
as potentially analogous to become instead
identical with an element of a particular
system, the model loses potency. (The
Making of Knowledge 250-51)
Because long-term memory is also external to this box,
however, the unlabeled box must be intended to represent
short-term memory. It is odd that short-term memory
does not appear in the model specifically, because, as
became clear in chapter three, it is a concept that is
essential to an understanding of the use of protocols in
cognitive science. The theory supporting the validity
of protocol analysis depends on a clear understanding of
what types of information will appear in short-term
memory (STM). For example, Ericsson and Simon ask "What
is heeded and thus reportable in memory retrieval
without direct recognition?"
The retrieval cues have to be heeded and
kept in STM. What are reportable in
recognition, according to our model, are
only the input (the cueing stimulus) and
249
the output (the accessed LTM structure).
If the answer is not immediately
recognized, the cueing stimulus may be
augmented by heuristic search and other
problem-solving efforts, and the
intermediate products of these efforts can
also be reported. Which retrieval cues
were effective in accessing the LTM
structure can only be reported in
retrospect by recall of these intermediate
products or by inference from the nature
of the information recovered (Protocol
Analysis 134).
As I have said, generation is the process which
retrieves information from long-term memory into short
term memory. The Flower and Hayes model, however, does
not label short-term memory, a fact which is clearly a
major oversight. Stephen North also notes another
related element that is missing from the model. He
finds "the Formal invisibility of the protocols
themselves" to be puzzling.
Having made the decision to rely so
heavily on protocols as the output data
from which to infer a model, one might
well expect them to show up somewhere in
that model. "Text Produced So Far," for
example, is a cumulative feature of the
"Task Environment"; why isn't there a
parallel "Protocol Produced So Far"? (The
Making of Knowledge 266)
Here one could argue that North is wrong, at least in
asking for a "protocol produced so far" box in the
model. The protocol is not part of the model because it
is not natural— it is part of the intrusion of the
250
methodology, and thus should not be considered "output
data." Flower and Hayes focus on the protocol in part
because of the "process, not product" creed; the process
is the object of study and the protocol is thought to
reflect it. However, North is correct in asking that
the model account for the production of the protocol,
which it could do if short-term memory were part of it.
We can not simply call the unlabeled box "short-term
memory" because it contains processes, such as reading.
which although they leave markers in short-term memory
are not entirely conscious.
Stephen North also points out that the textual
specifications of the Flower and Hayes model are "pretty
rudimentary." He notes that "the model ignores almost
entirely the sheer logistics of a text: the shape of the
letters, their spacing, their left to right, top to
bottom arrangement," and asks, "Are these to be handled
by TRANSLATING? EDITING? How?" fThe Making of
Knowledge 247-49).
And the text conceived as units of other
kinds gets only the broadest attention.
That is, the model will account for two
general kinds of text (notes and
non-notes); three forms of organization
(actually only relevant to notes); and
three grammatical structures (words,
sentence fragments, and sentences). What
about such features as parts of speech?
Paragraphs? Clauses? The implication,
perhaps, is that the model's text-analogue
output will be governed by a grammar and a
251
rhetoric— the EDITING function tries to
cover most of this ground— but the model
provides for the details of neither. If
text is really the output data this model
is designed to produce, it has a very long
way to go towards specifying its formal
properties (The Making of Knowledge 249).
There are a number of sound questions behind North's
complaints here. The procedures used to test the model,
which I described and analyzed in chapter three,
constitute an attempt to connect process labels to
segments of protocol discourse, through categorizing the
formal characteristics of the related sections of the
inscribed text. This procedure simply assumes both a
grammar and a rhetoric, with specific characteristics,
without describing them or locating them in the model.
Perhaps, looking at the flow chart, the rhetoric resides
in the organizing function, and organizing and goal-
setting together produce a plan which is loaded into the
monitor, although there is no arrow to imply this, only
a connecting line between the planning box and the
monitor box. And in fact, the purpose of the monitor
appears to be goal-managing in order to decide which
process takes priority at a particular time. It is thus
not clear how editing accesses the rhetoric, or even how
it accesses the particular plan which has been
generated. Can it be that Flower and Hayes consider
editing to be a self-contained process which accesses
252
only a grammar, does not refer to the writing plan, the
audience, or the project as a whole, and ignores
revision processes other than correction? They describe
editing as a "production system" or "an ordered sequence
of condition-action rules."
The conditions of productions have two
parts. The first part specifies the kind
of language to which the editing
production applies, e.g., formal
sentences, notes, etc. The second is a
fault detector for such problems as
grammatical errors, incorrect words, and
missing context. When the conditions of a
production are met, e.g., a grammatical
error is found in a formal sentence, the
action that is triggered is a procedure
for fixing the fault ("Identifying the
Organization" 17).
Thus editing seems to include both matters of grammar
and usage, and matters of rhetoric such as how much
context is required by a particular audience.
Note that the logical structure of this process is
that of the TOTE unit, test-operate-test-exit, and is
exactly the manner in which a computer would handle the
same task. This is a preliminary and incomplete attempt
to imagine what the computational level would look like.
Note also that their analysis at this point is not
supported by data. It is purely conjectural, based on
assumptions made about the functional processes and the
assumption that the mind is like a computer.
They note that the requirements of an editing
253
process may trigger a return to generating, and thus an
embedded instance of the whole writing process. This is
how complex interactive processes are handled in this
model. If, for example, editing finds that the plan is
not sufficient or specific enough, it triggers a return
to a previous process which takes over. Stephen North
observes that while Flower and Hayes make it clear that
these processes are not stages, "they do seem to
suggest, if only by their flowcharts, that the various
sub-processes are operationally exclusive— that they
occur one at a time only, albeit in no set sequence"
fThe Making of Knowledge 250). This again is related to
the test-operate-test-exit way of proceeding. Thus
arrows in the model that intuitively seem as if they
should depict interaction in both directions, such as,
for example, the arrow pointing between generating and
goal-setting. are uni-directional. Information flows
like trains in a rail system, in that although there are
switches and sidings, the choice of routes is severely
constrained by the track.
Flower and Hayes attribute the pattern of process
interrupts to the fact that (in their opinion) writing
is goal-driven.
The writer's intuitions and the
persistence of his or her actions suggest
that writing processes are controlled by
goals. Thus, if writers report that they
254
are trying to organize material, they will
persistently return to organizing
processes even when those processes are
interrupted by editing and generating
("Identifying the Organization" 19).
Note that by "goal" Flower and Hayes do not mean
purpose. The purpose for the writing remains outside
the domain of the model, and the focus remains on the
processes which accomplish that undefined purpose. In
arguing for these one-at-a-time processes, Flower and
Hayes do not consider the possibility that there are
background and foreground processes which operate
simultaneously, and that in their analysis they are
merely following the movement of attention. If this is
the case the main process is not "interrupted," it is
just that attention has shifted momentarily to a
subprocess. Emig's characterization of writing
processes as "laminated" actually better reflects this
possibility. Such parallel processes are not excluded
by the information-processing view, however. Ericsson
and Simon associate fast automatic processes with
parallel processing, and indeed, computers are capable
of both parallel processing and multi-tasking, or
running several programs at once ("Verbal Reports" 235)
One wonders if the model of the mind will become more
and more sophisticated as our computers become more
complex and powerful.
255
Editing is contained in the reviewing box, but
Flower and Hayes call them "two distinct modes of
behavior." They say, "The function of the reviewing
process is to improve the quality of the written text
. . . to detect and correct violations in writing
conventions and inaccuracies of meaning and to evaluate
materials with respect to the writing goals"
("Identifying the Organization" 16-17). In
distinguishing the two processes, they say,
On the one hand, EDITING is triggered
automatically and may occur in brief
episodes interrupting other processes.
REVIEWING, on the other hand, is not a
spur-of-the-moment activity but rather one
in which the writer decides to devote a
period of time to systematic examination
and improvement of the text. It occurs
typically when the writer has finished a
translation process rather than as an
interruption to that process ("Identifying
the Organization" 18-19).
Because editing has been defined as addressing not only
problems of code but also problems of rhetoric, the
distinction between reviewing and editing appears to be
based on two factors: 1) decision, and 2) dominance.
The distinction between automatic processes and a
decision to engage in a project of revision is perhaps
more useful than Emig's formal distinctions of
correcting, revising, and rewriting, but are editing
processes ever truly automatic or involuntary? Flower
and Hayes seem to be more concerned about the placement
256
of the subprocess in the larger process, and the
duration, than in distinctions about volition in the
realm of consciousness. The second basis for a
distinction is that editing is said to interrupt other
ongoing processes, while reviewing is a dominant mode in
itself. If this is the case, however, why do we need
two labels? Are these actually two distinct processes?
Other processes interrupt dominant modes without
changing their names. Why not this one? The intent of
this model is to isolate the fundamental processes of
planned discourse, out of which more complex processes
can be constructed. In this instance we seem to have
the same process operating in two different situations
with different names.
The flow chart of the reviewing process exhibits
the typical TOTE unit pattern. It has the following
boxes: "Read next segment of text, edit segment, text
done?" with a recursive arrow from "text done?" back to
"read next segment" ("Identifying the Organization" 17).
Here we see that editing is intended to be the active
operational component of the larger process of
reviewing. The simplicity of this string of operations
is deceptive, however. From the point of view developed
in chapter two, each box represents processes and tasks
of incomprehensible complexity. The first box invokes
257
the reading process, which must be at least as
complicated as the writing process itself. While de
coding and interpreting the text, the writer is both
addresser and imagined addressee. When editing the
text, the writer's attention is also split between the
immediate text and the projected goal with its
associated audience. Clearly, as we saw in the
discussion of Emig's model, the writer also has in mind
the text that comes before and after the segment being
edited. Once editing has been attempted and changes
made, most writers would test for correctness,
completeness and appropriateness through another reading
from these split perspectives (If I were constructing
this model, I would add another box.). It is perfectly
legitimate to isolate and label a subprocess that cannot
be disassociated from the major process. However,
because it is impossible to edit without reading or re
reading, the distinctions made by Flower and Hayes
concerning decision and dominance seem confused.
The "Text Done?" box that finishes off the
procedure seems absurd from a human perspective. It is
true that a computer proceeding in this fashion would
need to check after each segment to see if another one
existed. Humans perceive such objects as texts in
gestalts and do not need to perform this exit test after
258
each segment. Even if the current segment came at the
end of a page, in most cases a human editor would not
need to turn the page to discover if the next segment
existed, but would know that the text was incomplete
from other cues.
Similar problems can be seen in the flow chart of
the generating process. In general their concept of
generation is an improvement over Emig's concept of pre
writing, because any time the writer perceives that
information or structure is lacking, the generating
process is triggered. The flow chart which follows
appears in "Identifying the Organization of Writing
Processes" (13).
259
REPLACE CURRENT
MEMORY PROBE WITH
NEW PROBE
RETRIEVE USING
CURRENT MEMORY
PROBE
FAIL
SUCCEED
YES
RETRIEVED ELEMENT
==>CURRENT MEMORY PROBE
NOT
USEFUL ■EVALUATE^
RETRIEVED
.ELEMENT/
^ GOAL= ^
GENERATE?
NO
USEFUL
(IT
CONSIDER NOTE
YES
NO
WRITE
NOTE
YES ^ GOAL= ^
GENERATE?
NO
EXIT
Figure 4: The Structure of Generating
It is interesting that this process is called
"generating" when its primary function as described is
260
information retrieval. Defined in these terms, the
model appears to have no way to deal with insight or any
sort of creation. And, as Stephen North has observed,
the model is strongly connected to certain forms of
inscribed text, in this case a certain use of
notetaking. It appears from the chart that elements
which are retrieved but which do not generate a note
simply disappear. If, for example, an individual were
rehearsing what he was going to say when it became his
turn to speak in a meeting, would not the same sort of
retrieval process take place, perhaps without any
notetaking action? And of course, ancient orators were
able to memorize long speeches and recite them backwards
and forwards. Although the number of elements that
short-term memory can hold is limited, cognitive theory
allows that large amounts of information can be held at
ready access through the use of placeholding keys that
connect to larger information structures. Even modern
writers do not necessarily work from notes, but may
prefer to write drafts and revise them, or use other
strategies. The inclusion of this notetaking box in a
formal model is evidence that assumptions derived from
the specific rhetorical practices of particular
individuals have been incorporated.
This chart also illustrates what Flower and Hayes
261
mean by their assertion that the writing process is
"goal-driven." Each process or subprocess has a related
goal. At the end of each cycle, a decision is made
concerning the present status and priority of the
relevant process goal. If that goal is no longer
primary, the process exits to the monitor, which will
assign a new goal and initiate a new process. (The
"monitor" in this model is not at all similar to the
"monitor" in Stephen Krashen's Language Acquisition
model. This monitor is really a central processor,
which manages and orchestrates the various processes and
their inputs and outputs.) As we have seen, this
concept of "goal" is quite different from what is meant
by "rhetorical purpose" or "project." In this model the
"goal" and the process are nearly indistinguishable, or,
in other words, the "goal" is to do the process.
The chart above descends to a lower level of detail
than appears in the chart of the main writing processes,
where there is merely a box labeled "generating."
However, this is by no means the lowest level of detail.
We have only to consider the box labeled "evaluate
retrieved element" in order to see clearly that further
unseen and unexplained processes are involved. How does
this evaluation take place? What interpretive processes
are involved and what kinds of knowledge are accessed?
262
Unfortunately, these are more interesting questions for
writing teachers than the questions the model attempts
to address.
It is also important to note that while Flower and
Hayes have made some attempt to validate the main
features of their model (discussed in chapter three),
they have not tested these extensions of their model
into lower levels of detail. Protocol analysis does not
offer easy access to these sub-processes.
Implications of the Model for Teaching
In "The Dynamics of Composing: Making Plans and
Juggling Constraints," Flower and Hayes apply their
model to writers in action. They make two major points:
The first is that the act of writing is
best described as the act of juggling a
number of simultaneous constraints. This
is in contrast to seeing it as a series of
discrete stages or steps that add up to a
finished product. Second, we suggest that
one of the most effective strategies for
handling this large number of constraints
is Planning. Plans allow writers to
reduce "cognitive strain," that is, to
reduce the number of demands being made on
conscious attention. They also create a
nested set of goals that allow a number of
constraints to be satisfied at once
(31-32).
Flower and Hayes always disparage stage models of the
writing process, and indeed, one of the main strengths
263
of their model is the fact that although there are a
limited number of basic processes, they can interrupt
one another at various times under various circumstances
in order to account for considerable complexity.
Notice, however, that the concept of simultaneity is
introduced in regard to constraints rather than
processes. In practice, as we have seen, the three main
processes in the Flower and Hayes model tend to operate
one at a time in the traditional sequence.
As solutions for juggling competing constraints in
order to reduce "cognitive strain" Flower and Hayes
offer the following list: Throw a constraint away,
partition the problem, set priorities and "satisfice,"
draw on a routine or well-learned procedure, and plan.
They argue that planning is the best strategy for
reducing cognitive load. In this article, they
distinguish among "plans to do," which relate to the
rhetorical problem, "plans to say" which are
organizational and content schemes, and "composing
plans" which are supposedly process based and include
invention strategies. Selected protocols serve as
examples to illustrate their ideas and taxonomies,
rather than as proof or evidence ("The Dynamics of
Composing" 41-46).
Along with the taxonomy of plans they offer a
264
taxonomy of constraints: knowledge, language, and
rhetoric. "The first constraint we describe as the
demand for integrated knowledge; the second is the more
inclusive linguistic conventions of written texts; and
the third is the encompassing constraints of the
rhetorical problem itself" ("The Dynamics of Composing"
34). If Flower and Hayes can connect their model to
actual writing and teaching strategies, its power and
utility will be greatly enhanced. These constraints
relate to the model in the following ways— the first
concern, that of knowledge, is the domain of the
generating process, the second, that of linguistic code
and the conventions of usage, belongs to the translating
and reviewing processes. However, the third process,
the representation of the rhetorical problem, is not
clearly represented in the model, although Flower and
Hayes stress its importance a number of times.
Whatever writer's choose to say must
ultimately conform to the structures posed
by their purpose in writing, their sense
of the audience and their projected selves
or imagined roles. In essence, writing is
also a speech act and therefore subject to
all the constraints of any interpersonal
performance .... Although the
Rhetorical Problem is an added constraint,
it can't be conveniently "added" at the
end, because, in theory, it should direct
the entire process of generating knowledge
and language ("The Dynamics of Composing"
40) .
Here Flower and Hayes have identified perhaps the major
265
weakness of their model, but they do little to correct
it. All the confusion about goals, plans, and processes
derives from the fact that the rhetorical purpose is
missing from the model. All writing is aimed at doing
something. Of course, we might assume from the arrows
in the chart that the relevant information about the
rhetorical purpose in the task environment is passed to
the generating module, which then seeks relevant
information from long-term memory and passes all of this
to the goal-setting module to be synthesized. This is
only conjecture, however, and, as Patricia Bizzell
notes, even if this is how it works, "They correctly
identify goal-setting as the motor of the composing
process, its most important element, but in their model
they close it off in the most subordinate position (a
subdivision of a subdivision of the writing process)"
("Cognition, Convention, and Certainty" 227). If model
and theory cannot be reconciled, there is a problem with
one or the other or both.
The statements in this passage pose a number of
other difficulties. First, it is strange that Flower
and Hayes say that "in theory" the rhetorical purpose
should direct the entire process, when their research is
presented as theory-driven. They usually identify their
approach as "top-down" ("Writing as Problem-Solving 389;
266
"The Pregnant Pause" 2 34). Second, considering the
nature of the model created by Flower and Hayes, and the
research methodology they have adopted, it is surprising
that they characterize writing as a "speech act."
Speech Act Theory shows that context is essential to the
interpretation of many utterances. In this model we
have a task environment, but context is minimized. And
third, although they see writing as goal-driven, the
model they create does not incorporate a "project" as a
fantasized future act with a purpose (though they do
mention the writer's "projected selves," without
explaining what they mean). Perhaps this is why they do
not notice the shuttling back and forth of attention
between the immediate text and the larger goal that Emig
describes in the protocols of her subjects. In a sense
they are doing exactly what they say cannot be
done— simply adding a rhetorical concern at the end of
the analysis.
As noted above, Flower and Hayes consider knowledge
to be another constraint on writing.
Generally speaking, Knowledge is a
resource, not a constraint. However, it
becomes a constraint on the process when
it is not in an acceptable form ....
Much of the work of writing can be the
task of transforming incoherent thought
and loosely related pockets of information
into a highly conceptualized and precisely
related knowledge network ("The Dynamics
of Composing" 34).
267
Flower and Hayes separate knowledge, language and
rhetoric, yet this transformation they are talking about
is really the organizing of knowledge in language. for
someone else, for some purpose. As I noted in chapter
one, Schutz points out that the stock of knowledge may
be contradictory or incoherent because it is derived
from different situations or perspectives. But these
contradictions must also be resolved from a perspective,
and that perspective is the rhetorical situation. As
Bizzell says "Writing is always already writing for some
purpose that can only be understood in its community
context" ("Cognition, Convention, and Certainty" 227).
Flower and Hayes are trying to separate the problem of
"What do I know?" from "What do I want to say?" and
"What effect do I want to create and how do I achieve
it?" These problems can only be separated abstractly,
or, in other words, by brackets which are always
temporary. In attempting to construct their model out
of discreet processes, Flower and Hayes have created a
kind of atomism. Flower and Hayes recognize this
problem, but do not al1;er their model to support a more
interactive analysis. This leads occasionally to
discourse which is contradictory, in which what they say
they believe and what their model implies are quite
different.
268
Composition is a practical discipline. If a formal
cognitive model is to be useful to teachers, it has to
be connected to strategies that teachers can use. This
is one purpose of articles like "Plans that Guide the
Composing Process." Flower and Hayes argue
Good writers have a repertory of powerful
heuristics which might include
brainstorming. planning or simulating a
reader's response techniques. Weak
writers may lack these techniques or use
them ineffectively; or they may use other
techniques which are less effective or
even counterproductive ("Plans That Guide"
40) .
Here is a difference between weak writers and good
writers that relates to one of the composing processes
identified by the researchers. Perhaps the model can be
used to illuminate the differences between good writers
and poor writers in the planning module. This is an odd
list of heuristics, however. Brainstorming, which is
essentially a temporary suspension of the process of
judgment or evaluation, is the invention strategy that
seems to fit most neatly into the information retrieval
process that describes generating, and clearly, some
writers are unfamiliar with this strategy, or do not use
it well. On the other hand, planning is one of the
three main boxes that represent the major writing
processes. It is unlikely that a weak writer would
entirely lack the ability to engage in this process.
269
Thus when Flower and Hayes speak of "planning" as a
heuristic, they must have something else in mind,
something that can be influenced by instruction.
"Simulating a reader's response" is a powerful technique
related to audience awareness and purpose, but audience
awareness seems so basic to the act of writing that to
consider it a heuristic is to improperly minimize its
importance. Either the conceptualization or the
expression is fuzzy here.
The authors discuss the types of plans they have
discovered in protocols. Essentially, these are "Plans
for Generating Ideas" and "Plans for Producing a Paper,"
although they also have a category called "Plans for
Controlling the Composing Act."
In each category they have "procedural plans, 1 1 which
are supposed to be content-free, and "content-specific"
plans. Flower and Hayes say,
The chief advantage of . . . procedural
plans is that they provide a continuing
structure for the composing process. For
example, a plan to brainstorm or to
outline creates a high-level goal to which
the writer can return when ideas stop
flowing or when writing takes an
unproductive turn ("Plans that Guide" 43).
Cooper and Holzman have argued that Flower and Hayes use
the terms "goal" and "plan" indiscriminately ("Talking
about Protocols" 288). Flower and Hayes reply that "it
is the information which counts, not the label"
270
('•Response to Cooper and Holzman" 94) . Here it seems
that Cooper and Holzman are right. Most people would
see a "plan" as a means toward achieving a "goal." Here
it seems that the "goal" is to brainstorm, or create an
outline, and means and ends are confused. As we have
seen, this sort of confusion is reflected in the process
labels in the model itself, where process and goal are
blurred.
Perhaps the problem is that Flower and Hayes simply
don't mean by planning what most of us mean by it. In
"Images, Plans, and Prose" they say,
Planning is a generic cognitive process
that appears to encompass various
processes writers often describe as
incubation, getting it right with oneself,
finding a focus, or discovering what one
means. Planning, in our model, is the
purposeful act of representing current
meaning to oneself (124).
Thus it is clear that by "planning" they do not mean
anything like what Ricoeur would call "projecting."
They go on to say that "in one sense the act of writing
is simply an extension of planning— the process of
invention and testing continues— and it is also an
extension of the linguistic processes of speaking"
(150). These definitions are so broad as to be
confusing at best.
The content-specific plans are more closely tied to
the protocols, although again, protocol excerpts are
271
used to illustrate only, and no information is given
about who the writers are, except that they are
students. We do not know if this article is further
interpretation of earlier data, or new data.
Given the standard cognitive methodology discussed
in chapter three, what we might expect to find in this
study is something like this: First, it is customary to
have a description of the subject population, perhaps
"thirty students enrolled in freshman writing at
Carnegie-Mellon university in Fall, 1980." Second, we
expect coding categories developed from sample
protocols, which they have provided. For example, under
"Plans for Generating Ideas" they list: (1) Pursuing an
Interesting Feature; (2) Thinking by Conflict; (3)
"Saying What I Really Mean" and (4) Finding a Focus
("Plans That Guide" 42). Third, a procedure for
segmenting the protocols is necessary. Fourth, there
should be a procedure for having the segments coded
according to the category system by independent raters.
If these things had been done, then we could investigate
some interesting questions. For example, we could then
rank order the frequencies of the four strategies and
determine which strategies were most common in this
population. We could subdivide the population and
compare strategy frequencies in different groups, for
272
example older writers versus younger writers, or good
writers to poor writers (to do the latter we would have
to do an additional procedure to determine which were
which, perhaps by having the compositions rated).
Unfortunately, none of these things were done. The
categories themselves, illustrated with examples from
selected protocols, are the results.
Two of the categories given above sound as if they
originate in instruction: "Thinking by Conflict" and
"Finding a Focus." As an example of "Thinking by
Conflict," Flower and Hayes give "the procedure of
constructing parallel columns of pros and cons as an aid
in deciding what to include in an essay" ("Plans that
Guide" 44). They do not argue that this practice
reflects an underlying cognitive process, nor do they
speculate on where the writer acquired the practice.
They just notice that their subjects do it. They do
speculate a bit about "Finding a Focus."
Although it is hard to know what focus
actually meant to the writers we studied,
it seemed to be equated at times with the
fully articulated thesis statement which
textbooks often advise a writer to
formulate before beginning to write. Yet,
as these protocols show, writers typically
don't start with a thesis or well-focused
body of ideas. Instead they start with a
body of knowledge and set of goals, and
they create their focus by such complex
actions as drawing inferences, creating
relationships, or abstracting large bodies
of ideas down to what I really mean. If a
273
writer believed that the appropriate focus
was to be found by searching memory, or
searching a book, his plan to find a focus
might well fail ("Plans that Guide" 45).
There are many interesting questions here, but it is
also true that many of the problems inherent in the
cognitive method, and in the practice of Flower and
Hayes in using that method, are apparent. It is
necessary to know what focus means to these writers, but
to do so we must interpret what they say, and their
words are no longer strictly data as defined by Ericsson
and Simon. However, Flower and Hayes say that the
concept of focus these writers have seems to have
originated in textbook advice, and thus outside of
consciousness in the social world. Flower and Hayes do
not often speculate about where ideas come from, because
the focus of their interest is on cognitive processing.
Here, however, it seems clear that the protocols do not
make sense without reference to the structures of the
social world.
The observation that writers create their focus out
of a set of goals and a body of knowledge, using
processes of inference and association, is more
interesting, but they identify this creation of focus as
being equivalent to what I really mean, which is one of
the other categories they have generated. This analysis
leads one to believe that these two categories might
274
reflect the same process, which is perhaps in some cases
derailed by the textbook advice. Flower and Hayes seem
to have lost sight of their original goal to elicit and
describe the nature of cognitive processes through their
manifestations in the protocols.
At any rate, "Pursuing an Interesting Feature" and
"Saying What I Really Mean" are much more interesting
categories, which may, in fact, reflect actual cognitive
processes. Of "Pursuing an Interesting Feature" they
say,
The feature may be a word, or idea, or
event. Initially, the writer may have no
clear idea where a generating process
centered on this feature might lead. But
in the process of pursuing it, the writer
will usually develop more explicit sub
plans for exploring the feature and turn
to various generating techniques, such as
searching memory, drawing inferences,
reasoning from examples, or matching
current evidence to prior knowledge
("Plans that Guide" 44).
This seems to be the basic process of association, as
described by Ericsson and Simon. Analysis of such
episodes could lead to a more sophisticated concept of
how information is associated in long-term memory, and
how it is re-associated in the process of composition.
This would seem to be a valuable line of research.
"Saying What I Really Mean," which they also call
"What I Really Mean Is," or "WIRMI," is described as "a
plan writers appear to use when they want to abstract or
275
reduce a complex body of information to its essential
features" (44). In this sense it is a reverse of the
association process described above. It is the
discovery of a perspective that connects or unifies the
information and makes it meaningful for a particular
purpose. Again, episodes of this type would seem to be
profitable areas of investigation.
I
Flower and Hayes are not insensitive to the
phenomenological potentialities of the protocol
discourse. For example, they quote the following bit of
a protocol, and note that it "demonstrates the end of
the line experience writers frequently undergo when a
train of thought has been completed":
So, I don't really know what I've done. I
mean, I kinda talked about how I was
convinced a little bit, but I don't know i
that I really did it too precisely. And I |
don't know if this is really the best; it
seemed like an okay one— it kinda led
somewhere . . . .I'm trying to get myself
going again. I'm in a mire. I don't
know, I just stalled out, you know; my
mind just pooped out after it had all
those ideas. The reason I can't generate
anything at this point, I was just kinda
following things along, you know (43).
This end of the line feeling is a subjective experience
probably common to all writers, and certainly worthy of
t
study and analysis. Flower and Hayes simply attribute j
this feeling to the lack of a procedural plan, but it
seems just as likely that this "stalled out" feeling is
276
an unavoidable aspect of good writing, a signal that new
strategies, new perspectives, or new data are necessary.
It is interesting in another sense that the writer
uses the word "generate" when he is engaged in what
Flower and Hayes call the "generating" process. This
raises the question of whether or not the writer has
been exposed to the model before becoming a subject in
this research. In this regard it is also interesting to
note that the writer says "my mind just pooped out after
it had all those ideas." To my ear this is an odd
locution. Is it reading too much into this to wonder if
the writer is thinking of the abstract mechanistic mind
of the Flower and Hayes model, and the generating box?
Flower and Hayes do not raise the issue of how
instruction influences process or contaminates their
data.
Representations and the Hind
It is fair to say that later work by Flower and
Hayes is more rigorous, better presented, and more
philosophically interesting than the earlier, but still
very influential, articles I have been critiquing. For
example, in "Images, Plans, and Prose," published in
1984 in the first issue of Written Communication, they
begin by asking,
277
When a writer says "I know what I mean,
but I can't express it," what does that
writer know? Is that feeling of "knowing"
a mere illusion, or is her knowledge
really equivalent to the knowledge
expressed in a finished essay? The same
question, "What does the writer really
know?" can be asked about intuitions and
that elusive sense of a "felt difficulty"
that initiates problem solving .... In
sum, what do you know if you cannot say
it? and if you can say it, what do you
gain by saying it in prose? (122)
This is a productive way to ask these questions. They
begin with a common subjective experience, the struggle
to express something that is known. They reason that if j
something known is difficult to express in words, that
something must be represented in some fashion other than
in words, and the problem is one of translation. This
is why they call the "composing" process "translating."
i
The rationale behind this term is not really made clear
until this article. The interesting aspect of this
argument is that it originates in a phenomenological
experience. (Of course, as I noted in chapter three,
Derrida also argues that "we think only in signs," but !
he carries the argument farther until there is no
originary presence. For Derrida, there is no
"transcendental signified," no meaning that can be
separated from a semiotic representation. Therefore, (
I
the concepts of re-presentation, translation, even
signification, are all suspect.)
278
This argument also assumes the representational
theory of mind, a theory closely associated with
cognitive science but not without controversy within it.
Kim Sterelny argues
To the considerable extent that our
behavior is adaptively flexible and
informationally sensitive, to that extent
it must be directed by representations.
There can be no informational sensitivity
without representation. There can be no
flexible and adaptive response to the
world without representation. To learn
about the world, and to use what we learn
to act in new ways, we must be able to
represent the world, our goals and
options. Furthermore we must make
appropriate inferences from those
representations (Sterelny Representational
Theory 21).
Sterelny's example of a creature which cannot do this is
the robin, which will feed a cuckoo chick even after it
has killed the young robins and even though it is much
larger and does not look like a robin. He also
discusses the sphex wasp, which digs a burrow, paralyzes
a cricket and leaves it in front of the opening. The
wasp then enters the burrow for an inspection, returns
to the surface, and then installs the cricket as food
for the larvae. If a researcher moves the cricket after
it has been deposited in front of the burrow, the wasp
will return it to its place and inspect the burrow
again. The wasp will inspect the burrow as many times
as the cricket is moved.
279
Sterelny argues that the behavior of the robins and
of the wasp is not "adaptively flexible and
informationally sensitive." The robins cannot
differentiate between a robin chick and a cuckoo;
therefore, Sterelny would argue, they do not have a
representation of a robin chick or cannot compare their
representations. If the nest is empty, however, they
will cease their feeding behavior, so it seems that
their behavior is informationally sensitive to a degree.
The sphex wasp notices that the cricket is out of place,
and therefore it must have a representation of where it
put the cricket. It just cannot remember that it has
already inspected the burrow, a different matter
entirely. Both of these examples exhibit algorithmic,
computer-like behavior. The ability to create and
manipulate mental representations is offered here as an
explanation for why humans are able to do more than a
computer can presently do. Flower and Hayes' concept of
planning depends on this kind of comparison— the writer
takes the current internal representation of meaning and
compares it to the external representation to see what
must be done to bring the two representations into
closer similarity.
The form and structure of these mental
representations is also controversial. Sterelny defends
280
what he calls "the language of thought hypothesis."
The point of this hypothesis is to explain
how internal states can simultaneously
have the function of representing the
world, and of directing behavior in ways
that make sense given their
representational content. Very roughly,
according to this hypothesis, thoughts are
sentences in the head. They represent or
misrepresent the world, in virtue of
causal relations of some kind between
chunks of these sentences and individuals
and types of individual in the world.
When you think about beer, you have within
your head a mental sentence with a
mentalese expression for beer in it. It's
an expression for beer because it stands
in some kind of causal relation with the
stuff. Thoughts are language-like
representations of the world. But the
thoughts also direct behavior, and an
account of how they do that links the
representational theory of the mind to
computational theorizing about the mind.
There are plenty of uselessly loose
analogies between minds and computers
around, but in my view that model does
real work in explaining how internal
representational structures can direct
behavior (Sterelny Representational Theory
x) .
If mental representations are propositional and
language-like, there is also controversy about what
language they are in. Do we think in natural language,
or in "mentalese," or in a combination of the two at
different levels of consciousness? And are all
representations "language-like"? How do we remember
odors? Does the mind store images?
I
Flower and Hayes develop a notion they call the !
I
"Multiple Representation Thesis."
281
The essence of this thesis is simple: As
writers compose they create multiple
internal and external representations of
meaning. Some of these representations,
such as an imagistic one, will be better
at expressing certain kinds of meaning
than prose would be, and some will be more
difficult to translate into prose than
others. Much of the work of writing is
the creation and the translation of these
alternative representations of meaning
("Images, Plans, and Prose" 122).
Clearly Flower and Hayes are sidestepping a number of
controversial arguments in arriving at this position.
Many cognitive researchers have trouble with ideas such
as imagistic representations because it is difficult to
fit them into a computational theory of mind. On the
other hand, it is also hard to imagine that the sight,
smell, taste and texture of Sterelny's glass of beer are
represented internally as sentences.
The division of the class of representations into
internal and external leads to a view that rhetoric is
simply a body of knowledge which tells us of "the forms
public representations of meaning can or should take"
("Images" 123). The composing process, in this view,
involves a series of translations— from the stuff of
memory to the stuff of thought (which appears, in part,
in the protocol), to the written words of the external
public representation of meaning. The work or effort of
writing, Flower and Hayes argue, goes into making these
translations. The externalization of a representation
282
of meaning they call Minstantiation."
The first form of instantiation is one we
cannot see. It is the mental act of
transforming an internal, conceptual
representation (stored in whatever way the
brain encodes such things) into a verbal,
external analog ("Images’ ' 153) .
Flower and Hayes say that internal representations "are
instantiated in prose not by a kind of automatic
translation, but by an active rhetorical decision
process" ("Images" 153). Thus, it becomes clear that in
the Flower and Hayes model the generating process
produces the internal representation of meaning, which
is then re-represented or "instantiated" by the
translating process as an external analog of the
internal meaning. If we are to avoid a simplistic stage
model, however, we must also recognize that the internal
representation is dynamic, and that it is not isolated
from the social and rhetorical factors which influence
the process of creating the external representation. In
other words, the perspectives inherent in each type of
representation must cross-influence one another in a
dynamic way. This view encompasses considerable
complexity.
The products of the individual processes described
by the Flower and Hayes model are representations of
meaning, which are then handed off to other processes
for further operations. The only representation which
283
1
appears in the model as published, however, is "Text
Produced So Far," the external representation which is
the output of the whole process. This is a serious
deficiency, because it is difficult to understand what
these processes are supposed to be doing, if one cannot
see what they are operating on. Ultimately, from this
point of view, the composing process consists of sub
processes which create, compare, and perfect multiple
representations of meaning in different codes and for
different purposes, but this is not obvious from the
model as it has been published. The models presented in
Reading to Write. Linda Flower's latest work, are much
more specific about the role of representations in the
I
composing process.
Chapters three and four have dealt with the
philosophical implications of the cognitive view, the ;
methods and assumptions of cognitive science, and the
difficulties and implications of the research conducted
by Flower and Hayes, as well as the model generated out
of that research. What I have tried to emphasize are
the necessary assumptions which are behind such
I
research, assumptions which are clear to researchers in
cognitive science, but which are not always clear to
those in other disciplines who wish to use cognitive
research. We have for this reason dwelt long on the
i
284 i
object side of the subject/object cut. The final
chapter will return the emphasis to the phenomenological
perspective, to explore later work by Janet Emig and
Linda Flower, and to discuss the possibilities and
hazards of a synthesis of these various methods. One of
the crucial problems of this synthesis is the problem of
understanding and interpretation: the reference of
language to a world. In this regard, Ricoeur’s
phenomenological hermeneutics will be particularly
helpful. I will also discuss Composition as a Human
Science by Louise Phelps, especially regarding the
application of Ricoeur's thought to the problem of
defining composition and rhetoric as a field of study.
i
285
CHAPTER FIVE
FROM SIGN TO SEMANTICS:
THE PROBLEMS OF REFERENCE AND MEANING
As I noted in chapter one, the shift in focus from
text to process was a fundamental reorientation in
composition practice and research, largely initiated by
The Composing Processes of Twelfth Graders. By 1977,
when "Hand, Eye, Brain" was published, Janet Emig was
able to state a very strong process centered position
with new confidence, and to begin to lay out a new
process-centered research program.
Much of the current talk about the basics
of writing is not only confused but, even
more ironic, frivolous. Capitalization,
spelling, punctuation— these are touted as
the basics in writing when they represent,
of course, merely the conventions, the
amenities for recording the outcome of the
process. The process is what is basic in
writing, the process and the organic
structures that interact to produce it
(Emig The Web of Meaning 110).
Emig then asks about the nature of these organic
structures and their contributions. At first glance it
would seem that her interests are congruent with those
of Flower and Hayes. However, her candidates for these
286
"requisite structures" are the hand, the eye, and the
brain itself. In Emig's work, action, perception, and
cognition are never anything less than integrated.
In "Writing as a Mode of Learning," also published
in 1977, Emig is clearly influenced by Vygotsky and
Luria, and is interested in the issues of orality and
literacy. Here she says that writing is artificial, is
a technology. She quotes Vygotsky, noting that writing
makes a unique demand upon the writer, that the writer
must engage in "deliberate semantics . . . deliberate
structuring of the web of meaning" (Emig The Web of
Meaning 127). As her work develops, Emig's orientation
becomes less psychological and more social, until she
characterizes atomistic, non-social learning theories as
"magical thinking." She lists the assumptions of the
"magical thinking paradigm:"
1. Writing is predominantly taught rather
than learned.
2. Children must be taught atomistically,
from parts to wholes.
3. There is essentially one process of
writing that serves all writers for all
their aims, modes, intents, and audiences.
4. That process is linear: all planning
precedes all writing . . . as all writing
precedes all revising.
5. The process of writing is almost
entirely conscious.
6. Because writing is conscious, it can be
done swiftly and on order.
7. There is no community or collaboration
in writing; it is exclusively a silent and
solitary activity (Emig The Web of Meaning
140) .
287
Of course, Stephen Krashen would use "acquired" rather
than "learned" in this analysis. Emig follows this list i
with the "non-magical thinking paradigm," which turns
each of these assumptions on its head. Flower and Hayes
avoid most of these "magical" assumptions, or claim to
avoid them. They argue, for example, that there are
individual differences in writing processes, but as we
have seen, what they mean is that there are different
configurations of the monitor which controls the process
(they identify four). They state numerous times that
the writing process is not linear, but in practice this
means that there are momentary process interrupts, while
the main processes proceed in the traditional order. It j
t
is with point seven, however, as Marilyn Cooper and j
Patricia Bizzell have both pointed out, that the
greatest weakness of the cognitive approach is apparent.
Although it is clear that writing is both social action
and interaction, and that written communication takes
place in a community. the cognitive models of the
composing process developed so far do not represent this
reality.
Emig's response to these dilemmas is to search for j
a new paradigm. She characterizes her own development [
i
, I
as a researcher as first a focus on description of
i
phenomena, then analysis, and finally theorizing. All |
these activities, however, require a perspective, what
Emig calls a "governing gaze." In the essay, "Inquiry
Paradigms and Writing," Emig discusses what might be
called epistemologies, or ideologies, of disciplines.
She states that to qualify as an inquiry paradigm, an
endeavor must be informed by
1) a governing gaze; 2) an acknowledged,
or at least conscious, set of assumptions,
preferably connected with 3) a coherent
theory or theories; 4) an allegiance to an
explicit or at least tacit intellectual
tradition; and 5) an adequate methodology
including an indigenous logic consonant
with all of the above (Emig The Web of
Meaning 160).
She notes that a number of inquiry paradigms generally
co-exist in any given discipline, though not without
conflict.
By these standards, cognitive science certainly
qualifies as an inquiry paradigm. Confusions have
developed, however, out of the conflict between the
assumptions of cognitive science and those of
composition in general. If the object of study is
inappropriate to the methodology, the temptation is to
mix methods, or graft inappropriate assumptions onto the
results. As I have noted before, Emig says "A mature
inquiry paradigm . . . requires an appropriate
methodology .... Dissonance between methodology and
intent or methodology and content, is often easy to
289
discern" (Emig The Web of Meaning 166-67).
Emig compares positivism and phenomenology as
"governing gazes." She says that for the
phenomenologist, the width of the gaze must include the
field, whereas for the positivist there is no field,
only focus. The positivist engages in what Elliot
Mishler (Harvard Medical School, "Meaning in Context: Is
There Any Other Kind? Harvard Educational Review, 49
(Feb., 1979) 1-19) calls "context stripping." Emig notes
that the phenomenologist is concerned with understanding
human behavior from the actor’s own frame of reference. j
I
Emig includes her own case-study methodology and
Scribner and Cole’s ethnographic approach as examples of
phenomenological research (Emig The Web of Meaning
161-63). This amounts to an admission that the research
documented in The Composing Processes of Twelfth Graders i
is not as empirically-based as she may have thought at
the time, as well as a realization that it was better
that it wasn't so.
The concept of the composing process introduced by
Emig both liberated and ultimately limited composition
as a discipline. Louise Phelps, in her recent book
Composition as a Human Science, argues that the term
"process" seemed to be a "fruitful, necessary and even
obvious conception of subject matter in the field" when
290
it was introduced. She notes that the research inspired j
by this reconceptualization broke sharply with the past i
in claiming that teaching should rest on systematic
knowledge about writing as an activity, and in seeing
writing is a fundamental act of the human mind, as
opposed to the prevalent idea that writing other than
literature is trivial and mechanistic. In Phelps' view,
however, the founding of a discipline on the idea of
process was philosophically naive, because
process-oriented researchers did not understand that the
composing process was a concept and not an object of
study. Phelps notes that during this period the field
lacked all sense of irony about its terminology and
premises. "The idea of process promulgated throughout j
the 1960's and 1970's is almost entirely unanalyzed. j
The term functions indexically to call for empirical |
I
description of the named event rather than for |
1
theoretical construction of its meaning" (Phelps Human |
i
Science 42). j
i
One of the main tenets of Composition as a Human
i
i
Science is the argument that Composition as a field is J
now in the process of maturing and throwing off these !
i
misconceptions about its object of study. In a sense, j
however, these were productive misconceptions. Phelps !
I
lists some of the "recuperative" changes which came
i
I
I
291 i
about through the shift to process as a focus for
writing researchers and teachers:
1) The process shift implicitly downplayed
notions of language as system in favor of
language as event.
2) The focus on event introduced a
temporal/historical dimension into
composition.
3) Process restored the concrete subject
to written language and thus the
possibilities of agency, intention,
purpose, felt experience, and effective
action in a world.
4) Process reconnected thought to language
and prepared the way for overcoming other
oppositions.
5) Process initiated empirical j
investigations that challenged assumptions
and maxims of the tradition on the grounds
of their irrelevance to the actual
experience of individuals.
6) The process shift was a catalyst for
breaking down the intellectual/political
bondage of composition to literature
(Phelps Human Science 44).
!
It is interesting to note, however, that points one I
through four are not true in reference to cognitive
research in the Flower and Hayes style. In fact, the
assumptions of cognitive science lead in quite the
opposite direction. Thus, if the above list is an
accurate representation of the dominant currents in the
field, one can argue that many of the inconsistencies i
!
noted in the work of Flower and Hayes in chapters three j
t
and four derive from the fact that they are trying to be
cognitivists and compositionists too. Phelps observes
Initially, at least, and I think
characteristically as a mission-oriented
292
field, composition relied heavily on
multidisciplinary sources to build up a
broad if fragmentary knowledge base about
these matters [writing, reading, language,
thought, social-interaction], always
scrambling to keep up with its urgent
practical needs. In the process it opened
the door to the conflict of
interpretations. For composition has
incorporated from these disciplines not
only facts and theories, but also
perspectives on human life that are
fundamentally different from one another,
and in some cases apparently incompatible
(Human Science 184) .
Phelps also notes that while nonpositivist methods have
gained acceptance and broad philosophical support,
Composition, like other human sciences, has always been
fascinated by science, envious of the accomplishments of
the natural sciences, desirous of verifiable knowledge,
respectful of quantitative studies, and "regretful at
the possibility that it might not fit this mold" (Human j
Science 188).
Phelps calls the split between positivist methods
and subjective methods "a chasm that will not be leaped,
a wound at the heart of the young discipline that will
not heal."
Each methodology stakes its claim
imperialistically and with a fine contempt
for any other way of knowing. Indeed, in
the extreme, positivist science does not
even grant the status of knowledge to
experiential forms of understanding, while
literary relativists bring into doubt not
only verifiable knowledge but the very
possibility of human knowledge, rational
inquiry, and communication (Human Science
i
1
293 j
_ . - - - t
188-89).
Phelps' project in this chapter of her book, however, is
to find a way, through Ricoeur's hermeneutics, to bridge
this gap, and to argue that both ways of knowing are
necessary. I will return to her argument a bit later.
At this point I want to examine two research methods,
each related to protocol analysis, in the context of
this conflict.
Reflective Methodologies
In Writer's Block: The Cognitive Dimension. Mike
Rose articulates a research program which is cognitive
in nature and indeed based largely on the model
developed by Flower and Hayes. However, he is
dissatisfied with its mechanical rigidity.
Perhaps these very important operations of
"recursiveness" and "priority interrupts"
could be accounted for in a less
mechanical way— the mechanical orderliness
of Flower and Hayes' rendering possibly
being rooted in the hierarchical model of
Miller and his associates, and too, in
Flower and Hayes' method of gathering data
(having writers speak aloud while
composing). It is conceivable that when a
writer speaks as he writes, he articulates
a more ordered flow of thought than would
naturally occur (Rose Writer's Block 9).
Rose goes on to suggest that "opportunism," a concept
developed by Barbara and Frederick Hayes-Roth, better
represents the recursiveness and priority interrupts
294
described by Flower and Hayes. Opportunism, as he
presents it, is a view in which the planner's decisions
and observations suggest numerous opportunities for
development. Subseguent decisions follow up on selected
opportunities. Sometimes the decision sequences appear
to be orderly, but others are less orderly, even
chaotic. Rose says
Applied to writing, opportunism suggests
that the goals, plans, discourse frames,
and information that emerge as a writer
confronts a task are not always
hierarchically sequenced from most general
strategy to most specific activity ....
While editing a paragraph, a writer may
see that material can be organized in a
different way or as a writer writes a
certain phrase, it could cue other
information stored in memory. This
fundamental reciprocity between intent and
discovery, goal orientation and goal
modification is anecdotally documented by
professional writers, and the notion of
opportunism provides a cognitive science
operation to account for it (Rose Writer1s
Block 9).
Rose is looking for a way to use scientific methods and
descriptions while preserving the felt writerly sense
that key decisions are being made on the fly, or, as he
notes James Britton says, shaped "at the point of
utterance" (Writer's Block 10). !
In her discussion of the "paradoxes of coherence" I
i
Louise Phelps makes a distinction between "flow" and
"design." She defines "flow" as "the sense of coherence j
j
295 j
as energy or process, . . . the felt power to integrate
that fills readers as they move through a text,
successfully making sense of it." "Design," on the
other hand, is a text’s "fully realized and relatively
fixed coherence as a meaning object" (Human Science 171-
72). The design of a text becomes stable only in
retrospect, and even then is subject to
reinterpretation. In this section she is discussing the
reading of a text, not writing, and she notes that it is
only the first reading of a text that is experienced as
mostly flow. In this sense the writing of a text and
the reading of it are complementary processes. Phelps
argues
The power to integrate progressively is
very much tied to time, in two senses.
The process takes place in real time,
which subjects readers to a number of
cognitive and contextual limitations (for
example, fatigue and interruption). The
reader also participates in a virtual
discourse time that (ideally) carries him
or her forward on a moving point of
focused attention, alternately looking
backward to the text's "past" and ahead to
its "future." This virtual time is often
inscribed in the discourse through such
phrases as "earlier," "soon," or "for the
moment," referring to the discourse as
action (Human Science 172).
(To be absolutely clear we should say that there are
three modes of time here: empirical time or clock time,
subjective time, and discourse time.) Phelps argues
that in the experience of flow offered by a first
296
reading of something, attention is mostly local and the
reader is preoccupied with direction, order, linearity,
and connections between consecutive items. This is the
perspective which Rose wishes to preserve on the
writer's side with the concept of opportunism.
Protocols exhibit this ongoing experience of the shaping
at the edge of the utterance, so it is in a sense
curious that Rose rejects this methodology. Rose
wonders if "when a writer speaks as he writes, he
articulates a more ordered flow of thought than would
naturally occur" (Rose Writer's Block 9), but as we have
seen in the protocol provided by Emig, the movement of
attention from the point of utterance, to previous text,
to future goals and meanings, and back to utterance, is
recorded. In actuality it is the fact that the Flower
and Hayes model oversimplifies this movement that
troubles Rose. Rose is getting at this problem when he
says that the
Continual distinction between top-down and
bottom-up behavior— at least as far as
composing is concerned— is misleading.
Even superficial examination of writers at
work reveals the enactment, even the
transaction, of both orientations. To
posit one or the other as being the norm
(or both as being the only possibilities)
is to reduce the complexity of composing.
Thus the notion of opportunism— with its
emphasis on shifting between top-down and
bottom-up behavior and shifting, as well,
"horizontally" among executive operations,
composing subprocesses, dimensions of
297
knowledge, and elements of the task
environment— seems much truer to the way
writers write" (Rose Writer's Block 11).
Design, as opposed to flow, is "characteristically
timeless, independent both of the virtual time in which
the thought process and action of a discourse evolve,
and of the real time lived by the reader and writer"
(Human Science 172-73). (Again, clock time and
subjective time are being collapsed here into "real"
time.) This is close to what Flower and Hayes mean by a
"representation" of the meaning. Flow is the process
view of coherence; design is the static timeless view.
Ironically, in attempting to build a model of the
process, Flower and Hayes obscure its nature as a
process in time. Phelps clearly gains these
distinctions and insights through bracketing different
aspects of time and performing a phenomenological
analysis, as Ricoeur does in Freedom and Nature (see
Chapter Two, p. xx, of this dissertation). As I noted
in chapter two, the only way to analyze a process is to
bracket time.
Rose's model is based on that of Flower and Hayes,
and shares many of that model's basic features:
Domain Knowledge: facts and propositions about
myriad topics stored in long-term memory, some of
it in non-linguistic fashion.
298
Composing Sub-processes: "Linguistic, stylistic,
rhetorical, sociolinguistic, and process rules,
interpretive as well as intersentence to discourse-
level writing plans, discourse frames and
attitudes, all of which select, shape, organize,
and evaluate domain knowledge." Rules are either
"flexible and multi-optional or one-directional,
rigid, inflexible."
Executive Operations; "High-level, often
assumption-based, strategies" which "select,
organize, and activate composing sub-processes."
This is equivalent to Flower and Hayes' monitor.
Task Environment: "Includes a particular writing
project, all attendant materials, and words-on-the-
page the writer has converted from thought to
written language" (Rose Writer's Block 10).
These definitions are not too different from those of
Flower and Hayes, except in the category of "Composing
Sub-processes." Here the types of processes which
exist, the kinds of rules which operate, and the types
of knowledge accessed are much broader and more various.
The concept of flexible and inflexible rules is alien to J
the Flower and Hayes model, or at least to its j
underpinnings in Artificial Intelligence. Yet in a
common sense way, we know that Rose must be right; j
299
writers must access this knowledge and do these things.
It is easy to see why Rose was attracted to a
functional cognitive model. His object of investigation
is writer's block. It is logical to assume that
writer's block may be the result of a dysfunctional
writing process. If we can divide the writing process
into separate functions, we may be able to determine
which part of the process is not working and do
something to remedy it. It is also possible that the
study will work in the other direction and help refine
the model itself. The study of the normal through the
abnormal, or the functional through the dysfunctional
has a long tradition in psychology. Emig herself
recommends studies of people with specific and
generalized disabilities, such as the blind, the deaf,
and the brain-damaged, in order to establish what the
essential components and structures of the composing
process are, although she recognizes that such research
has been criticized fWeb of Meaning 111). (Ricoeur
specifically rejects this sort of approach, and
criticizes psychology for using it.)
Rose's chosen methodology for his investigation is
"stimulated recall," which he describes as
A decision-making, problem-solving
research procedure pioneered by Benjamin
Bloom. During the procedure, an event is
audio-taped and then played back to the
300
participant(s) soon after the completion
of the event. The key assumption is that
the replay will stimulate recall of mental
processes occurring during the event in
question" (Rose Writer's Block 23).
In Rose's procedure the writer is videotaped by two
cameras, one focused on the page, and one on the writer.
Rose wishes to avoid what is commonly seen as the main
problem of protocol analysis— the fact that the
experimental procedure may influence the thought
processes it is designed to study. Of course, as we
have seen in chapter three, Eriksson and Simon believe
that the problems with retrospective accounts are more
serious than those encountered in concurrent accounts.
As Rose is well aware, the presence of the videotape
equipment also may influence the results by making the
scene less than natural. One could argue that
stimulated recall, from an empirical point of view,
introduces more variables than protocol analysis and is
therefore a less desireable method.
However, from a phenomenological point of view
there is an important and interesting difference between
protocol analysis and Rose's method. Think-aloud
protocols do not usually involve reflection. Stimulated
recall does. Alfred Schutz says
Meaning, as has been shown elsewhere, is
not a quality inherent in certain
experiences emerging within our stream of
consciousness but the result of an
interpretation of a past experience looked
at from the present Now with a reflective
attitude. As long as I live in my acts,
directed toward the objects of these acts,
the acts do not have any meaning. They
become meaningful if I grasp them as
well-circumscribed experiences of the past
and, therefore, in retrospection. Only
experiences which can be questioned about
their constitution are, therefore,
subjectively meaningful (Schutz The
Problem of Social Reality 210-11).
Schutz goes on to speculate that there may be
spontaneous experiences, physiological reflexes and the
like, which cannot be grasped by reflection and are not,
therefore, subjectively meaningful. The temporal
structure of subjective meaning does not call into
question the validity of think-aloud protocols; it
simply makes them more "objective." Stimulated recall,
on the other hand, is more "meaningful," because we get
the subject's interpretation of the meaning of his or
her own actions. To distrust or disregard this type of
information is to exhibit a prejudice that is truly
inappropriate to a discipline like composition, or even,
as Jerome Bruner argues, to psychology.
Much of the distrust of subjectivism in
our explanatory concepts has to do, I
think, with the alleged discrepancy
between what people say and what they
actually do. A culturally sensitive
psychology . . . is and must be based not
only upon what people actually do, but
what they say they do and what they sav
caused them to do what they did. It is
also concerned with what people sav others
did and why. And above all, it is
302
concerned with what people sav their
worlds are like. Since the rejection of
introspection as a core method of
psychology, we have been taught to treat
such "said" accounts as untrustworthy,
even in some odd philosophical way as
untrue (Bruner Acts of Meaning 16).
Bruner goes on to say that such a distrust of subjective
information "implies that what people do is more
important, more 'real,' than what they sav. or that the
latter is important only for what it can reveal about
the former" fActs of Meaning 16).
In this regard Ricoeur quotes from philosopher Jean
Nabert's article in the Encvclooedie francaise:
It is therefore true that in all of the
domains where spirit reveals itself as
creative, reflection is called on to
retrieve the acts which works conceal,
because, living their own life, these
works are almost detached from the
operations that have produced them. It is
a question of bringing to light the
intimate relationship between an act and
the significations in which it is
objectified (Ricoeur Cl 218-19).
This is the beginning of Ricoeur's turn from
phenomenology to hermeneutics, or the beginning of his
attempt to graft the one onto the other. Patrick
Bourgeois argues that Ricoeur learned from Nabert to
link "deeds which reflect our effort and the notion of
sign needing interpretation."
In his third Terrv Lecture [see Cl 211-
222] the link becomes the link between
reflection and interpretation which is
worked out with the concrete examples of
303
decision. This link is what goes beyond
the method of Le volontaire et
11involontaire [published in English as
Freedom and Nature1 or sees what it
implies. For now Ricoeur understands this
reflection as interpretation, i.e., he
understands a link between reflection and
interpretation of the meaning of willing
in the phase of deciding with motives . .
. now Ricoeur sees the motive as the sign
of the intention of willing” (Bourgeois
Extension 100-01).
Thus it is clear that if Rose points to a certain point
in the writing of a text and says something like, "You
paused at this point, what were you thinking about?", he
is asking the subject to do more than simply recall. He
is asking the writer to engage in reflection, which is
by its nature both a semiotic and a hermeneutic process.
He is asking the writer to interpret his actions, to
represent his motives, intentions and purposes in
semiotic code. This does not make the data suspect; it
simply makes it different, and more suitable than think-
aloud protocol data for certain purposes.
According to Ricoeur, it is in fact impossible for
an individual to understand himself or his actions,
except in signs.
I stand fast with Nabert in saying that
understanding is inseparable from
self-understanding and that the symbolic
universe is the milieu of
self-explanation. This means that there
is no longer a problem of meaning unless
signs are the means, the milieu, and the
medium thanks to which a human existent
seeks to situate, project, and understand
304
himself. In contrast, however, there is
no direct apprehension of the self by the
self, no internal apperception or
appropriation of the self's desire to
exist through the short cut of
consciousness but only by the long road of
the interpretation of signs. In short, my
philosophical working hypothesis is
concrete reflection, i.e., the coqito as
mediated by the whole universe of signs
(Ricoeur Cl 170, see also Cl 264).
Of course, even a think-aloud protocol consists of
signs. In a concurrent verbalization, however, the mode
is not reflective, and in a strange sense, the signs
expressed by the subject are not being interpreted by
the subject as actions with motives and intentions, but
are expressive signs, in Husserl's sense, (see chapter
one, page xx) which have been verbalized, passing beyond
the boundary of consciousness only through the
artificial distortion of the investigative procedure.
This, at least, is the ideal; as we have seen, in Emig's
data the protocol discourse is at times rhetorically
addressed.
Rose quite rightly argues that the key issue with
composing process self-reports is "the availability of a
particular composing act to personal observation"
(Writer's Block 18). For the first stage of his study,
Rose used a questionnaire, tested and modified through
six versions, to categorize his subjects as "high-
blockers" or "low-blockers." This questionnaire asked
305
respondents to report behaviors or experiences
determined through pilot interviews as likely to be
relevant to writer's block. Rose argues that this data
is accessible to questionnaire inquiry, because
"reporting on them does not necessarily involve
exploration of covert mental processes or subtle
influences (such as the sound of the language) on those
processes" (18). Although this is a valid argument, the
use of two different methodologies as different elements
in the same study, both attitude survey and stimulated
recall, complicates the nature of the data considerably.
The survey allows Rose to divide the subjects into two
populations, which can then be compared to see if
differences are apparent in the stimulated recall
sessions. The validity of the study would have been
easier to evaluate if the distinctions had been made
solely through the analysis of the stimulated recall
data. However, Rose's procedures work for his purposes.
His questionnaire meets the proper statistical criteria,
and high-blockers and low-blockers exhibit significant
differences in the stimulated-recall protocols.
A similar reflective method is the "discourse-based
interview" used by Lee Odell, Dixie Goswami, and Ann
Herrington in their study of workplace writers. In this
procedure, writing is collected from a workplace
306
environment, and the researchers identify key rhetorical
decisions in the texts. In a subsequent meeting with
the writer of a particular text, possible changes in
wording are suggested for these key locations. The
writers are asked whether these changes are acceptable,
and to explain why or why not. The purpose of these
interviews is "to probe a worker's store of knowledge of
the rhetorical context for writing done on the job"
("The Discourse-Based Interview" 227) . The authors note
that while Flower and Hayes are interested in studying
writers' underlying composing processes, and so have
made use of experimental tasks "that are designed to
pose unique and unexpected demands for a writer," Odell
and company wish to study writing in familiar, natural
contexts, and so choose a different methodology. The
discourse-based interview concerns actual writing with a
real purpose, situated in a real world with all of the
social connections intact.
Odell, Goswami, and Herrington discuss the
reliability of interview data and argue that interviews
are not reliable for investigating composing processes,
because when information is transferred from short term
memory to long term memory it is simplified and
restructured. "However, we are not using interviews to
obtain information about mental processes. We are using
307
interviews to identify the kinds of world knowledge and
expectations that informants bring to writing tasks and
to discover the perceptions informants have about the
conceptual demands that functional, interactive writing
tasks make on them" (Odell et al. "The Discourse-Based
Interview" 228). In fact, the authors find that
composing-aloud protocols often omit information about
decisions that seem important during the interviews
(234) . This amounts to a subtle, but useful, critique
of Flower and Hayes, although the critical aspect is
hidden well.
Both Rose and Odell, Goswami, and Herrington
consider the Flower and Hayes talk-aloud protocols to be
more "objective" than the data gathered through their
own reflective methods. However, both rightly feel that
talk-aloud protocols are inappropriate to their research
goals. This is an important point. Reflective
methodologies should not be rejected simply because they
are reflective, or because they require the subject to
interpret his or her own actions and words.
I discussed the theoretical grounding of such
methodologies in chapter one (see page 31-32??). George
Psathas argues that Ethnomethodology assumes "that there
is a world of everyday life known by members in and
through the practices they themselves use for revealing,
308
constructing, and describing it" (Psathas 87). It is
this world and these practices that Rose, Odell,
Goswami, and Herrington are investigating, and a
reflective methodology is the best way to do this.
Composition is about communication and social action.
Cognitive processes and semiotic systems may be relevant
to Composition as a field, but they are not at its
center.
From Reflection to Critical Consciousness
Louise Phelps makes clear the essential role
reflection plays in composition studies. She quotes
Dewey's definition of reflection: 1 1 Active, persistent,
and careful consideration of anv belief or supposed form
of knowledge in the light of the grounds that support it
and the further conclusions to which it tends." but
notes that while Dewey conceives of reflection as a
mental act, she is more concerned with the social nature
of reflection and the formative role of discourse and
textual symbolization in facilitating deliberative
consideration of subjects. Reflection, as Ricoeur
argues, brings into being the subject-object split, and
establishes an effortful relationship of knowing between
human and world. Reflection in this sense is associated
with criticism and a concept of the negative, what
309
Ricoeur calls the "dubitative word." Phelps argues that
postmodernism neglects the positive/generative aspects
of reflection, intuition and insight (Phelps Composition
as a Human Science 68-69).
The critical power of reflection is an important
theme in Phelps’ book. (See Edlund "Review of
Composition as a Human Science" 191-95). In the chapter
entitled, "Literacy and the Limits of the Natural
Attitude," Phelps attempts to reconnect some of the
major opposing camps in composition. In doing so she
borrows a number of concepts from phenomenology, among
them the concept of the life-world, "the domain of
common sense, everyday experience into which we are all
born and live prereflectively, before science or
philosophy." This is the realm of Ricoeur's "natural
attitude," a pre-critical non-reflective consciousness
that accepts the world as it is given, and language and
culture as "natural." Talk-aloud protocols tend to
reveal this world, while reflective methods encourage
analysis and a critical attitude.
Phelps articulates a "myth of natural literacy"
which is relevant to the domain of the natural attitude,
though it is important to note here that "myth" is not
used in a negative or pejorative sense. She identifies
three major variations on the theme of natural literacy
310
in composition, each with different theoretical roots,
research programs and pedagogies: the biological, the
romantic, and the contextual.
The biological orientation is organismic. The
human being is seen as an organism adapted by evolution
to an environment that includes language. The
underlying metaphor is the genetic code, and the
predominant mode of inquiry is scientific. Chomsky and
Piaget are offered as examples of biologically oriented
theorists who have been adapted to the purposes of
literacy study. Stephen Krashen is another theorist
whose orientation is basically organismic. The romantic
(or vitalist) perspective, like the biological,
emphasizes the potentialities of the individual, but in
this view there is a sense of mystery about the creative
process and the imaginative and symbolizing capability.
For this reason the romantic tends to reject scientific
inquiry in favor of more philosophical approaches. The
contextualist, on the other hand, focuses on social
interaction rather than on the individual.
Both the biological and the Romantic
approaches focus on the individual and
generate extremely child-centered ideas of
literacy learning, almost to the exclusion
of direct cultural instruction. In
contrast, the newly vigorous contextualist
or ecological orientation to literacy
recognizes that the learner lives in a
cultural and specifically linguistic
world, and thus highlights the
311 i
— 1
interpersonal dimensions of natural
literacy learning .... Children appear
as active agents of their own learning,
but the contextualist emphasizes not
constructing codes but participating in
literacy events (114).
All three of these seemingly divergent perspectives are
immersed in the natural attitude. Together they assume
that literacy is acquired from the environment through
the natural potential of the human organism, though each
perspective has its own emphasis. The natural attitude
is, however, a naive state of being:
In terms of the myth, each person— writer
and reader, teacher and learner— acts
naively, that is, within the natural
attitude, which Ricoeur expresses as "lost
and forgotten in the world, lost in the
things, lost in the ideas, lost in the
plants and animals, lost in others, lost
in mathematics .... What I call living
is hiding myself as naive consciousness
within the existence of all things" (115).
The project of phenomenology has been to get beyond
this naivete through a "critical moment" which leads to
an enlarged understanding, a "critical consciousness."
Phelps points out that many scholars claim that literacy
is an agent for "disembedding consciousness from the
life-world context, distancing it from an immediacy
often pictured as restrictive and illusory, and
empowering critical reflection and subtle thought"
(111). How then can literacy be both naturally acquired
and an agent for critical consciousness? This
312
contradiction leads Phelps to ask two important
questions:
First, in what sense, if at all, does the
practice of literacy in my culture exceed
the natural or naive participation in
language and become self-reflective and
critical? And second, insofar as it does,
how should cultural instruction respond so
as to overcome the naivete of a natural
pedagogy without losing its depth? (112)
These questions, framed in a contextualist manner, are
at the heart of what composition as a discipline is
about. Phelps' answer is threefold: there is a
progression from natural literacy, rooted in a
biological base and contextual motivations, to craft
literacy, where "mastery involves attention to form and
technique as instrumental, and then renaturalization of
these as tacit" (126), to critical consciousness in
Paulo Freire's sense. Thus in one sense literacy
acquisition is as natural and inevitable a response to
the cultural environment as language acquisition itself;
in another it is an instrument through which action is
accomplished and cultural belonging is established; in
yet another it is the means by which culture is both
recognized and called into question. From another
perspective, however, literacy is all these things at
once. Phelps writes, "After analysis, the challenge
remains: how— in a spirit of hope, if not of innocence—
to find, or perhaps merely to seek again and again, that
313
fragile and fleeting balance whereby we can surpass the
natural attitude and yet reclaim it?" (127)
The Contextualist Perspective
For Louise Phelps, the contextualist perspective
encompasses many other perspectives, and thus solves
many of the inherent contradictions of narrower views.
She notes that discourse is a common metaphor in
contextualist psychological theories, which model
reality on the communicative relation. She argues that
the Latin root of the word "discourse," meaning "to run
to and fro," "expresses the intuition that this relation
is primarily diachronic or in motion, projecting the
temporality of utterance even on system" (50).
According to Phelps, on this view, objects, events,
entities, and structures can be described as:
1. Semiotic: organized as sign/meaning
relationships; mediated, triadic in
Peirce's sense.
2. Transactive: deriving any autonomous or
stable units from a holomovement, the
parts constantly influencing, defining,
and modifying one another. 3. Holistic:
functioning as a system in which all
elements are interconnected and
interdependent.
4. Dialogic: developing conversational
relations such as question/answer,
statement/qualification,
hypothesis/implication, and so on.
5. Dialectic: Characterized by conflict,
opposition, competition. A further
specification of dialogue: the two are
often equated. Both dialogue and
314
dialectic are frequently broadened from
the face-to-face dyad, which produces the
vocabulary and images of duality, to the
metaphors of conversation, as in Burke's
cocktail party or Bakhtin's multivocal
dialogue.
Phelps notes that these qualities are often summarized
as "rhetorical" or "semiotic" depending on whether the
emphasis is interpersonal or systemic. From this point
of view, even a scientist's relation with nature can be
characterized as a type of dialogue or interaction
(50-51).
Phelps argues that contextualism supplies the basis
for treating discourse and the life-world as
"intertwined realities." She notes that contextualist
theories tend to be realist in that they acknowledge
"the facticity of phenomena in a world presenting itself
to human consciousness as given," but that phenomena are
presented "not only materially but as a web of
signification." This common-sense realism may appear at
first "to be at odds with contextualist emphasis on
transience, illusion, indeterminacy, and difference."
However, Phelps argues that contextualist theories "tend
to multiply realities rather than deny them." In
addition, contextualist theories "do not perpetuate the
subject-object distinction that poses the problem of
realism to begin with. Like Heideggerian phenomenology,
they simply begin with the fact of the
315
person-in-the-environment" (54-55).
Clearly, the contextualist view opens up so many
elements and relationships to investigation that it is
impossible to proceed with any study without careful
bracketing. Phelps argues that choices must be made.
For example, Phelps argues that because composition is a
praxis, it ought to adopt the common sense view of the
reality of texts and communication outlined above.
Because not all contextualist theories
profess a frank everyday realism, such
realism represents for composition a
choice— not merely intellectual but moral
in nature. As experience, discourse is
real; so are thought and the world to
which discourse refers and addresses
itself. Because composition is a praxis
and studies a praxis, it is concerned with
human life as concretely lived by
individuals in perceptual, social,
economic, and political worlds where
literacy (or illiteracy) and texts have
both vivid consequences and explanatory
contexts. Written discourse suffuses
cultural experience in a literate society
and requires that we understand it in
relation not only to abstract thought,
personal development, learning, and
literary art, but also to work and power
(55) .
Phelps notes that most disciplines in the humanities
display "an underlying uncertainty about the existence
of their objects of study'1 especially with respect to
texts, and that this uncertainty "shakes the foundations
of their disciplines" (55).
316
Ricoeur and the Problem of Reference
As noted above, Phelps considers the difference between
the "rhetorical" and the "semiotic" to be a matter of
perspective or emphasis. This move is based in part on
Ricoeur1s incorporation of semiotics into his own
phenomenological approach. Ricoeur argues that lanaue
has sense, but not reference.
It is in discourse that language has a
reference. To speak is to say something
about something. . . . Frege showed
precisely that the aim of language is
double: the aim of an ideal sense (that
is, not belonging to the physical or
psychic world) and an aim of reference: if
the sense can be called inexistent,
insofar as it is a pure object of thought,
it is the reference— Bedeutuna— which
roots our words and sentences in reality
(Philosophy of Paul Ricoeur 114).
Thus when we speak or write we articulate "a signifying
intention that breaks the closure of the signs." When
Fodor says, as I noted in chapter three, that cognitive
science may lack any means to go beyond the study of
syntactic processing to the study of semantic
properties, it is this leap that he is referring to.
Yet the study of language is not whole without both
sides.
Ricoeur clearly rejects this split, or this
"choice," as he calls it..
The moment when the turning from the
ideality of sense to the reality of things
is produced, is the moment of the
317
transcendence of the sign. This moment is
contemporaneous with the sentence. It is
on the level of the sentence that language
says something; short of it, it says
nothing at all. ... It is not
necessary, therefore, to oppose two
definitions of the sign, the one as
internal difference of the signifying to
the signified, the other as external
reference of sign to thing. There is no
need to choose between these two
definitions. One relates to the structure
of the sign in the system, the other to
its function in the sentence (Philosophy
of Paul Ricoeur 115).
Ricoeur recommends that we seek new models of
intelligibility that make possible the synthesis of
structure and history, both synchronic and diachronic
perspectives, lanae and parole. As a first move toward
the development of such "instruments of thought,"
Ricoeur characterizes the phenomenon of language as
"neither structure nor event, but the incessant
conversion of one into the other in discourse"
(Philosophy of Paul Ricoeur 116).
In order to understand the importance of this
redefinition of what language is, it is perhaps
necessary to delve back into what the practice of
phenomenology is like. In chapter one I discussed Don
Ihde's "hermeneutic" rules, which were:
1. Attend to the phenomena of experience as they
appear.
2. Describe, don't explain.
318
3. Horizontalize or equalize all immediate
phenomena (don't attempt to judge what is "real" or
"more real").
(Ihde Exoerimenta1 34-36)
The purpose of these directives is to enable the
observer to eliminate pre-conceived notions,
presuppositions and prejudices which arise from
sedimented knowledge and cultural biases. I would like
to apply them briefly to the phenomenon of language use.
As I write this text, I am consciously aware of a
large number of considerations which influence the
choice of words and forms at the "edge" of my utterance.
I am aware of previous selections of words and
structures, in varying degrees of clarity and obscurity,
the previous word, previous phrases, previous sentences,
and choices made in earlier paragraphs and chapters. I
am aware of choices I could have made, but did not.
Some of these choices I doubt, and they nag at me to
return and choose again. If I so return, there is a new
"edge" to my utterance, a new present. I am aware of a
future of my text, possible choices, possible
destinations, possible goals and effects. I am aware of
the rhythms and sounds of my words, their consonances
and assonances, but I choose my words, not only for
their sounds, but also for their significance. I am
319
aware of references, things in the world, things in the
past, words in books, inferences, ideas,
interpretations. I am aware of audiences, the members
of my committee, my colleagues, my students. I am aware
of perspectives, positions, arguments, evidence. I am
aware of my physical surroundings, the noise of my
computer, the flashing of the cursor, the sense of my
body. All this, and more, is held in the mind at once.
If we begin our investigations with this moment at
the edge of the utterance, we cannot say that it is the
system that is more real, or more essential, than the
event. We cannot choose in this way, because all is
present together. We can bracket parts of the
experience, because there is too much to understand at
once, but we cannot prioritize. When, as is often the
case for my students, a writer must write in an
unfamiliar language, when the problem is with the
semiotic system, beginning to write is difficult.
However, if I am writing something new, for an audience
I do not know (as recently happened when I had to write
a press release) I am also uncomfortable, and I do not
know how to begin. We can argue that signs are
meaningless outside of syntactic structures, but they
are also meaningless outside of social structures,
outside of events. Ricoeur argues
320
The word is much more and much less that
the sentence. It is much less because
there is not yet any word before the
sentence. What is there before the
sentence? Signs, that is, differences in
the system, values in the lexicon. But
there is not yet any meaning, any semantic
entity. Insofar as it is a difference in
the system, the sign says nothing . . .
Thus, the word is, as it were, a trader
between the system and the act, between
the structure and the event .... But it
is more than the sentence from another
point of view. The sentence, we have
seen, is an event: as such, its actuality
is transitory, passing, vanishing. But
the word survives the sentence (Philosophy
of Paul Ricoeur 118-119).
Think-aloud protocols, and the related reflective
methodologies discussed in this chapter, are both ways
of investigating what is happening at the edge of the
utterance, at the conjunction of system and event. In
my view, this is the proper locus for composition
research.
It has not been my intention in this dissertation
to disparage empirical research orientations, or to
argue that phenomenology is the only path to knowledge.
I do believe, however, that only phenomenological
insights and methods can dissolve the gap between
objective data and subjective experience, and between
system and event. I hope that this study is useful in
beginning an integration and evaluation of the various
kinds of knowledge that have been developed through a
variety of methods over the past twenty years in
321
composition research.
322
BIBLIOGRAPHY
Augustine, Dorothy. "Geometries and Words: Linguistics
and Philosophy: A Model of the Composing Process."
CE (1981): 221-231.
Barritt, Loren S. and Barry M. Kroll. "Some
Implications of Cognitive-Developmental Psychology
for Research in Composing." In Research on
Composing: Points of Departure. Ed. Charles R.
Cooper and Lee Odell. Urbana: NCTE, 1978. 49-57.
Bartholomae, David. "Inventing the University." In
When a Writer Can't Write. Ed. Mike Rose. New York:
Guilford Press, 1985. 134-65.
Beach, Richard, and Lillian S. Bridwell, eds. New
Directions in Composition Research. New York:
Guilford, 1984.
Bereiter, Carl and Marlene Scardamalia. The Psychology
of Written Composition. Hillsdale, NJ: Lawrence
Erlbaum Associates, 1987.
. "Levels of Inquiry in Writing Research."
Mosenthal, Research on Writing 3-25.
Berger, Peter L. and Thomas Luckman. The Social
Construction of Reality. New York: Doubleday,
1966.
Berlin, James A. "Contemporary Composition: The Major
Pedagogical Theories." CE 44:8 (1982): 765-77.
Writing Instruction in Nineteenth-Centurv American
Colleges. Carbondale: Southern 111. Univ. Press,
1984.
Rhetoric and Reality: Writing Instruction in
American Colleges. 1900-1985. Carbondale: Southern
111. Univ. Press, 1987.
"Rhetoric and Ideology in the Writing Class." CE
50:5 (1988): 477-94.
Berthoff, Ann E. Forming/Thinking/Writing. New Jersey:
Boynton/Cook, 1982.
323
Bizzell, Patricia. "Cognition, Convention, and
Certainty." Pre/Text 3:3 (1982): 213-43.
. "College Composition: Initiation into the Academic
Discourse Community." Curriculum Inquiry 12
(1982): 191-207.
"Foundationalism and Anti-Foundationalism in
Composition Studies." Pre/Text 7:1-2 (1986):
37-56.
Bourgeois, Patrick L. Extension of Ricoeur1s
Hermeneutic. The Hague: Martinus Nijhoff, 1975,
Bracewell, R.J., C.H. Frederiksen, and J.D. Frederiksen.
"Cognitive Processes in Composing and Comprehending
Discourse." Educational Psychologist 17 (1982):
146-64.
Britton, James. "The Composing Processes and the
Functions of Writing." Cooper and Odell, eds.
Research on Composing 13-29.
Britton, James, Tony Burgess, Nancy Martin, Alex McLeod,
and Harold Rosen. The Development of Writing
Abilities fll-18). London: Macmillan Education,
1975.
Bruffee, Kenneth A. "Collaborative Learning and the
Conversation of Mankind." CE 46 (1984): 635-52.
Bruffee, Kenneth A. "Social Construction, Language, and
the Authority of Knowledge: A Bibliographic Essay."
CE 48 (1986): 773-90.
Bruner, Jerome. Acts of Meaning. Cambridge, Mass.:
Harvard Univ. Press, 1990.
Burke, Kenneth. A Grammar of Motives. 1945. Berkley:
Univ. of Calif. Press, 1969.
Carter, Michael. "Problem Solving Reconsidered: A
Pluralistic Theory of Problems." CE 50 (1988) 551-
565.
. "The Idea of Expertise: An Exploration of
Cognitive and Social Dimensions of Writing." CCC
41 (1990): 265-286.
Charme, Stuart L. "Paul Ricoeur as Teacher: A
324
Reminiscence." Pre/Text 4 (1983): 289-294.
Cooper, Charles R. and Lee Odell, eds. Research on
Composing: Points of Departure. Urbana: NCTE,
1978.
Cooper, Charles and Sidney Greenbaum, eds. Studying
Writing: Linguistic Approaches. A Written
Communication Annual. Vol. 1. Beverly Hills: Sage,
1986.
Cooper, Marilyn M. "The Ecology of Writing." CE 48
(1986): 364-375.
. "The Pragmatics of Form: How do Writers Discover
What to Do When?" In Beach and Bridwell, eds. New
Directions in Composition Research.
Cooper, Marilyn and Michael Holzman. "Talking about
Protocols." CCC 34 (1983): 284-93.
de Beaugrande, Robert. Text Production: Toward a
Science of Composition. Vol XI, Advances in
Discourse Processes, ed. Roy O. Fleedle. Norwood,
NJ: Ablex, 1984.
. "Psychology and Composition: Past, Present, and
Future." In Nystrand, ed. What Writers Know.
"Psychology and Composition" CCC Feb. 1979:
50-57.
Deloach, Bill. "On First Looking into Ricoeur1s
Interpretation Theory: A Beginners Guide."
Pre/Text 4 (1983): 225-236.
Derrida, Jacques. Speech and Phenomena and Other Essavs
on Husserl1s Theory of Signs. Trans. David B.
Allison. Evanston: Northwestern Univ. Press, 1973.
. Of Grammatology. Trans. Gayatri Spivak. Baltimore:
Johns Hopkins Univ. Press, 1976.
Dobrin, David N. "Protocols Once More." CE 48 (1986):
713-26.
Dreyfus, Herbert L. What Computer1s Can't Do: The Limits
of Artificial Intelligence. Rev. ed. New York:
Harper Colophon, 1979.
325
Dreyfus, Hubert L. and Harrison Hall, eds. Husserl.
Intentionalitv. and Cognitive Science. Cambridge,
Mass.: MIT Press, 1982.
Dreyfus, Hubert L. and Stuart E. Dreyfus. Mind Over
Machine: The Power of Human Intuition and Expertise
in the Era of the Computer. New York: The Free
Press, 1986.
Edie, James M. Edmund Husserl’s Phenomenology: A
Critical Commentary. Bloomington: Indiana Univ.
Press, 1987.
Edlund, John R. "Review of Composition as a Human
Science: Contributions to the Understanding of a
Discipline by Louise W. Phelps." The Writing
Instructor 8 (1989): 191-95.
Embree, Lester E. ed. Life-World and Consciousness:
Essays for Aron Gurwitsch. Evanston: Northwestern
Univ. Press, 1972.
Emig, Janet. The Composing Processes of Twelfth
Graders. Urbana, 111.: NCTE, 1971.
. The Web of Meaning: Essays on Writing. Teaching.
Learning, and Thinking. Upper Montclair, NJ:
Boynton/Cook, 1983.
Engelhart, H. Tristram, Jr. "Husserl and the Mind-Brain
Relation." Ihde, Don, Richard M. Zaner eds.
Interdisciplinary Phenomenology p. 51-70.
Ericsson, K. Anders and Herbert A. Simon. "Verbal
Reports as Data." Psychological Review 87 (1980):
215-251.
. Protocol Analysis: Verbal Reports as Data.
Cambridge, MA: The MIT Press, 1984.
Faigley, Lester. "Competing Theories of Process: A
Critique and a Proposal." CE 48 (1986): 527-542.
Faigley, Lester, Roger Cherry, David Jollifree, and Anna
Skinner. Assessing Writer's Knowledge and
Processes of Composing. Writing Research Series.
Norwood, NJ: Ablex, 1983.
Flower, Linda. "Writer-Based Prose: A Cognitive Basis
for Problems in Writing." CE 41 (1979): 19-37.
326
. Problem Solving Strategies for Writing. New York:
Harcourt Brace Jovanovich, 1981.
. "Revising Writer-Based Prose." Journal of Basic
Writing 3 (1981): 62-74.
. "Response to Anthony R. Petrosky, Review of
Problem-Solving Strategies for Writing, by Linda
Flower. CCC 34 (1983): 233-35." CCC 35 (1984):
96-97.
. "Interpretive Acts: Cognition and the Construction
of Discourse." Poetics 16 (1987): 109-130.
"The Construction of Purpose in Writing and
Reading." CE 50 (1988): 528-50.
"Taking Thought: The Role of Conscious Processing
in the Making of Meaning." In E. Maimon, B.
Nondine, and F. O'Connor eds. Thinking. Reasoning,
and Writing. New York: Longman, 1989.
. "Cognition, Context, and Theory Building." CCC 40
(1989): 282-311.
. "The Role of Task Representation in
Reading-to-Write." Reading-to-Write: Exploring a
Cognitive and Social Process. Linda Flower et al.
New York: Oxford UP, 1990.
Flower, Linda and John R. Hayes. A Process Model of
Composing. Document Design Project Technical Report
No. 1. Pittsburgh: Carnegie-Me11on, 1979.
. "The Dynamics of Composing: Making Plans and
Juggling Constraints." Cognitive Processes in
Writing. Ed. Lee W. Gregg and Erwin R. Steinberg.
Hillsdale, NJ: Lawrence Erlbaum Associates, 1980.
3-50.
. "The Cognition of Discovery: Defining a Rhetorical
Problem." CCC 31 (1980): 21-32.
. "Identifying the Organization of Writing
Processes." Cognitive Processes in Writing. Ed.
Lee W. Gregg and Erwin R. Steinberg. Hillsdale, NJ:
Lawrence Erlbaum Associates, 1980. 3-50.
. "The Pregnant Pause: An Inquiry Into the Nature of
Planning." RTE 15 (1981): 229-243.
327
. "A Cognitive Process Theory of Writing." CCC 32
(1981): 365-87.
"Plans That Guide the Composing Process."
Writing; The Nature. Development, and Teaching of
Written Communication. Vol. 2. Eds. C.H.
Frederiksen and J.F. Dominic. Hillsdale, NJ:
Erlbaum, 1981. 39-58.
. "Images, Plans and Prose: The Representation of
Meaning in Writing." Written Communication 1
(1984): 120-60.
. "Response to Marilyn Cooper and Michael Holzman,
•Talking About Protocols.'" CCC 36 (1985): 94-99.
Flower, L.S., Hayes, J.R. and Swarts, H. Revising
Functional Documents: The Scenario Principle
Technical Report No. 10 Pittsburgh: Carnegie-Mellon
Univ., 1980.
Flower, Linda, L. Carey and J. Hayes. Diagnosis in
Revision: The Expert's Option. Communication
Design Center Tech. Rep. Pittsburgh: Carnegie
Mellon, 1986.
Flower, Linda, John R. Hayes, Linda Carey, Karen
Schriver, and James Stratman. "Detection,
Diagnosis, and the Strategies of Revision." CCC 37
(1986): 16-55.
Flower, Linda, Karen A. Schriver, Linda Carey, Christina
Haas, and John R. Hayes. Planning in Writing; The
Cognition of a Constructive Process. Technical
Report No. 34. Berkeley: Center for the Study of
Writing at University of California, Berkeley and
Carnegie Mellon University, 1989. See also S.
Witte, N. Nakadote, and R. Cherry, eds. A Rhetoric
of Doing. Carbondale: Southern Illinois Univ.
Press, in press.
Flower, Linda, Victoria Stein, John Ackerman, Margaret
J. Krantz, Kathleen McCormick, and Wayne C. Peck.
Reading-to-Write: Exploring a Cognitive and Social
Process. New York: Oxford UP, 1990.
Fodor, Jerry A. The Language of Thought. New York:
Thomas Y. Crowell, 1975.
. Representations: Philosophical Essavs on the
328
Foundations of Cognitive Science. Cambridge,
Mass.: MIT Press, 1981.
. The Modularity of Mind: An Essay on Faculty
Psychology. Cambridge: Mass.: MIT Press, 1983.
. A Theory of Content and Other Essays. Cambridge,
Mass.: MIT Press, 1990.
Fodor, Jerry A., T. G. Bever, M. F. Garrett. The
Psychology of Language: An Introduction to
Psycholinguistics and Generative Grammar. New
York: McGraw-Hill, 1974.
Frederiksen, Carl H. "Cognitive Models and Discourse
Analysis." In Charles Cooper and Sidney Greenbaum,
eds. Studying Writing: Linguistic Approaches
Written Communication Annual. Vol. 1. 1986, 227-67.
Frye, Northrop. Anatomy of Criticism: Four Essays.
1957. Princeton, New Jersey: Princton Univ. Press,
1971.
Gadamer, Hans-Georg. Truth and Method. New York:
Crossroad Pub., 1988.
Gardner, Howard. The Mind's New Science: A History of
the Cognitive Revolution. New York: Basic Books,
1985.
Geertz, Clifford. The Interpretation of Cultures. New
York: Basic Books, 1973.
. Local Knowledge. New York: Basic Books, 1983.
Gregg, Lee and Erwin Steinberg, eds. Cognitive
Processes in Writing: An Interdisciplinary
Approach. Hillsdale, NJ: Lawrence Erlbaum, 1980.
Gurwitsch, Aron. Phenomenology and the Theory of
Science. Ed. by Lester Embree. Evanston:
Northwestern Univ. Press, 1974.
Haas, Christina and Linda Flower. "Reply to Ruth Ray
and Ellen Barton." CCC 40 (1989) 482.
. "Rhetorical Reading Strategies and the Construction
of Meaning." CCC 39 (1988) 167-83.
Hairston, Maxine. "The Winds of Change: Thomas Kuhn and
329
the Revolution in the Teaching of Writing." CCC 33
(1982) 76-88.
Harvey, Charles W. Husserl1s Phenomenology and the
Foundations of Natural Science. Athens: Ohio Univ.
Press, 1989.
Hayes, J.R. Cognitive Psychology: Thinking and
Creating. Homewood, 111.: Dorsey Press, 1978.
Hayes, J.R. and Simon, H.A. "Understanding Written
Problem Instructions." In L.W. Gregg ed. Knowledge
and Cognition. Potomac, MD: Lawrence Erlbaum
Associates, 1974.
Hayes, John R. and Linda S. J. Flower. "Uncovering
Cognitive Processes in Writing: An Introduction to
Protocol Analysis." Mosenthal, Peter, Lynne Tamor,
Sean A. Walmsley ed. Research on Writing:
Principles and Methods. New York: Longman, 1983.
206-219.
. "Writing as Problem Solving." Visible Language
XIV:4 (1980): 388-99.
Hayes, J. R., L. Flower, K. Schriver, J. Stratman, and
L. Carey. "Cognitive Processes in Revision." In S.
Rosenberg ed. Advances in Applied
Psycholinguistics: Reading. Writing, and Language
Processing. Cambridge: Cambridge Univ. Press,
1987.
Heath, Shirley Brice. Wavs With Words. Cambridge:
Cambridge Univ. Press, 1983.
Heelan, Patrick A. "Hermeneutics of Experimental Science
in the Context of the Life-World." Don Ihde and
Richard M. Zaner eds., Interdisciplinary
Phenomenology. 7-50.
Hillocks, George, Jr. "Inquiry and the Composing
Process: Theory and Research." CE 44:7 (1982):
659-73.
Husserl, Edmund. Husserl: Shorter Works. Ed. Peter
McCormick and Frederick Elliston. Notre Dame:
Univ. of Notre Dame Press, 1981.
Husserl, Edmund. Logical Investigations. Vol. I & II.
Trans. J. N. Findlay. New Jersey: Humanities Press
330
Inc., 1970.
Husserl, Edmund. The Crisis of European Sciences and
Transcendental Phenomenology. Trans. David Carr.
Evanston: Northwestern Univ. Press, 1970.
Ihde, Don. Hermeneutic Phenomenology: The Philosophy of
Paul Ricoeur. Evanston: Northwestern Univ. Press,
1971.
. Listening and Voice: A Phenomenology of Sound.
Athens: Ohio University Press, 1976.
. Experimental Phenomenology: an Introduction. New
York: Putnam, 1977.
. Conseguences of Phenomenology. New York: State
University of New York Press, 1986.
Ihde, Don and Richard M. Zaner, eds. Dialogues in
Phenomenology. Selected Studies in Phenomenology
and Existential Philosophy 5. The Hague: Martinus
Nijhoff, 1975.
. Interdisciplinary Phenomenology. Selected Studies
in Phenomenology and Existential Philosophy 6. The
Hague: Martinus Nijhoff, 1977.
Kaufer, David S., John R. Hayes, and Linda Flower.
"Composing Written Sentences." RTE 20:2 (1986):
121-141.
Kearney, Richard. Dialogues with Contemporary
Continental Thinkers: the Phenomenological
Heritage. Manchester: Manchester Univ. Press,
1984.
Kemp, T. Peter and David Rasmussen eds. The Narrative
Path: The Later Works of Paul Ricoeur. Cambridge,
Mass.: MIT Press, 1989.
Kennedy, Mary Lynch. "The Composing Process of College
Students Writing from Sources." Written
Communication 2:4 (1985):434-456.
Klemm, David E. The Hermeneutical Theory of Paul
Ricoeur. Lewisberg: Bucknell Univ. Press, 1983.
Kuhn, Thomas S. The Structure of Scientific
Revolutions. 2nd ed. Chicago: Univ. of Chicago
331
Press, 1980.
Langer, Judith A. "Reading, Writing, and Understanding:
An Analysis of the Construction of Meaning." WC
3:2 (1985): 219-267.
Lanigan, Richard L. "The Phenomenology of Human
Communication." Philosophy Today 23 (1979): 3-15.
Lapointe, Francis H. "Ricoeur and His Critics: A
Bibliographical Essay." In Studies in the
Philosophy of Paul Ricoeur. ed. Charles E. Reagan.
Athens, OH: Ohio Univ. Press, 1979. 163-77.
Lauer, Janice M., J. William Asher. Composition
Research: Empirical Designs. New York: Oxford
Univ. Press, 1988.
Lawlor, Leonard. "Event and Repeatability: Ricoeur and
Derrida in Debate." Pre/Text 4 (1983): 317-334.
LeFevre, Karen Burke. Invention as a Social Act.
Carbondale: Southern 111. Univ. Press, 1987.
Lentricchia, Frank. After the New Criticism. Chicago:
Univ. of Chicago Press, 1980.
Lewontin, R.C. "Fallen Angels." Rev. of Wonderful Life:
The Burgess Shale and the Nature of History by
Stephen Jay Gould. New York Review of Books 14
June, 1990: 3-7.
Matsuhashi, Ann. "Pausing and Planning: The Tempo of
Written Discourse Production." RTE 15:2 (1981):
113-134.
, ed. Writing in Real Time: Modeling Production
Processes. Norwood, New Jersey: Ablex, 1987.
Mohanty, J.N. Edmund Husserl’s Theory of Meaning. The
Hague: Martin Nijhoff, 1976.
Mosenthal, Peter, Lynne Tamor, Sean A. Walmsley ed.
Research on Writing: Principles and Methods. New
York: Longman, 1983.
Murray Donald M. "Response of a Laboratory Rat— or,
Being Protocoled." CCC 34:2 (1983): 169-72.
. "Internal Revision: A Process of Discovery." In
332
Cooper and Odell, eds., Research on Composing.
Natanson, Maurice, ed. Phenomenology and the Social
Sciences. 2 vols. Evanston: Northwestern Univ.
Press, 1973.
. Edmund Husserl: Philosopher of Infinite Tasks.
Evanston: Northwestern Univ. Press, 1973.
. Phenomenology. Role and Reason: Essavs on the
Coherence and Deformation of Social Reality.
Illinois: Charles C. Thomas, 1974.
Ney, James. "Notes Toward a Psycholinguistic Model of
the Writing Process." RTE 8:2 (1974): 157-69.
Nicholson, Graeme. Seeing and Reading. Atlantic
Highlands, NJ: Humanities Press International,
Inc., 1984.
Nisbet, Richard E. and Timothy DeCamp Wilson. "Telling
More Than We Can Know: Verbal Reports on Mental
Processes." Psychological Review 84:3 (1977):
231-59.
North, Stephen M. The Making of Knowledge in
Composition: Portrait of an Emerging Field. Upper
Montclair, NJ: Boynton/Cook, 1987.
Nystrand, Martin, ed. What Writers Know: The Language.
Process, and Structure of Written Discourse. New
York: Academic Press, 1982.
Odell, Lee, Dixie Goswami, and Ann Herrington. "The
Discourse-Based Interview: a Procedure for
Exploring the Tacit Knowledge of Writers in
Nonacademic Settings." Mosenthal, Research on
Writing 220-235.
Ong, Walter J. Oralitv and Literacy: The Technologizing
of the Word. London: Methuen, 1982.
. The Presence of the Word: Some Prolegomena for
Cultural and Religious History. New Haven: Yale
Univ. Press, 1967.
Perl, Sondra. "Composing Processes of Unskilled College
Writers." RTE 13:4 (1979): 317-36.
. "A Look at Basic Writers in the Process of
333
Composing." In N. Kasden and Daniel Hoeber, eds.,
Basic Writing. 1980.
. "Understanding Composing." CCC XXXI:4 (1980):
363-69.
Peterson, Linda. "Repetition and Metaphor in the Early
Stages of Composing." CCC 36:4 (1985):429-443.
Petrosky, Anthony R. "Review of Problem-Solving
Strategies for Writing, by Linda Flower." CCC 34
(1983): 233-35.
Petrosky, Anthony R. and David Bartholomae, eds. The
Teaching of Writing. Eighty-fifth Yearbook of the
National Society for the Study of Education.
Chicago: NSSE, 1986.
Pettigrew, Joan. "Studies of the Composing Process: a
Selective Review." Research in Composition
Newsletter 3:1 (1981): 1-28.
Phelps, Louise Wetherbee. "Possibilities for a
Post-Critical Rhetoric: A Parasitical Preface 6."
Pre/Text 4 (1983): 201-213.
. Composition as a Human Science: Contributions to
the Self-Understanding of a Discipline. New York:
Oxford Univ. Press, 1988.
Pianko, Sharon. "Reflecting: A Critical Component of
the Composing Process." CCC XXX:3 (1979): 275-78.
"A Description of the Composing Processes of
College Freshman Writers." RTE 13:1 (1979): 5-22.
Polanyi, Michael. Personal Knowledge: Towards a
Post-Critical Philosophy. Harper Torchbooks, 1964.
The Tacit Dimension. New York: Anchor Books, 1967.
Psathas, George. "Ethnomethodology as a Phenomenological
Approach in the Social Sciences." Don Ihde and
Richard M. Zaner eds., Interdisciplinary
Phenomenology♦ 73-98.
Rasmussen, David. Mvthic-Svmbolic Language and
Philosophical Anthropology: a Constructive
Interpretation of the Thought of Paul Ricoeur.
Nijhoff, 1971.
334
Ray, Ruth and Ellen Barton. "Response to Christina Haas
and Linda Flower, 'Rhetorical Reading Strategies
and the Construction of Meaning'" CCC 40 (1989)
480-481.
Reagan, Charles. "Hermeneutics and the Semantics of
Action." Pre/Text 4 (1983): 239-255.
Richards, I. A. Principles of Literary Criticism. New
York: Harcourt Brace, 1925.
Practical Criticism. New York: Harcourt Brace,
1955.
Ricoeur, Paul. Freedom and Nature: The Voluntary and the
Involuntary. Trans. Erazim V. Kohak. Evanston:
Northwestern Univ. Press, 1966.
. Fallible Man. Trans. Charles A. Kelbley. Chicago:
Henry Regnery, 1965.
The Symbolism of Evil. Trans. Emerson Buchanan.
New York: Harper and Row, 1967.
Freud and Philosophy: An Essay on Interpretation.
Trans. Denis Savage. New Haven: Yale Univ. Press,
1970.
. History and Truth, tr. Charles A. Kelbley.
Evanston: Northwestern Univ. Press, 1967.
. Husserl: An Analysis of His Phenomenology. Trans.
Edward G. Ballard and Lester E. Embree. Evanston:
Northwestern Univ. Press, 1967.
The Conflict of Interpretations: Essays in
Hermeneutics. Ed. Don Ihde. Evanston:
Northwestern Univ. Press, 1974.
"Ideology and Utopia as Cultural Imagination."
Philosophic Exchange. 2 (1976): 16-28.
Interpretation Theory; Discourse and the Surplus
of Meaning. Fort Worth: Texas Christian Univ.
Press, 1976.
• The Philosophy of Paul Ricoeur: an Anthology of
His Work. Ed. Charles E. Reagan and David Stewart.
Boston: Beacon Press, 1978.
335
. The Rule of Metaphor: Multi-Disciplinary Studies
of the Creation of Meaning, tr. Robert Czerny.
London: Routledge and Kegan Paul, 1978.
. Main Trends in Philosophy. New York: Holmes &
Meier, 1979.
. Hermeneutics and the Human Sciences: Essays on
Language. Action and Interpretation, ed. and tr.
with an Introduction by John B. Thompson. London:
Cambridge Univ. Press, 1981.
. "A Critique of B.F. Skinner's 'Beyond Freedom and
Dignity.'" Philosophy Today 17 (1983): 166-75.
Time and Narrative. Vol. 1, tr. Kathleen
McLaughlin and David Pellauer. Chicago: Univ. of
Chicago Press, 1984.
. Time and Narrative. Vol. 2, tr. Kathleen
McLaughlin and David Pellauer. Chicago: Univ. of
Chicago Press, 1985.
Time and Narrative. Vol. 3, tr. Kathleen
McLaughlin and David Pellauer. Chicago: Univ. of
Chicago Press, 1988.
"Toward a 'Post-Critical Rhetoric'?" Pre/Text 5
(1984): 9-15.
Lectures on Ideology and Utopia. Ed. George H.
Taylor. New York: Columbia Univ. Press, 1986.
Roche, Maurice. Phenomenology. Language, and the Social
Sciences. London and Boston: Routledge & Kegan
Paul, 1973.
Rodrigues, Raymond. "Moving Away from Writing Process
Worship." English Journal 74:5 (1985): 24-27.
Rorty, Richard. Philosophy and the Mirror of Nature.
Princeton: Princeton Univ. Press, 1979.
Consequences of Pragmatism. Minneapolis: UNiv. of
Minnesota Press, 1982.
Rose, Mike. "Rigid Rules, Inflexible Plans, and the
Stifling of Language." CCC XXXI:4 (1980): 389-401.
. Writer's Block: the Cognitive Dimension.
336
Carbondale: Southern 111. Univ. Press, 1984.
, ed. When a Writer Can’t Write. New York: Guilford,
1985.
. "The Language of Exclusion: Writing Instruction at
the University." CE 47:4 (1985): 341-59.
Schutz, Alfred. The Problem of Social Reality. Vol. I
of Collected Papers Ed. Maurice Natanson. The
Hague:Martinus Nijhoff, 1967.
Schutz, Alfred, and Thomas Luckmann, The Structures of
the Life-World. Vol. 1. Trans. Richard M. Zaner
and H. Tristram Engelhart, Jr. Evanston:
Northwestern Univ. Press, 1973.
. The Structures of the Life-World. Vol. 2. Trans.
Richard M. Zaner and David J. Parent. Evanston:
Northwestern Univ. Press, 1989.
Scribner, Sylvia and Michael Cole. The Psychology of
Literacy. Cambridge, MA: Harvard Univ. Press,
1981.
Silverman, Hugh J. "For a Hermeneutic Semiology of the
Self." Philosophy Today 23 (1979) 199-204.
Simon, H.A. and Hayes, J.R. "The Understanding Process:
Problem Isomorphs." Cognitive Psychology 8
(1976): 165-190.
Stack, George J. "Husserl's Concept of the Human
Sciences." Philosophy Today 17 (1973): 52-61.
Steinberg, Erwin R. "Protocols, Retrospective Reports,
and the Stream of Consciousness." CE 48 (1986):
697-712.
Sterelny, Kim. The Representational Theory of Mind: An
Introduction. Cambridge, Mass.: Basil Blackwell,
1990.
Sternglass, Marilyn. The Presence of Thought:
Introspective Accounts of Reading and Writing.
Advances in Discourse Processes XXXIV. Norwood,
NJ: Ablex, 198 9.
Swearingen, C. Jan. "Between Intention and Inscription:
Toward a Dialogical Rhetoric." Pre/Text 4 (1983):
337
257-271.
Sweeney, Robert D. and Louise Wetherbee Phelps.
"Rhetorical Themes in the Work of Paul Ricoeur: A
Bibliographical Introduction." Pre/Text 4 (1983):
215-223.
Thompson, John B. Critical Hermeneutics: A Study in the
Thought of Paul Ricoeur and Jurgen Habermas.
Cambridge: Cambridge Univ. Press, 1981.
Tomlinson, Barbara. "Talking About the Composing
Process: The Limitations of Retrospective
Accounts." Written Communication 1:4 (1984):
429-45.
Van Den Hengel, John W., S.C.J. The Home of Meaning:
The Hermeneutics of the Subject of Paul Ricoeur.
Washington: Univ. Press of America, 1982.
Van Leeuwen, Theodor Marius. The Surplus of Meaning:
Ontology and Eschatologv in the Philosophy of Paul
Ricoeur. Amsterdam: Editions Rodopi, 1981.
Voss, Ralph F. "Janet Emig's The Composing Processes of
Twelfth Graders: A Reassessment." CCC XXXIV:3
(1983): 278-83.
Wellek, Rene and Austin Warren. Theory of Literature.
3rd ed. New York: Harcourt Brace, 1956.
Willard, Dallas. "The Absurdity of Thinking in
Language." The Southwestern Journal of Philosophy
41 (1973): 125-32.
Willard, Dallas. Logic and the Objectivity of
Knowledge: A STudv in Husserl's Earlv Philosophy.
Athens, Ohio: Ohio Univ. Press, 1984.
Winterowd, W. Ross. "'Topics1 and Levels in the
Composing Process." CE 34:5 (1973): 701-09.
Winterowd, W. Ross., ed. Contemporary Rhetoric: A
Conceptual Background with Readings. New York:
Harcourt Brace Jovanovich, 1975.
Winterowd, W. Ross. "Transferable and Local Writing
Skills." Journal of Advanced Composition 1:1
(1980):1-3.
338
Winterowd, W. Ross. Composition/Rhetoric; a Synthesis.
Carbondale: Southern 111. Univ. Press, 1986.
Winterowd, W. Ross. "Viewpoints: Dramatism and the
Composing Process." RTE 20:3 (1986): 317-327.
Witte, Stephen P. and Roger D. Cherry. "Writing
Processes and Written Products in Composition
Research." In Charles Cooper and Sidney Greenbaum,
eds. Studying Writing: Linguistic Approaches. A
Written Communication Annual. Vol. 1. Beverly
Hills: Sage, 1986.
Witte, Stephen. "Pre-Text and Composing.", CCC 38
(December 1987): 397-425.
339

Linked assets