00001.tif

PDF

Download

Open document

Transcript (if available)

Content INTERPRETIVE SCHEMATA AND LITERARY RESPONSE
by
Terry Lynn Beers
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 1986
UMI Number: DP23109
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.
Dissertation Publishing
UMI DP23109
Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author.
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 CAUFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CAUFORNIA 90089
Ph
t
' 8 kz>
BH 15
3/ yjr£J?-£f
This dissertation, written by
Terry Lynn Beers
under the direction of h .f.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 T O R O F P H IL O S O P H Y
^ 'D e a n of Graduate Studies
D a te Max..15,,.1986.
DISSERTATION COMMITTEE
Chairperson
For my parents, Orville and Marjorie
ii
ACKNOWLEDGEMENTS
I wish to thank the members of my committee, Dallas
Willard for his encouragement and discerning criticisms,
Charles Berryman for his careful reading of especially the
final chapter, and most of all Marilyn Cooper for her
tireless guidance and her remarkable editorial acumen.
I owe, as well, an intellectual debt to my teachers
and to my fellow students at the University of Southern
California, especially Louise Wetherbee Phelps and Robert
Reichle for many hours of conversation and debate.
xxi
TABLE OF CONTENTS
Page
DEDICATION............ ii
ACKNOWLEDGMENTS...................................... iii
LIST OF ILLUSTRATIONS................................. v
INTRODUCTION . . . .......................................... 1
Chapter
I. READER-CENTERED APPROACHES TO LITERATURE:
AWAY FROM THE T E X T ........................... 8
II. TOWARD AN ECOLOGICAL THEORY OF RESPONSE . . . 32
III. SCHEMA THEORETIC MODELS OF READING: PROCESS
BECOMES DESIGN ................................. 64
IV. INTERPRETIVE SCHEMATA AND LITERARY RESPONSE . 106
V. CRITICISM AND THE POETRY OF ROBINSON
JEFFERS............................................156
N O T E S ............ . 200
WORKS CONSULTED............... 206
i v
ILLUSTRATIONS
Figure Page
1. Kinneavy’s Model of Discourse ........... 148
2. A Temporal Model of Discourse . . . . . . . . . 148
v
INTRODUCTION
The shift away from the New Criticism and its cult of
textual authority is partly due to literary theorists
dissatisfied with critical models that ignore the role of
the reader in interpretive activity. Many of these
scholars have refused to accept the New Critic's view of
the text as the container and sole arbiter of meaning;
consequently, they also reject the assumption that
underlies such a metaphor, that readers and writers are
projections of the text. Instead, reader-centered critics
prefer to view the physical text as an invitation, an
opportunity for readers to make meaning through the reading
process. However they pay a price for abandoning the
metaphor of text as container, for when the reader is held
to have primary power to constitute meaning — and
according to some critics the text itself — meaning
becomes a subjective experience beyond the power of the
text to control, constrain, stabilize, or validate.
Theoretically, the problem becomes how to describe the
means that relate the text, the reader, and the social
context that surrounds them in order to account for the
fact that competent readers legitimately generate
interpretations of texts that agree and disagree in various
1
ways; practically, the problem becomes how to choose the
most valid among competing interpretations.
Some scholars, for example Stanley Fish and E.D.
Hirsch, rely on cultural, social, and/or philosophical
constructs — conventions, strategies, and normative
interpretive behaviors — to describe the constraints
working within readers and to provide criteria for
evaluating the validity of particular readings. But, for
the most part, these constructs have remained controversial
since some, for example Fish’s notion of "interpretive
community" (Text 171), fail to describe the means that
relate texts, readers, and social contexts in specific
ways; and some, for example Hirsch's concept of "intrinsic
genre" (Validity 86), are assumed to be relatively stable,
despite the fact that they are subjectively realized by
individual readers with idiosyncratic goals. Thus, in
place of the objective, textual standard of the New
Critics, these "reader-centered" critics have sought a new
standard by which the constraints upon the subjectivity of
readers can be explained. However, the nature of this
standard has proved to be the knottiest problem for
reader-centered theorists.
In the following pages I evaluate these problematic
aspects of reader-centered theories of reading literature
in light of the theoretical work of reading researchers and
2
cognitive scientists who work in the province of schema
theory, a theory of how conceptual knowledge is represented
to consciousness. Schema theory — the name is derived
from Kant (B180-181) — is endorsed by a consensus of
theorists in reading research and represents, I believe,
the best theoretical account of how readers come to
understand extra-literary texts; by extension, I believe it
has extraordinary potential to aid literary critics in
understanding how readers come to understand and therefore
judge literary texts as well.
The two chapters that follow establish the need for a
more specific theoretical account of reading literature by
showing how reader-centered literary theorists —
represented primarily by Roland Barthes, E.D. Hirsch,
Stanley Fish, and Louise Rosenblatt — account for the
relationships among the different constituents of the
rhetorical transaction of reading, namely the reader, the
text, and the social environment. A subsequent chapter
will focus on research into schema theory to establish its
major tenets by examining the work of Marilyn Adams and
Allan Collins, Walter Kintsch and Teun van Dijk, Roger
Schank and Robert Abelson, and David Rumelhart (among
others) to show how these researchers have contributed to a
coherent theory of the relative roles of readers, texts,
and social contexts in the constitution of meaning.
3
Additionally, this chapter will offer a critique of schema
theory based on the inadequacies of the machine metaphor
upon which it is founded, citing relevant work in the
fields of linguistics, cognitive psychology, social theory,
and philosophy; specifically my intention is to show that
schema theory can be modified to accomodate a fluid,
dynamic, and anti-mechanistic view of complex human
acitivities like reading by forsaking the machine metaphor
of its genesis to embrace a more humanistic metaphor of
design.
Together, these chapters will reveal the need and the
opportunity for a different approach to the problem of how
readers are constrained. At least one critic, Steven
Mailloux, has attempted to fulfill this need by describing
a typology of literary conventions (Interpretive
Conventions 150); however, I believe that a critical theory
that synthesizes schema theory and literary criticism will
prove theoretically more compelling and practically more
useful. Both the aesthetic and cognitive strands of
reading research rely on similar assumptions, specifically
that textual form has a temporal dimension, that readers
constitute texts and their meanings, and that there can be
no subject/object dichotomy of reader and text upon which
to found critical theories. Still, the two approaches are
complementary: the work of reader-centered literary
4
theorists focuses on the aesthetics and social contexts of
texts (as well as the standards of valid readings), and the
work of reading researchers in schema theory focuses on how
conventions and interpretive strategies are represented to
consciousness to enable the experience of reading.
Finally, I evaluate the resultant critical approach in
the context of a significant literary problem, the adequacy
of the critical approaches to the narrative poetry of
Robinson Jeffers. After the publication of his first major
narrative "Tamar," Jeffers was held in high esteem among
contemporary critics. Mark Van Doren felt that few volumes
of any sort had struck him with as much force as Jeffers'
book (268), and James Rorty wrote "Nothing as good of its
kind has been written in America" (1). But with the
hegemony of the New Critics, Jeffers' reputation declined,
partly because his verse was difficult to characterize in
terms of the currently accepted categories of literature:
many critics continued to praise their "lyrical"
expression, while others lamented their "prosaic"
language. Yvor Winters' comments are typical of the
antipathy of New Critics to Jeffers' work:
Mr. Jeffers' verse continues to miss the virtues
of prose and verse alike: it is capable neither
of the fullness and modulation of fine prose nor
of the concentration and modulation of fine verse.
There is an endless, violent monotony of movement,
wholly uninteresting and insensitive....
("Review" 684)
5
But as William Everson has pointed out, Jeffers' long
poems do not move in the way that Winters believes that
narratives should (Fragments 116), and other categories
need to be considered. I believe that the wide discrepancy
among Jeffers’ critics is a critical problem that can be
explained by a schema-theoretic approach, since an
exploration of the relationships among reader/critic,
social context, and literary text will demonstrate how
different aspects of the same text become more valuable for
some readers than for others, according to their preferred
schematic categories. Second, even though Jeffers’
criticism is represented by a number of critical approaches
(including the text centered criticism of Winters, the
religious/philosophic criticism of Everson and Arthur
Coffin, the mythic criticism of Robert Brophy, and a source
study by Radcliffe Squires), the persistence of this
critical discrepancy points to the need for a new
explanation. Specifically, I examine Jeffers' narrative
poem "Thurso's Landing," the occasion of some of the most
wide-ranging critical response. Winters charged that the
book was "composed almost wholly of trash" ("Review" 684),
yet Babette Deutsch, an important contemporary critic of
Jeffers, found much to admire: "Jeffers continues to employ
his long, deep-breathed, billowing line; avoids rhyme; here
is more than usually chary of ornament. The quality of the
6
poetry — music and high imagination -- remains" ("Hunger"
7). Thus "Thurso's Landing" offers a typical instance of
the critical problem of reading Jeffers' narrative verse,
one which a reader-based, schema-theoretic approach might
well resolve.
Chapter I
READER-CENTERED APPROACHES TO LITERATURE:
AWAY FROM THE TEXT
Reader-response criticism, as Jane Tompkins notes, is
the usual term critics use to describe "the work of critics
who use the words reader, the reading process, and response
to mark out an area for investigation" (ix). I use the term
reader-centered criticism to include the work of those
critics described above and also to include the work of
critics, notably E.D. Hirsch, who are usually not
associated with the reader-response movement, since their
main interest is in the role of authorial intention in the
interpretation of texts. However, according to Hirsch, the
means of recognizing this intention depends upon the
re-cognitive acitivities of the reader, making it necessary
for the critic to recognize the activity of the reader as
well as the constraints of authorial intention. Since all
reader-centered critics take active interest in the role
readers play in the understanding and evaluation of
literature, they can be understood to be opponents of the
tendency of American New Critics who focus critical
attention almost exclusively upon the text.
8
This "New Criticism" was both a theoretical and a
practical enterprise. The goal was an understanding of
literary texts through the practical study of these texts
as verbal artifacts that presented an unchanging pattern of
objects and emotions. Thus John Crowe Ransom called for an
"ontological criticism" that would study these artifacts —
specifically poems — to determine their unique structures,
their peculiar order of poetic content (280). The better to
achieve this understanding, New Critics courted an almost
scientific objectivity in order to differentiate among the
work itself — the practical object of study — and a host
of correlative issues, including the intentions of the
author and the effects upon the audience. Failure to
maintain these distinctions was anathema, the consequences
of which were clearly described by Wimsatt and Beardsley in
two influential essays. "The Intentional Fallacy" resulted
from attempting to derive the standards of criticism from
the psychological causes of a work’s origin; "The Affective
Fallacy" resulted from attempting to derive the standards
of criticism from the psychological effects of a work upon
an audience. Wimsatt and Beardsley believed that both
approaches lead to an unscientific relativism because "the
poem itself, as an object of specifically critical
judgment, tends to disappear" ("Affective Fallacy" 345).
9
The hegemony of the New Criticism, however, was
largely maintained thanks to its practical rather than
theoretical contributions, for it did much to foster an
effective pedagogical method. By focusing almost
exclusively on the work itself, and by assuming that it was
this work that contained meaning and literary quality,
teachers of literature were free to ignore or discount
spurious responses by students to literary works and were
not obliged to dwell overlong on biographical, social,
cultural, or psychological issues that were not actually
contained within the text; according to Terry Eagleton,
"Distributing a brief poem for students to be perceptive
about was less cumbersome than launching a Great Novels of
the World course" (50). Teachers could assume a central
focus of study — the text — which was assumed to contain
evidence that supported or refuted critical
interpretations, thus streamlining the critical apparatus
necessary for teaching (and taking) a successful course in
literature. Moreover, New Criticism, thanks to its
theoretical objectivity, tended to model itself upon, and
thus to compete with, the hard sciences; it was a means by
which literary criticism could become accepted as a more
professional academic discipline (Eagleton 49).
However, any theory of literary criticism that argues
for the text as the primary container and arbitor of
10
meaning (including the New Criticism of Wimsatt and
Beardsley) must also accept the theoretical assumptions
that underlie such a position: as the practical focus of
investigation for the critic, the text becomes an isolated
object through which all aspects of the discourse situation
must be explained. Thus the text is inherently "poetic" or
"non-poetic" (that is literary language is qualitatively
different from ordinary, standard language); the text is an
object that exists in the world regardless of whether or
not it is read; the validity of an interpretation of a text
is a matter of demonstrating that the interpretation
conforms to evidence that exists in the text; and, as a
result, a text has the primary power to determine how it is
read and experienced by an audience. If the New Critics
achieved a positive result, it was to rescue the text from
critical obscurity; the corollary negative result was that
readers and writers were treated as if they were artifacts
of the text, becoming just about neglected in the practice
of criticism.
As Robert de Beaugrande has written, a critical theory
is both a description and an advocacy of a way to make use
of evidence ("Critical Theories" 539). As such,
reader-centered theories of literary criticism explicitly
reject the New Critics' reliance on the text as an
independent object and container of meaning with the
11
primary power to determine how it is read and understood by
an audience. Some reader-centered critics justify this
rejection by appealing to the work of philosophers who have
questioned the Cartesian split between subject and object.
Louise Rosenblatt depends upon the work of John Dewey
(''Viewpoints" 99) whose theory of aesthetic response
emphasizes the human contribution,^ and E.D. Hirsch invokes
2
Husserl's phenomenology as a justification of his
distinction between meaning and its significance to human
consciousness ("Meaning and Significance" 204). But some
less speculative work in the fields of linguistics and
pragmatics also serve to question the objective stance of
text-centered critics by denying the essential difference
between literary and non-literary language — and thus
denying the independent existence of their preferred object
of study, the text as art object. Once this split is
rejected, critical theory must redefine its practice.
Mary Louise Pratt asserts that the belief in
inherently poetic and non-poetic texts implied by the
stance of the New Critics is the main area of overlap
between New Criticism and European Formalism (xiv-xv), and
her arguments against the Formalist "dogma" apply as well
to New Criticism. Pratt maintains that literary discourse
is a use rather than a kind of language (xiii), pointing
out that the assertions of Formalists about the qualities
12
of literary language — for example Mukarovsky's assertion
that poetic language is "foregrounded" by its essential
difference from standard language (43) -- are weakened by
their failure to supply examples from non-literary
discourse free of literary qualities (5). In fact, such
examples of purely non-poetic language may not exist,
William Labov has shown how language in an extra-literary
setting has qualities associated with literary discourse
(Language in the Inner City); Erving Goffman argues
persuasively that frames which define and organize our
expectations about discourse situations are at least
partially responsible for the way we view discourse — for
example, as a joke, a description, or as a performance
(10); and George Lakoff and Mark Johnson have shown that
metaphor, supposedly a distinctly literary use of language,
is typically used to define abstract concepts in
extra-literary settings (196). By demonstrating that the
existence of a distinct "standard" dialect against which
literary language may be compared and evaluated is a
dubious theoretical proposition, at the very least, the
arguments of Pratt, Labov, Goffman, and Lakoff and Johnson
suggest that the linguistic distinctions supposedly
contained in a text are in some measure dependent upon the
participation of readers. This is exactly the point of
Louise Rosenblatt who maintains that the essential
13
difference between the literary and the non-literary text
is the adoption of a particular stance by the reader —
either aesthetic or efferent — that determines the quality
of the reading experience (Reader 102).
Moreover, once the role of the reader is acknowledged
in the distinction of literary and non-literary qualities,
the assumption that the text is an independent object that
exists apart from reading is untenable. According to
Rosenblatt, the reading process is a temporal, lived
experience, attention to which is constituative of the
aesthetic stance (Reader 102). Necessarily, the text is no
longer an isolated object, but an entity that is part of
the entire temporal discourse situation whose elements
combine to affect the perceived quality of the text. This
temporal element — itself a part of most models of
3
reader-response criticism — changes the ontological
status of what heretofor had been called the text from an
independent spatial array to a dependent, impermanent
construction of the reading process.
The next assumption of the New Critical stance — that
the validity of an interpretation of a literary work
depends upon its fidelity to evidence in the text — is
also weakened if the arguments above are accepted. Since
qualities of the text can be held to be dependent upon the
activity of the reader, and since this activity takes place
14
over time in a particular discourse situation, the text can
hardly be considered to conatain objective evidence. A
highly significant corollary to this assertion is that
critics themselves are not independent, disinterested, and
detatched individuals capable of dispassionate judgment.
Most reader-centered critics recognize to a greater or
lesser extent their own contribution -- and their
responsibilities — to the texts they evaluate: "for rather
than being merely a player in the game, [the critic] is a
maker and unmaker of its rules" (Fish, Text 367). Such a
view has at least one theoretical advantage over the New
Criticism and Formalist approaches to texts: it can account
for the different interpretations that readers produce,
since individual readers are responsible for constituting
the rules by which interpretations are made and texts are
no longer held to be containers of a particular, objective
meaning.
However, the refusal to accept the text as the
container and arbitor of meaning has at least one practical
drawback: if the text no longer contains meaning, and if
meaning, literary quality, even the text itself are
products of active readers who themselves constitute the
"rules" of interpretation, then how are critics, deprived
of any objective standard, to validate any particular
interpretation? Stanley Fish believes such an objective
15
standard is impossible; moreover, he claims for his brand
of reader-centered criticism a moral superiority over
critical theories that believe in the objectivity of such a
standard: "I would rather have an acknowledged and
controlled subjectivity than an objectivity which is
finally an illusion" (Text 49). It is the nature of this
"controlled subjectivity" that has proved to be the
knottiest problem for reader-centered theorists.
The apparent opposition of reader-centered theorists
to the tenets of the New Criticism does not, however, imply
that they are unified in their approaches to literary
texts. In fact, they differ greatly in approach -- with
the exception that they do not isolate the text from its
readers. Stephen Mailloux has proposed a taxonomy of these
approaches that discriminates among three broad categories
— psychological, intersubjective, and social — which is
useful in delineating the differences and in discovering
the relative drawbacks and virtues among reader-centered
approaches to literary texts (Interpretive Conventions 22).
The two best known models that fall within Mailloux's
first category, the psychological, are those of David
Bleich and Norman Holland. Both critics strongly reject the
notion of the text as an independent object, but Bleich’s
position is the more radical in its movement away from
textual constraints upon the reading process. For Bleich,
16
the issues are ultimately epistemological; he holds that
knowledge is never independent of a system of individual
motives and collective beliefs. Therefore, "if we are to
speak of what we want to know, we have to first identify
ourselves, and then ask ourselves 'why do I in particular
want to know this in particular'" ("Pedagogical Directions"
456). Individual motives and beliefs thus influence our
perception and understanding of knowledge and objects,
making the knower and the known inseparable (456).
Bleich founds upon this subjective orientation a
critical model of three stages. During the "response"
stage readers perceive the text as an object and associate
the individual forms they perceive that object to have with
concepts, symbols, and meanings according to their own
individual motives and beliefs (Subjective Criticism 96).
The second stage, "resymbolization," is the process of
understanding that response within the context of
individual motives (Sj3 39, 88). For Bleich, like Fish, this
is a moral issue; admitting to the subjective basis of
individual interpretations results in "an increase in
mutual enlightenment both with regard to how one interprets
and how each participant is predisposed beforehand"
("Pedagogical Directions" 457). The third stage,
"negotiation," is the communal aspect of the model, where
individual interpretations are negotiated with a responding
17
community that itself is producing or has produced
resymbolizations of the same text (Sh3 151-9). The three
stage model that Bleich has developed defies the objective
status of the text both by focusing on the subjective
aspects of reader response and by being necessarily a
temporal model where the initial symbolization and the
resymbolization is a constant, dialectical process. Thus
resymbolization emerges against a background of past images
that reveal contradictions and contrasts, to borrow
Wolfgang Iser's terminology (Act 148). Negotiation takes
place after resymbolization to eventually transform the
impermanent symbolization into a more permanent socially
supported interpretation. But because these are
negotiations, not demonstrations, they too are in flux,
changing with time, and as such inconsistent with the idea
of permanent meaning contained within a static, objective
text.
Like the New Critics, Bleich has shown an
extraordinary interest in the application of his model to
the teaching of literature and an extraordinary awareness
of the effects a theory of reading can have on the way
students respond to literature, on classroom procedures,
and on the authorization of interpretations (Tompkins
xxi). And like the New Critics, perhaps Bleich's most
important contribution is to the pedagogy of literature,
18
for his model of reading turns the classroom into a
community of interpreters that negotiates among individual
subjective responses, guaranteeing the critical, evaluative
participation of everyone — student and teacher alike.
However, this practical advantage has, according to
Mailloux, a theoretical disadvantage: "Bleich has presented
himself with an apparently impossible task: to account for
interpretive agreement after having established the primacy
of the individual interpreter" (I_C 33). Consequently,
Bleich's model contributes to an understanding of how
differences themselves are adjudicated by negotiation, but
it does little to explain why such adjudication is not
always, or is only partially, necessary.
Like David Bleich, Norman Holland centers his critical
theory on the identity of the individual reader. But while
Bleich focuses on the subjective role of the individual,
Holland’s model attempts to describe more nearly a
transaction between the text and the reader (Mailloux, I_C
24). Holland’s central tenet is that identity recreates
itself: "That is, all of us, as we read, use the literary
work to symbolize and finally to replicate ourselves"
("Unity" 124). This process procedes by a process Holland
describes with the acronym DEFT, a combination of
idiosyncratic defenses, fantasies, and transformations that
constitute an individual’s "identity theme" ("Unity" 127).
19
In short, according to Holland, readers react to a text in
idiosyncratic ways — to obtain what they most wish and
fear. First, they shape the work so it will pass through a
network of adaptive and defensive strategies for coping
with the world; second, they re-create from it the
particular kind of fantasy and gratification they
individually respond to; finally, they transform this
fantasy into a total, unified experience of the literary
work (''Unity'1 125-6). In effect, Holland's DEFT formula is
a more detailed specification of the subjective elements
inherent in Bleich's model, offering a more precise theory
of how the individual achieves the initial symbolization
(based upon the defense/fantasy constraint) and how the
individual interprets this symbolization through a unified
transformation consistent with the individual's identity
theme.
But unlike Bleich, Holland attempts to mitigate the
subjective bent to his theory by giving power to the text.
In Poems in Persons, Holland speaks of the text as a raw
material that itself constrains the reading process (96).
This position is consistent with Holland's attempt to make
of reading a transaction between text and reader, but as
Mailloux points out, the assertion that the text constrains
the reading process in some way is left unexplained by
Holland, and more importantly, it is inconsistent with the
20
emphasis he places upon the identity theme of the reader,
for if readers constitute texts through their identity
themes, it is difficult to see how the text could constrain
such a process a priori (I_C 27). This inconsistency is a
major theoretical thorn because if Holland had succeeded in
explaining how the text could constrain the reading
process, he could account for the substantial agreements
that do proliferate among textual interpretations.
Bleich has accused Holland of operating within an
objectivist paradigm. He maintains that "Holland speaks as
if defenses, expectations, fantasies, and transformations
are discrete items perceivable by anyone studying response"
("Pedagogical Directions" 464). In fact, Bleich's criticism
4 .
is overzealous — Holland s formulas, for the sake of
analysis and explanation, must needs seem more objective
than he would like, and his analysis, for example, of the
identity theme of Robert Frost ("Unity" 127-30),
necessarily procedes within Holland's own identity theme,
thus refuting the accusation of an unacknowledged
objectivist bias. But Bleich's objection raises a salient
point about psychological models of reader-centered
criticism: in their enthusiasm to explain the differences
among interpretations through the adoption of a subjective
criticism -- to "close the dualism that has dominated
systematic thought since Descartes: the belief that the
21
reality and meaning of the external world exist alone,
independent of the perceiving self" (Holland, "Unity" 130)
— they neglect to develop a strong intersubjective
foundation for a model of communication. They have lost
sight of the fact that, with some exceptions, communication
does occur, that agreement — as well as disagreement —
must be constituents of a model of literary criticism.
Mailloux’s second category, the inter-subjective,
describes the phenomenological model of Wolfgang Iser and
the critical model Stanley Fish calls "affective
stylistics." These models are less subjective than those
of Bleich and Holland, for both tend to pay a good deal of
attention to the text even though they place meaning in the
mind of the reader. The attention these models pay to the
text is similar to the transactive criticism of Holland,
and the reason for this is much the same, for by attending
to the role of the text in the reading transaction, these
theorists attempt to claim for their models an
inter-subjective base which will explain agreement among
interpretations. But unlike Holland, Iser and Fish are
less concerned with attempting to describe the specific
dynamics of individual personalities during their response
to literary texts. Instead, they attempt to account for
differences among interpretations by emphasizing the
22
temporal element in the reading process and thereby
challenging in this way the objective status of the text.
Iser begins by recognizing that the whole text can
never be perceived at any one time (Act 108); as a result,
the object of the text can only be imagined by way of
different consecutive phases of reading.
The relation between text and reader is therefore
quite different from that between object and observer:
instead of a subject-object relationship, there is
a moving viewpoint which travels along inside that
which it has to apprehend. This mode of grasping an
object is unique to literature. (Act 109)
He further guarantees the ability of his model to handle
difference by asserting that the reading process itself is
selective; subsequent readings of the same text produce
different interpretive shades because the reader attends
more closely to some aspects of the text and less carefully
to others in consecutive readings ("Reading Process" 55).
If this process of selective attention is true of a single
reader, it is true as well for a community of readers, each
of whom will attend more carefully to different aspects of
the text.
Iser's careful attention to the role of the text
forces him to confront, as Holland does not, the specific
problem of what the text is. If it is not one polarity of
the subject-object duality, as Iser insists, how can it
present particular aspects to which a reader may attend?
23
To answer this question, Iser relies on a concept developed
by Roman Ingarden in The Literary Work of Art,
concretization. According to Iser, "The text itself simply
offers ’schematized aspects’ through which the subject
matter of the work can be produced, while the actual
production takes place through an act of concretization"
(Act 21). Thus the text offers certain schemata which
constrain the reading process and guide the reader’s
process of selection which in turn concretizes the work.
More specifically:
...the text mobilizes the subjective knowledge
present in all kinds of readers and directs it
to one particular end. However varied this know
ledge may be, the reader’s subjective contribution
is controlled by the given framework. It is as if
the schema were a hollow form into which the reader
is invited to pour his own store of knowledge.
(Act 143)
Iser has made an important explicit distinction that was
implicit in the work of Bleich and undeveloped in the work
of Holland: that among the text as a physical object, the
reader, and the work that is "concretized" as a result of
the reader’s transaction with the text.
However, Iser asserts that the text not only contains
schemata (or "gaps" that require fulfillment ["Reading
Process" 55]), but as a result of the structure of these
schemata, the text requires a particular kind of reader,
the "implied reader" which makes it possible to describe
24
literary effects (Act 38). Such an assertion links Iser
with neo-Aristotelean critics such as Wayne Booth and
Walker Gibson, each of whom theorizes about how specific
elements of texts imply particular sorts of readers."* And
it betrays the main weakness of Iser's theory; even though
he provides a site (the text as physical object) for the
inter-subjective element of a critical theory necessary to
explain agreement, he does not say how the structure of the
text limits the activity of readers (Tompkins xv) or
precisely how a text can thus imply a particular sort of
reader. Frank Kerraode's review of Iser’s The Act of
Reading uncovers the problem:
What one misses, I think, is any developed theory
of competence. How does the reader come to be able
to do all this [fill gaps, follow the instructions
of the text], and by what criteria are subjective
readings tested against that inter-subjective
structure? A fully developed theory of competence
...is needed on all hands. (300)
Another of Kermode’s objections to Iser's work is that
he is more concerned with the interaction of text and
reader than with the testimony of actual historical
readings (298-9). Such a charge cannot be leveled against
Stanley Fish who developed his model of affective
stylistics out of his concern for the actual reading of
literary texts, especially the works of John Milton.
25
In Surprised by Sin; The Reader in Paradise Lost, Fish
asserts that a correct reading of Milton depends on three
assumptions: that the reading experience takes place in
time; that the habit of moving eyes along a page and back
is never abandoned, making the line as a unit a resting
place; and that a mind asked to order a succession of
detail (mental or physical) seizes on the simplist scheme
of organization (23). In effect, Fish is advancing a
critical reading model that gives the reader the power to
organize detail (and by implication the text) and that
depends upon temporal processes to effect that
organization. In fact, Fish goes so far as to assert that
the site of the struggle of Paradise Lost is not Eden, or
even the text, but the mind of the reader, and thus that
the poem itself demonstrates the fallen state of man
because the reader is confronted with evidence of his own
corruption through his inability to respond adequately to
the spiritual conceptions of the poem (S_S ix). Such a
position denies the objective status of the text, since
meaning is dependent upon the reading process; yet such a
position is not truly subjective because Fish is attempting
to preserve the status of the text as an entity that
somehow manipulates a reader in specific ways.
Fish’s method of reading Milton becomes a more
specific model for criticism in his essay "Literature in
26
the Reader: Affective Stylistics.” Specifically attacking
Wimsatt and Beardsley's conception of the affective
fallacy, Fish argues that affect is a necessary constituent
in determining the meaning of any text (Text 25).
Recognizing that these effects are the result of a process
of reading that takes place over time, Fish advocates
slowing down the reading process and attending to all of
the events that occur during the process of reading. By
doing so, Fish maintains, the critic is less likely to skip
over prominent stylistic features to a direct explanation
of their meaning that ignores the question of whether or
not their prominence has any direct relationship to their
effect upon the reader (Text 41).
Like Iser's model, Fish's model of "affective
stylistics" is overtly phenomenological, attempting to
relate specific elements of texts to reader reactions and
then organizing and explaining those reactions as effects
that were temporally ordered. Fish claimed superiority for
his method on the basis that it recognized the subjective
role of the reader as opposed to an illusive objective
reality, and he claimed for it (and still does) a
pedagogical effectiveness. However, Fish's claims for his
method contained a drawback, perhaps characterized best by
Fish himself in a later commentary upon his essay:
27
In place of the objective and self-contained text
I put "the basic data of the meaning experience" and
"what is objectively true about the activity of
reading"; and, in order to firm up this new "bottom
line" I introduced the notions of the "informed
reader" — designed to take account of, by stigma
tizing, all those readers whose experiences were
not as I described them — and "literary competence"
— designed to stabilize the knowledge the informed
reader is presumed to have. (Text 22)
In essence, Fish had made a claim for the superiority
of his method over objectivist approaches to the text, yet
he paradoxically attempted to describe an objective
approach to the reading experience. This internal
contradiction led Fish to abandon his claims for the
priority of his criticism over Formalist criticism
(Mailloux 3X 23); nevertheless, his model of the informed
reader, who has internalized literary conventions — who
has literary competence — is a first step toward a truly
inter-subjective model of criticism, one based not upon the
text itself but upon social knowledge which organizes and
evaluates texts through the reader and the reading
process.
The psychological and the inter-subjective models of
reader-response — those of Bleich, Holland, Iser, and Fish
— have two principal advantages over New Critical or
Formalist models: they better account for legitimate
differences among interpretations and they can account for
the temporal nature of reading and the effects this has
28
upon the reading process. What these models cannot do is
explain agreements among interpretations nor account for
the undoubted fact that some interpretations are more
useful than others. In the case of psychological models
this is because they focus on the subjective nature of the
individual, practically ignoring the potential of a priori
constraints that would guarantee agreement and provide
criteria for validity; in the case of inter-subjective
models, this is because they either place schema or gaps
within the text without explaining how they can constrain
readers, or because they paradoxically assert the objective
status of the reader's response. Reader-response models
that fall into Mailloux's social category take a step
toward overcoming these theoretical objections.
Social reader-response models account for agreement
among interpretations by assuming a system of socially
derived conventions, rules, and interpretive behaviors that
are internalized by readers. Thus, the a priori
constraints that guarantee the possibility of
communication, agreement, and therefore criteria for the
judgment of validity are not held to be aspects of an
objective text but are considered to work through the
reader who becomes their agent, applying them to the work
during the reading process. For example Jonathan Culler
develops an account of reading literature ultimately
29
derived from the semiotic theories of Saussure who believed
that the meaning of signs — linguistic primitives, either
written or spoken — depended upon the conventional and
ultimately arbitrary association of signifiers with
particular signified concepts.^ Because the association of
signifier and signified is a matter of social convention
(as is the discrimination of signfiers and signifieds),
structuralist critics are interested in a reversal of
perspective away from the interpretation of individual
works and towards the task of formulating a comprehensive
theory of literary discourse which would require, as Culler
points out in Structuralist Poetics, a model of literary
competence based upon a set of conventions for reading
literary texts (118). Thus structuralist critics like
Culler would seem to question the subject/object
orientation from the outset; yet often they paradoxically
reintroduce that very distinction in their practical
discussion of criticism. Culler warns, "Interpretation is
always interpretation of something, and that something
functions as the object of the subject-object relation,
even though it can be regarded as the product of prior
interpretations" (Deconstruction 74).
The persistence of the subject/object orientation in
structuralist literary theory may derive partly from the
model of structuralist linguistics, originally proposed by
30
Saussure, that focuses on supposedly "synchronic,*1
a-temporal linguistic conventions that govern the
association of signifiers and signifieds (101-139). By
assuming these conventions to be stable objects,
structuralist critics often seem to imply that texts, which
depend upon these conventions for their constitution, are
stable as well. Nevertheless, Culler is right to assert
that the shift in perspective to the social conventions of
reading literature leads to a stress upon literature's
dependence upon modes of reading (Structuralist Poetics
128). This shift of perspective has contributed to
reader-response theory the foundation for a consistent
explanation of a priori interpretive agreement without
threatening the subjective aspects that explain
interpretive differences, since these social aspects work
through the subjective consciousness of individual
readers. Four critics who have focused on social
foundation of interpretation are E.D. Hirsch, Roland
Barthes, Louise Rosenblatt, and the later Stanley Fish. An
examination of their theories — perhaps the most
influential socially based models of criticism — will
reveal the theoretical and practical problems for which
reader-centered models of criticism have yet to account.
31
Chapter II:
TOWARD AN ECOLOGICAL THEORY OF RESPONSE
The critical models of E.D. Hirsch and Roland Barthes
focus respectively on two opposing poles of the discourse
situation, the writer and the reader. Where Hirsch is
interested in criteria for validity that depend upon
authors limiting linguistic possibility by "actualizing and
specifying some of those possibilities” (Validity 47),
Barthes is interested in readers resisting the notions of
authorial intention and limits on linguistic possibility;
he believes that to give an author to a text is to impose a
stop clause, to furnish it with a final signification, to
close the writing and seemingly to deny that "the true
locus of writing is reading" ("Death" 11). Nevertheless,
though Hirsch seems to demand the closure of writing that
Barthes denies is desirable — or even possible -- both
critics recognize the importance of social and/or cultural
constraints upon the reading process.
Hirsch’s model is based upon a distinction between
what he terms meaning and significance:
Meaning is that which is represented by a text;
it is what the author meant by his use of a partic
ular sign sequence; it is what the signs represent.
Significance, on the other hand, names a relation
ship between that meaning and a person, or a con-
32
ception, or a situation, or indeed anything imaginable
....Significance always implies a relationship, and
one constant, unchanging pole of that relationship is
what the text means. (Validity 8)
Hirsch was helped to this formulation by expanding
Gottlob Frege's distinction between "Sinn and Bedeutung"
(or between sense and reference) and Husserl's
phenomenology.* Of the two, Husserl's work seems the more
important to Hirsch, for he draws on Husserl's distinctions
among an object as perceived by consciousness (the
intentional object), the actual act of perception (the
intentional act), and the real object that exists
independent of perception; Hirsch argues that different
intentional acts can intend the same, identical object and
that meaning is itself an intentional object that remains
self-identical (Validity 218). Hirsch's easy assertion that
meaning is an object — which appears to be almost an act
of faith — is actually made by Husserl:
They [meanings] are an ideally closed set of
general objects, to which being thought or being
expressed are alike contingent. There are there
fore countless meanings which, in the common,
relational sense, are merely possible ones, since
they are never expressed, and since they can, owing
to the limits of man's cognitive powers, never be
expressed. (Logical Investigations 333)
Hirsch's distinction between meaning and significance
allows him to explain both agreement and disagreement among
readings. By following Husserl and considering meaning to
33
be a stable object of consciousness, he can promise that
his theory will account for validity and agreement; and by
complementing this idea with that of significance he can
account for differences among interpretations since the
significance of a text is free to vary from individual to
individual. However, the main burden of Hirsch's work has
been to explain how it is that meaning becomes an object of
consciousness in the first place, and here he emphasizes
the role of the author in manipulating social constraints.
Hirsch's model in some ways parallels structuralist
models derived from Saussure's analysis of the sign. He
elaborates the concept of meaning by defining it more
particularly as "a willed type which an author expresses by
linguistic symbols and which can be understood by another
through those symbols" (Validity 49). Hirsch says that a
"type" can be represented by more than one instance of
signification but that it remains self-identical, that it
has specific boundaries (Validity 50). For example, "Sir
Walter Scott" and "the author of Waverly" are two instances
of signification that refer to the same type. The
differences between the actual signifiers serve to
guarantee differences in significance, not meaning. Thus,
verbal meanings -- or willed types (Validity 51) — are
signified by particular linguistic signs that are similarly
selected, or willed, by an author. Their understanding is
34
guaranteed because knowledge of types and their habitual
associations with particular arrays of linguistic
signifiers are shared by authors and readers through a
system of conventions (Validity 92).
The concept of sharability introduces readers into
Hirsch's theory, for although Hirsch insists on the
relative importance of authors by giving them the power to
will types and to select the linguistic signs to indicate
them, interpretation depends on the reader associating the
signs with the specifically willed type (meaning) by means
of shared conventions. Hirsch holds that this process is
aided by the selection and rejection of '’genres" which are
heuristic and constitutive of verbal meanings. These
genres are much like Goffman's concept of frame; they are
"...that sense of the whole by means of which an
interpreter can correctly understand any part in its
determinacy" (Validity 86). Hirsch points out that readers
can make misakes in their assumption of a genre and calls
such wrong guesses extrinsic genres, genres not motivated
by the text. Conversely, the reader's gradual refinement
of the text yields the intrinsic genre (Validity 89). This
distinction is Hirsch’s characterization of the classic
hermeneutic circle: readers guess at the whole to determine
what role a passage plays in the overall interpretation of
a text; yet readers depend upon elements of the text to
35
motivate and confirm that guess. Hirsch's model of
reading, which for him is the process of extracting verbal
meaning, is dependent upon the reader's apprehension of the
intrinsic genre of a text, itself constituted by autonomous
textual elements:
...an idea of the whole must arise from an encounter
with parts. But this encounter could not occur if
the parts did not have an autonomy capable of
suggesting a certain kind of whole in the first place.
A part — a word, a title, a syntactical pattern — is
frequently autonomous in the sense that some aspect of
it is the same no matter what whole it belongs to.
(Validity 76-77)
The process of selecting and rejecting genres is
possible because these genres are part of the shared
knowledge between authors and readers, stable reading
conventions that exist prior to and independent of the
text. In fact, since these genres are independent,
themselves partially constitutive of meaning, Hirsch can
assert that the givens of the reading situation, including
the individuality of the reader, do not determine verbal
meaning (though they may play a role in its significance)
because the essential contextual elements necessary for the
constitution of meaning are part of the intrinsic genre
(Validity 87). Thus for Hirsch, the encounter of reader and
text leaves little room for the negotiation of meaning: a
reader recognizes significant elements of a text which
independently refer to specific genres (part of a social
36
store of knowledge common to author and reader) aiding the
reader in constituting the verbal type — or meaning —
willed by the author.
The determination of meaning, however, is not the only
goal of readers; they must also interpret that meaning by
giving to it a particular significance. Hirsch therefore
creates a complementary distinction between understanding
and interpretation, stating that understanding verbal
meaning is prior to interpretation which he holds to be the
explanation of the significance of verbal meaning (Validity
129). Hirsch thus distinguishes between verbal meaning,
which is determinate, and intepretations, which are
variable. Thus different interpretations may be valid as
long as they seek to interpret an identical verbal meaning;
these differing interpretations are in turn adjudicated by
means of any and all available independent evidence
(Validity 181) and by measuring their consistency with the
authorial intention as signified in the text. Hirsch sees
this adjudication as a practical matter, and acknowledges
that actual correctness may be impossible; nevertheless,
some interpretations, by their accord with available
evidence, will be more probable (and thus more valid) than
others, making correctness at least a goal, if not a
realized possibility (Validity 173).
37
Hirsch's model is related to reader-response models in
that his insistence on the role of authorial intention
questions New Critical and Formalist assumptions about the
independence of texts and re-introduces the activity of the
reader as a specific critical concern in the act of
interpretation. However, this relationship is often
unrecognized because Hirsch insists on the self-identity
and stability of verbal meaning, an insistence confused
with a similar attitude about texts. Fish makes this
mistake in his "Preface" to Is There a Text in This Class?:
"There isn't a text in this or any other class if one means
by text what E.D. Hirsch and others mean by it, 'an entity
which always remains the same from one moment to the next'"
(vii). In the passage Fish cites, Hirsch was actually
referring to verbal meaning (Validity 46), and his position
is more complicated than Fish allows; Hirsch asserts that
the object of interpretation is verbal meaning, dependent
upon the activity of readers, not an independent, objective
text. By following Husserl, Hirsch maintains a certain
objectivity of meaning while acknowledging the role of the
reader in the constitution of that meaning. Hirsch thereby
modifies the subject/object dichotomy in the direction of
reader-response models though he doesn't reject it
altogether.
38
In his critique, Fish might better have focussed on
Hirsch’s insistence on the priority of meaning over
significance. Hirsch cannot explain how readers select
individual elements in a text (and thereby make them
significant) which lead them to particular guesses about
the intrinsic genres constitutive of meaning. This is
partly because Hirsch is interested only in the
after-the-fact adjudication of conflicting
interpretations. But the processes of reading literature
engender alternatives which can lead to dispute (de
Beugrande, "Critical Theories" 541), dispute which is
possibly the result of reader decisions about which
elements in a text were significant for the constitution of
meaning in the first place. Teun van Dijk's concludes from
his work in reading research that theoretical, cognitive,
and contextualized procedures are actualized at different
levels of the reading processs ("Cognitive Processing"
148-9) and that elements of the text are valued and made
significant constantly during the reading process. Van
Dijk's research implies that readers make a myriad of
decisions ranging from local decisions about word meanings
to more global decisions about genres and the expectations
that readers form from them. Considering the great range
and number of these decisions, it would be remarkable if
readers would ever agree exactly about verbal meaning. The
39
process is no doubt dialectical with no clear boundary
between its beginning and its end; but Hirsch’s focus on
prior meaning and after the fact significance distorts its
scope and thereby denies the importance of the temporal
process of reading response.
Hirsch has come to recognize some of the limitations
of his meaning/significance dichotomy. In recent articles
he expands the notion of verbal meaning to include not only
the historically defined willed "type" of the author but
possible future implications that may come to be associated
with that type. Still, Hirsch strongly maintains that
inclusion of future implication in verbal meaning must be
consistent with the willed type of the author ("Past
Intentions" 81) so as to retain the concept of authorial
intention as a stabilizing factor in interpretation. Such
an expansion changes the boundary between meaning and
significance, but Hirsch nevertheless continues to insist
on the basic distinction ("Meaning" 211). Moreover,
sensitive to the objection that his model is exclusive of
social and cultural aspects, he has rightly insisted that
such aspects are already included: the notion of authorial
intention is simply a preferred principle for deciding
which system of cultural conventions should be applied to
the text ("Past Intentions" 80). What Hirsch still fails to
recognize, however, is the relative power of those social
40
and cultural conventions. He maintains that the death of
the historical author proclaimed by Barthes is the
resurrection of the ad hoc author of one's choice ("Past
Intentions" 81), thus reaffirming that the only powerful
constraint upon interpretation is authorial intention. But
Barthes does recognize a powerful constraint, that of
social and cultural "codes," and thereby he questions the
value of Hirsch's distinction between an historical and ad
hoc author.
Unlike Hirsch — and like Fish — Barthes has not held
a consistent theoretical position. During his career he
moved across, without ever completely inhabiting, a number
of different intellectual fashions. At one time he
appeared to be a committed structuralist, interested in a
science of literature, at another time a
post-structuralist, interested in the reader's response to
textual signifiers. However, underpinning all his approach
positions is a semiological means of description derived
2
from the structuralist linguistics of Saussure. As a
result, all of his work pertaining to written language
focuses on the role of readers in the association of
signifiers and signifieds.
The most influential of Barthes' contributions to
reader-centered criticism has been his work on the
activities of readers in reading the novella Sarrasine by
41
Balzac. In "Style and Its Image," Barthes takes a position
opposed to that of New Critics; speaking of a sentence from
Sarrasine, he asserts that it is an interweaving of several
codes — linguistic, rhetorical, actional, hermeneutic, and
symbolic — so that the text is no longer a binary
structure of form and content, "...within it there are only
forms without a content" (5-6). This assertion is
consistent with the semiotic distinction between signifier
and signified; since texts are comprised only of signifiers
(form), they cannot then contain meaning (content), for
meaning depends upon a reader to make the association of
signifer and signified. Barthes summarizes his position by
comparing the text with
...an onion, a construction of layers (or levels,
or systems) whose body contains, finally, no
heart, no kernal, no secret, no irreducible princ
iple, nothing except the infinity of its own enve
lopes — which envelop nothing other than the unity
of its own surfaces. ("Style" 10)
*
To accept Barthes' assertion that there is no
form/content dichotomy within a text, it is necessary to
accept the role of readers in creating content. Such a
position is not inconsistent with Hirsch's position.
However, Barthes is more radical than Hirsch, for along
with the notion of textual determination of content, he
wishes to reject the notion of the power of authors to
determine meaning.
42
Barthes asks of Sarrasine who is speaking, the story’s
hero, the author Balzac, the man Balzac, universal wisdom?
He believes it is impossible to know, for the reason that
all writing consists of several indiscernible voices
("Death" 7). These voices are the product of the social use
of language; that is, Barthes is making much the same point
as M. Bakhtin, that language is not a neutral medium that
"passes freely and easily into the private property of the
speaker’s intentions; it is populated — over-populated —
with the intentions of others" (294). Thus the power to
control the signifying activity of the text does not reside
in the historical author who selects the signifiers, but is
shared between the social conventions outside the text and
the readers who use them to associate those signifiers with
particular concepts. Consequently, the social foundation
of language works with the reader to eclipse the role of
the historical author. In S/Z Barthes develops this
position, by asserting the essential identity of the
activity of writing with that of reading.
S/Z does not present so much a model for criticism
that purports to explain the way readers do read so much as
a demonstration of the way one reader — Barthes — has
read the text of Sarrasine. Nevertheless, in the midst of
this demonstration Barthes is forced to assert a number of
explanatory critical arguments consistent with his interest
43
in semiology and with his insistence on the priority of
form. The most important of these arguments concerns the
relative conventionality of texts. Barthes distinguishes
between the "writerly" (scriptible) text and the "readerly"
(lisible) text. The readerly text Barthes defines as the
classic text, the text whose signifieds are arranged in
such conventional manners that the reader has little power
to control the signification; she is intransitive, left
only with the power to accept or reject the text. His
contrasting category, the writerly, describes texts which,
by their unconventional nature, invite the reader to
participate in their writing; they force the reader to be
active. Barthes explicitly values the writerly text more
highly, because he believes that "the goal of literary work
(of literature as work) is to make the reader no longer a
consumer but a producer of texts" (S/Z 4). Barthes'
distinction is analogous to the distinction Umberto Eco
makes between "open" and "closed" texts; open texts (like
Ulysses) are highly unconventional, requiring,
paradoxically, a knowledgeable and bright reader who is
able to cope with the openness; closed texts (like James
Bond novels) are very conventional, requiring no particular
sort of reader, since the conventions necessary for
understanding the text are accessible to almost all readers
(8-9). But Barthes' distinction actually goes beyond Eco's,
44
for he paradoxically undermines them — even as he asserts
them — by his unconventional reading of a classic text.
For Barthes, the reader that comes to a text is
already a plurality of codes (S/Z 10); reading, then, is
the application of these codes to the text. As Barthes
demonstrates, however, the combination of codes and their
application is idiosyncratic, so that the text itself — or
at least its division into flexible units of reading — is
dependent upon the activity of the reader over time. This
idiosyncratic nature of reading questions the categories of
readerly and writerly that Barthes originally proposed: as
Culler has noticed, the paradox of S/Z is that its
categories explicitly denigrate classic, readable
literature and yet its analysis of the classic novella by
Balzac shows it to be powerful and complex. tfS/Z helps to
produce an intellectual climate in which lovers of Balzac
can try to rescue his novels from an appreciative classical
reading and treat his works as writing that explores its
own signifying procedures" (Barthes 88). Thus S/Z
demonstrates that no text, however "closed," is tolerant of
only one approach; such an anaylsis not only makes of
literature an open category, but it celebrates the
differences among readings. As the repository of codes,
readers create the texts they read; but since the codes are
socially and culturally derived, it is possible that there
45
will be similar features across readings. However, Barthes
warns that the imposition of any originary meaning on a
text is impossible, chiding that researchers who attempt to
equalize texts "under the scrutiny of an in-different
science, forcing them to rejoin, inductively, the Copy from
3
which we will make them derive" (S/Z 3).
Still, S/Z has not become a paradigm for reader
response, which is, at first glance, surprising considering
the model of reading that it implies. It is temporally
based, explicitly forming "lexias" (units of reading given
shape by the reader) in the process of reading, thereby
denying the text a stable form; it accounts well for the
differences in reading, since individual readers will
attend to the text differently and so constitute different
patterns of reading units; and it has a socially derived,
inter-subjective base in the form of codes that account for
agreement among interpretations. But these codes are
problematic in that they are ad hoc; there is no necessary
reason why any particular set of codes should be given
principal power to control readings. The problem is set
aside by Barthes; caught up in the play of signifiers, the
pleasure of the text, in the constitution of Balzac's
novella, he is unconcerned with the basis of the social
constraints he describes. As a result, his descriptions of
them are seemingly unmotivated by any particular or
46
relevant criteria. Thus he convincingly demonstrates how
interpretive disagreements and agreements might occur, but
because the codes that he selects to explain these
phenomena are ad hoc conceptions, Barthes offers,
ultimately, only his own idiosyncratic categories for
characterizing agreement and for judging validity.
Where Barthes and Hirsch are most at odds, then, is in
their focus upon different poles of written discourse.
Hirsch would focus on the intentions of authors and the
social constraints they wish to signify by means of
intrinsic genres; Barthes would focus on the ways social
codes work through readers as they constitute texts. In
common, of course, is their acknowledgement of the role of
these social constraints, and taken together, their
different approaches imply that these constraints are
relatively powerful in determining the discourse
situation. Thus, despite Hirsch's claim that Barthes was
rejecting an historical author to embrace an ad hoc author,
in fact, Barthes demonstrated the power of social codes in
reading literature, merely arguing that readers have the
power to write the text themselves. But with the
recognition of that power comes the need to explain its
form. Hirsch's concept of intrinsic genre makes a
contribution by asserting the organizational nature of some
social constraints; but he conceives of genres as fixed
47
entities that according to Hirsch remain self-identical
across the temporal distance between author and reader.
Such an assertion is inconsistent with the fact that these
genres are internalized by individuals and so operate
idiosyncratically in the reading process. Barthes’ codes
are more flexible and thus more able to explain individual
response; but since he fails to explain why these codes
should be accepted, they are thereby less powerful
explanations of agreement and validity. Thus Hirsch and
Barthes have each contributed to reader-centered criticism
an inter-subjective base by demonstrating the explanatory
power of socially derived constraints upon the reading
process; but their separate concerns, for agreement and
disagreement respectively, have led them to neglect the
description of how social constraints operate. The work of
Fish and Rosenblatt makes a significant contribution to the
task of describing these constraints.
During his career, Fish has moved from a
phenomenological perspective in affective stylistics to a
social perspective which allows him to focus on the social
structures that determine the shape of the reading
process. In "Interpreting the Variorum" he calls these
social structures "interpretive communities," establishing,
through his explanation of their character, a solidarity
48
with Barthes' position on the essential identity of the
writing and reading processes:
Interpretive communities are made up of those
who share interpretive strategies not for
reading (in the conventional sense) but for
writing texts, for constituting their proper
ties and assigning their intentions. In other
words, these strategies exist prior to the act
of reading and therefore determine the shape of
what is read rather than, as is usually assumed,
the other way around. (Text 171)
This position has several advantages over Fish's
affective stylistics model. First, it explains a priori
agreement without depending upon the text as a container of
reading constraints, since members in the same interpretive
community will already share procedures for constituting
texts and their meanings; second, it explains disagreement,
since individuals are free to move from community to
community, adopting new procedures and rejecting old ones
to conform to each community; and third, it implicitly
challenges the subject/object dichotomy, since it assumes
texts to be intentional objects, the products of human
interactions with physical documents. Fish himself is not
always clear about this distinction; Jerome McGann accuses
Fish of letting the word "text" slip its moorings so that
it refers not only to physical documents but also to the
psychologically realized entity that is created by readers
of such documents. Nevertheless, the distinction is
49
crucial to Fish’s model, for the constraints that members
of an interpretive community share are directed toward a
physical document, though this document itself contains
nothing but forms that give readers "the opportunity to
make meanings (and texts) by inviting them to put into
execution a set of strategies” (Text 173). The forms
themselves do not encode meaning, nor do they include the
directions for their own decoding, since they only become
directions to those who already have particular strategies
in the first place; consequently, the significance of marks
on a page, according to Fish, is a function of an
interpretive community, for only its members will imbue the
marks with value (Text 173).
Thus, Fish's newer model of reading is consistent in
many ways with Barthes' model, since the reading process is
dependent upon social aspects (interpretive communities and
codes respectively) and since both models challenge the
objective status of texts by making them a product of the
interaction of documents, social constraints, and readers,
who "write” the text by associating its signifiers with
socially derived meanings. However, Fish sees the
implications of such a model perhaps more clearly than
Barthes. Rosenblatt criticizes Barthes for taking for
granted that a work of art, that is, a distinctively
literary text, is involved in his analysis, leading
50
Rosenblatt to the conclusion that Barthes ultimately gives
too much power to the text by assigning it an essentially
literary or non-literary value (Reader 172-4). Fish remains
more consistent to his model, recognizing that if texts are
the result of reader's applications of socially derived
procedures, than texts themselves inherently have no
literary value. In "How to Recognize a Poem When You See
One," Fish asserts that the act of recognizing a poem — or
any literary document — is not a function of formal
characteristics in the poem that require a specific quality
of attention, but of a specific quality of attention that
compels certain formal features to emerge (Text 326), Fish
illustrates this general point with an example from his own
classroom. While teaching two courses in stylistics, Fish
wrote a reading assignment for his first class on the
blackboard as follows:
Jacobs-Rosenbaum
Levin
Thorne
Hayes
Ohman (?)
These were the names of authors meant to represent the
selections Fish would have his students read; the question
mark after "Ohman" was Fish's way to indicate he wasn't
sure about the spelling of Richard Ohmann's name. After
his first class ended but before his second class could
meet, Fish drew a frame around the assignment and wrote at
51
its top, "p. 43." When his second class arrived, Fish told
them it was a religious poem he had been studying; his
students began to interpret the "poem" as such, offering a
number of justifications for the formal features they
perceived in the text (Text 323-4). Thus the formal
features that distinguish literary qualities are themselves
merely the product of specific activities of perception and
of reading; perhaps more clearly than any previous
reader-centered theorist, Fish strips the text of power in
the discourse situation while retaining an inter-subjective
social base to accomodate a priori agreement and to save
his model from unconstrained subjectivism. Nevertheless,
as Fish recognizes, disagreements do occur over the
interpretation of texts, not least because the membership
of interpretive communities — and thus the rules and
procedures that members share — are constantly in flux.
Adjudicating such disagreements is the greatest problem for
Fish’s model of criticism, specifically because, as Fish
sees, "One cannot appeal to the text, because the text has
become an extension of the interpretive disagreement that
divides them [conflicting interpretations]" (Text 340).
Fish attempts to solve this paradox by making a
distinction between two sorts of critical models,
persuasion and demonstration. The latter is a model
derived from an analogy with scientific inquiry and the
52
procedures of logic where interpretations are confirmed or
disconfirmed by facts independently specified; the former
model, the one which Fish supports, argues that the facts
that one cites are available only because of the activity
of interpretation and have no objective independence (Text
365). Such a distinction is inevitable within
reader-centered models of criticism which assume the text
to be constituted in some manner by the activities of
readers; however, even a model of persuasion must account
for how evidence is judged, and in this respect, Fish's
model is severely deficient.
Which is not to say that Fish is unconcerned with the
nature of critical persuasion or validity. Fish attempts
to account for the compelling quality of some
interpretations on the grounds that some interpretations
are better suited to the rules and procedures for
constituting texts, rules and procedures that readers
derive through their membership in particular interpretive
communities; critics will therefore evaluate
interpretations according to criteria within those
communities to which they belong. Although such a position
recognizes that evidence is only evidence within a
particular frame, it also shifts the grounds of the
argument, because to argue about the validity of an
interpretation is to argue about the validity of the
53
evidence used to constitute it, which is to argue about the
nature of the frame that dictates the procedures and
requirements for evaluating evidence in the first place.
Different interpretive communities will vary in these
procedures and requirements so that the argument is not
between individuals who would support different
interpretations but between different communities that
would support different rules of evidence.
However, such a position is finally inadequate. It
depends ultimately upon accepting Fish's notion of
interpretive community; this notion does have the advantage
of explaining how an inter-subjective social model of
reading might account for a priori agreement, but since
Fish declines to specify how these communities are
constituted, this notion offers no way to justify the
procedures of one community or another. Thus Fish finally
fails to justify any particular means for recognizing
evidence, and therefore he fails to provide any specific
criteria for judging the interpretations that depend upon
the quality of that evidence. As Mailloux observes, Fish's
theory is both revolutionary and inconsequential; it
completely changes the way we understand the activities of
literary study, yet it changes nothing in critical practice
since there is no justification for any particular critical
procedure and thus no reason for critics to alter their
54
habitual methods of criticism (_IC 63). For example, Fish
continues to do affective stylistics in his practical
criticism, though he recognizes that his social model of
reading offers no particular theoretical justification for
continuing to do so. Of course by refusing to specify the
possible nature of interpretive communities Fish seemingly
avoids the problems Hirsch and Barthes face with their
unyielding genres and ad hoc codes. Consequently, however,
Fish refuses the opportunity to give to reader-centered
criticism a truly practical characterization of a socially
derived inter-subjective base. Nevertheless, his theory
goes well beyond those of Hirsch and Barthes by exploring
the theoretical consequences of interpretive communities,
showing the explanatory power of such a concept and its
potential for a theory of criticism.
The work of Stanley Fish is in many ways consistent
with that of Louise Rosenblatt whose interest in the work
of John Dewey has led her to reject the objectivist
paradigm altogether. Thus she rejects, like Fish, the
notion of an objective text, instead preferring to conceive
of the object of interpretation — in her terminology, the
literary work or poem — as a product of the transaction
between the text — a series of signs interpretable as
verbal symbols — and the reader ("Event" 126-7). But
unlike Fish, her "transactional" model foregrounds the
55
distinction between text and evoked literary work. She
sees the text and the reader as two poles which determine
the work of art, asserting the necessity of paying close
attention to the words of a text as well as to what the
words stir up within readers (Reader 137).
All other aspects of her model depend upon this
distinction. Like Fish, she asserts that any text is open
to readings that will evoke literary qualities; these
qualities are dependent upon the way the reader approaches
the text, either as an "efferent" reader — who focuses
attention on public meaning, abstracting what is to be
retained from the reading process and later recalled,
paraphrased, and analyzed — or as an "aesthetic" reader —
whose attention is focused on what is being lived through
as the text is being read, on the cognitive and affective
experience (Reader 101-2). All readings are mixtures of the
two approaches — the categories are never pure. The
efferent focuses on the end result, on the information that
readers create through their transaction with texts; the
aesthetic focuses on the actual experience during the
reading event (Reader 27-8). Regardless of the quality of
that transaction, however, the reader is always engaged in
a creative activity of selection, synthesis, and
interpretation, much of which is below the threshold of
awareness (Reader 52).
56
Moreover, the similarity to Fish's model extends to a
concern about the role of social constraints in the reading
process, Rosenblatt implies that readers share membership
in "cultural communities" (Reader 144n):
Despite the inevitable uniqueness of each life,
readers... may have acquired the language under
similar conditions, had a similar literary training,
read the same books, participated in the same social
milieu, and acquired similar ethical and aesthetic
values. Such a body of readers may thus be able
to communicate easily with one another about their
still, to some extent, diverse individual responses
to a text, (Reader 128-9)
Such a formulation of social constraints anticipates
Fish's concept of interpretive community, and it provides
the same explanatory function: to account for a priori
agreement while allowing for individual differences. But
unlike Fish, she extends the process by explicitly
asserting that members within a community will be more
likely to come to a common judgment about which of the
individual readings within a community is most valid
because of the common values held among the members (Reader
129). Thus, unlike Fish's model where validity was a
function of critics ultimately arguing for and against the
priority of interpretive communities and their procedures
for reading, Rosenblatt's model assumes that the argument
will take place within a community, for other communities
have no common ground from which to disagree. Moreover,
57
Rosenblatt accords the literary critic a special role in
such disagreements, because a critic is ua fellow reader
who earns the interest of others by his special strengths
in carrying out the processes essential to the literary
transaction” (Reader 147). Thus validity of interpretation
is dependent upon a priori social constraints (reminiscent
of Fish’s interpretive communities, Barthes' codes, and
even of Hirsch’s genres) and after the fact adjudication
among readers and critics who share agreement about the
priority of some cultural communities (similar, perhaps, to
Bleich’s concept of negotiation). However, in keeping with
her emphasis on both poles of the reading transaction —
text and reader — Rosenblatt is not satisfied with this
formulation; she extends the inter-subjective base among
readers from the social milieu to the text itself,
recalling the work of inter-subjective critics who focus on
the text as the means of insuring agreement.
Rosenblatt calls the signs to which a reader responds
"cues” (Reader 55) which are of various sorts, including
cues of language — which evoke a semantic code — and cues
of literary convention, which may alert the reader to adopt
the aesthetic stance or construct a particular genre, for
example a poem or play (Reader 55-6). Thus her theory of
reading involves the response to cues, the adoption of the
efferent or aesthetic stance, the development of a
58
tentative framework or guiding principle of organization
(based upon the knowledge readers glean from cultural
communities), the arousal of.expectations based upon the
framework, and the fulfillment or reinforcement of those
expectations which itself leads to a recursive process of
revision (Reader 54).
Thus Rosenblatt builds into her work two
inter-subjective elements that explain agreement and
provide criteria for validity. In a sense, her model
consolidates the close attention to textual elements of
affective stylistics with the social aspects of
interpretive communities, thus realizing the advantages of
both positions -- active attention to the process of
reading and the quality of the textual elements that drive
it, and the guarantee of agreement through a priori social
constraints. Unfortunately, her model also incorportates
the disadavantages of those positions. Specifically,
Rosenblatt gives a good deal of power to the text in the
reading transaction; despite her assertion that efferent
and aesthetic stances adopted by the reader determine the
literary or non-literary quality of the text, she gives
cues in the text the power to trigger one or the other
stance; for example, Rosenblatt says such openings as "once
upon a time" may signal the adoption of the aesthetic
stance, since it is usually a convention of some types of
59
literary genres like fairy tales (Reader 81). Her
characterization of the process of reading makes of the
reader a tabula rasa and it ignores the possibility that
cues themselves may be the products of prior
interpretations. According to Terence Hawks:
Professor Rosenblatt's main difficulty lies in the
rather over-simplified notions of "reader" and "text"
that she so readily takes on board. The presumption
of readerly innocence inscribed in her account of
starting "from scratch with the text on the page'"is
one example of the sort of unrecognized problem that
develops. (566)
Moreover, such a characterization of readers and the
cues with which they interact is at odds with her
formulation of cultural communities; since the membership
of readers in these groups is prior to any encounter with
texts, it necessarily affects the way readers approach any
text for the first time. And finally, as with Fish's
description of interpretive community, Rosenblatt does not
describe the specific rules, conventions, and values that
are constituents of her notion of cultural community.
Nevertheless, Rosenblatt's model is useful because she has
been more inclusive than other reader-centered critics in
her particular combination of the phenomenological and
social aspects of reader-centered criticism, thus giving to
the field a framework that lends some sense of the scope of
what is needed.
60
Taken together, the theories of Hirsch, Barthes, Fish,
and Rosenblatt present a number of common threads that
weave together a basic pattern for a meta-critical theory
of reader-response. All four theories reify the
subject/object dichotomy by focusing on the temporal,
evoked work — Hirsch’s concept of meaning, Barthes’
concept of lexias, Fish's concept of text, and Rosenblatt's
concept of evoked work or poem — and the actual process of
its psychological constitution; all four accommodate
disagreement by acknowledging either the idiosyncratic
relationship of such an evoked entity to the reader or by
acknowledging the idiosycratic strains in its constitution
by the reader; and all four recognize, in various degrees,
the power of socially derived procedures, rules, and
conventions in the reading process, asserting that these
social constraints are prior to the act of reading and
predispose readers who share,them to recognize large areas
of agreement.
Moreover, each of these theories make important
individual contributions. Hirsch's theory argues that not
only readers but authors are constrained and in turn
productively use existing social conventions to create
meaning, so that a theory of reading needs to extend its
scope to include authors as readers in an interpretive
community. Barthes extends this argument by asserting the
61
essential identity of reading and writing by shifting the
focus of power away from authors to the social codes that
readers use to constitute texts. Fish radicalizes the
approach by stripping all power to constrain meaning away
from the text, demonstrating the explanatory power of
social communities from which interpretive procedures are
derived. And finally Rosenblatt incorporates a theoretical
concern for social institutions and their values that guide
the reading process with a practical concern for validity
almost as strong (and perhaps as problematic) as Hirsch’s.
Weaving together these common threads of all four
theories with the strands of their individual patterns,
contributes to a more comprehensive and thus a more
ecological perspective. The New Critics projected the
writer, the reader, and the social environment into the
text, creating an impoverished model of discourse, as if
the text can exist apart from the environment of its
creation. Psychological and inter-subjective
reader-response critics returned the reader to the
discourse situation and acknowledged the temporal aspects
of form; but their focus was largely directed inside the
reader, ignoring the social aspects of the discourse
situation already in place, surrounding and penetrating the
reading (or writing) process. Like the inter-subjective
models, the social models above, with the exception of
62
Hirsch's, do privilege the reader, but that reader is
considered to be a product and a creator of the social
forces that work through her and with her in the process of
constructing texts. Ultimately these models question the
subject/object dichotomy not only by making the text a
temporal product of consciousness rather than its object,
but by showing that all the elements of the discourse
situation — the text (or document), the evoked work, the
reader, and (thanks to Hirsch) the author — are part of a
social environment.
But despite their theoretical promise these social
models of criticism fail through their inability to
describe the structures of the social constraints that
drive them and thus to provide an explicit characterization
of the means by which evidence is selected and valued, of
the means by which readers might choose, in a principled
manner, among competing interpretations of literary works.
Such a characterization needs to accomodate all aspects of
the discourse environment while describing how social
aspects of this environement are realized by readers. The
work of reading researchers and cognitive psychologists in
the field of schema theory suggests a principled means of
describing the way in which social structures are realized
by individual readers, contributing to a meta-critical
theory of reading literary texts.
63
Chapter III
SCHEMA THEORETIC MODELS OF READING:
PROCESS BECOMES DESIGN
Theorists in the fields of education, psychology,
linguistics, and cognitive science who investigate the
precise nature of the reading process have come to similar
conclusions as those literary critics who investigate the
role of the reader in the constitution of literary texts
and their meanings: that textual meaning depends heavily on
the contribution of the reader and the knowledge that
enables readers to process texts. However, unlike a
literary semiologist like Roland Barthes, who takes the
study of structural linguistics as his model for human
understanding, or unlike a literary critic like E.D,
Hirsch, who takes the phenomenology of Husserl as his model
for human understanding, these researchers often depend
upon the computer and the constituents of its programing to
provide them with a paradigm for human comprehension.
Models of reading developed under this influence are
typically termed information processing models; early
examples were usually stage models, with a succession of
events or encoding processes that took place over time and
64
proceeded from some sort of stimulus input at their
beginning to a response output at their conclusion.
Different models emphasized different encoding processes,
but they all assumed some hypothetical events moving along
in time (Gibson and Levin 481). An influential example that
continues to be discussed'*' is Phillip Gough's "One Second
of Reading," a synthesis of empirical work on such matters
as eye fixation and letter recognition; according to Gough,
"...the Reader is not a guesser. From the outside, he
appears to go from print to meaning as if by magic. But I
have contended that this is an illusion, that he really
plods through the sentence, letter by letter, word by word"
(532), Gough's model resembles a flow chart, with discreet
stages and directional vectors mapping the processing
flow. However, it stops short of decribing how readers
actually comprehend what they read: Frank Smith wryly
observes that Gough relegates comprehension to a "mystical
area of the brain called TPWSGWTAU (the place where
sentences go when they are understood)" (231-32). Gibson
and Levin argue that Gough’s model leaves unexplained the
use of higher-order structures that is characteristic of
the economical behavior of adult readers (449). In short,
Gough's model appears to depict reading as a linear,
one-way activity with no component that allows the reader
to assimilate the process of reading to her knowledge of
65
how the world is or might be organized (de Beaugrande,
"Linearity" 47). To use the parlance of cognitive
2
scientists, Gough’s model is exclusively "data driven."
An important alternative to the information processing
approach taken by Gough is Kenneth Goodman's
psycholinguistic model of the reading process. Goodman
opposses the "common sense" notion that reading is a
precise process involving "exact, detailed, sequential
perception and identification of letters, words, spelling
patterns, and large language units" (497). Instead, he
argues:
Reading is a selective process. It involves partial
use of available minimal language cues selected from
perceptual input on the basis of the reader's expect
ation. As this partial information is processed,
tentative decisions are made to be confirmed, reject
ed, or refined as reading progresses.
More simply stated, reading is a psycho-
linguistic guessing game. It involves an interaction
between thought and language. Efficient reading does
not result from precise perception and identification
of all elements, but from skill in selecting the few
est, most productive cues necessary to produce guesses
which are right the first time. (498)
Goodman subsequently proposes a more flexible,
non-linear model of reading which distinguishes among
eleven separate and recursive phases that succeed one
another in no necessary order. Goodman's model is
theoretically more ambitious than Gough's since it attempts
to describe some of the processes likely to be necessary to
66
the explanation of comprehension: the third phase of his
reading model specifies that the reader picks up graphic
cues, guided by constraints set up through prior choices,
his language knowledge, his cognitive style, and the
strategies he has learned; the fifth phase includes a
memory search for syntactic, semantic, and phonological
cues to revise the reader’s selection criteria; the sixth
phase includes semantic analysis to make a decoding guess
consistent with the graphic cues already picked up (507).
Thus for Goodman, reading is a selective, evaluative
process guided and supplemented by the reader’s knowledge.
According to Gibson and Levin, such a model functions as
analysis by synthesis: the reader constructs meaning for
himself as his eyes move over the page, forms hypotheses
about what to follow, and pauses for fixation occasionally
to confirm what he has been predicting. It is a far more
flexible model than the information processing model
proposed by Gough; however, it does not specify how the
reader knows when to confirm his guesses or where to look
to do so (481). Goodman has rejected the information
processing approach (and with it the machine metaphor which
lends that approach its descriptive precision) to realize
one principal advantage: his model is not bound by linear,
one-way processes. Unfortunately, in realizing this
advantage, Goodman is forced to accept a corollary
67
disadvantage: his model loses precision because it has no
way to characterize the knowledge base upon which it
depends, consequently depicting reading as an almost
totally idiosyncratic, almost whimsical process. Goodman’s
model is "concept driven," with no way to show how concepts
constrain the process of reading.
In sum, an information-processing approach like
Gough's is rigid, specifying the same processes for all
readers, regardless of differences in goals, values, and
knowledge; a constructivist approach like Goodman's does
not specify how knowledge is used and how it is
represented, Gibson and Levin argue that a reading model
needs to be flexible, to account for 1) the fact that
reading processes adapt to the demands of the text and the
goals of the reader, and 2) the fact that readers ignore
irrelevant input by processing the largest units that are
appropriate for the reading situation and thus processing
the least amount of information necessary (481-82), A
reading model also needs to account for how readers agree
substantially about the texts they read, and it needs to
account for why, sometimes, they don't agree in both
trivial and substantial ways. Some reading researchers
have attempted to accomodate both the
information-processing and constructivist strains implied
by the requirements of such a model by adapting the work of
68
cognitive scientists and psychologists in the province of
schema theory, which David Rumelhart sees as the most
promising way to talk about how knowledge is represented to
consciousness (Processing 164).
Much of the contemporary work in schema theory derives
from researchers in artificial intelligence (AI) who are
interested in how to program a computer so that it can
understand and interact with the world around it (Schank
and Abelson 1). Understanding the world includes
understanding discourse in the world, and workers in AI
test the ability of computer programs to understand
discourse by feeding their programs short texts and testing
their comprehension through their ability to summarize
those texts. Such tasks require computer programs to
emulate the human conceptual mechanisms that deal with
language; however, according to Schank and Ableson,
cognitive scientists have found such an approach is
impossible without
providing the computer with extensive knowledge of the
particular world in which it must deal. Mechanistic
approaches based on tight logical systems are inade
quate when extended to real-world tasks. The real
world is messy and often illogical. Therefore arti
ficial intelligence... has had to leave such approaches
and become much more psychological. (2)
In a sense, such programs are a later-generation
version of information processing models like Gough's, for
69
they run over some particular period of time bounded by
discourse input and summary output. But unlike such
models, these programs are actively contructivist in that
they incorporate a knowledge component — as it happens,
conceptually consistent with some psychological and
philosophical descriptions — against which the input is
compared and evaluated for its "goodness of fit" to
3
particular components of its knowledge base. Moreover,
such programs provide a new paradigm for reading research
that promises to yield a model consistent with the
requirements set by Gibson and Levin, in part by returning
to the metaphor of human as machine. Of course, such a
metaphor — despite its descriptive precision — is
dangerous to extend. Humans are not machines, and their
idiosyncratic goals and motivations need to be addressed in
any theory of reading. In general, schema-theoretic
accounts do not do so, largely because of their AI genesis;
nevertheless, these accounts do contain, I believe,
potential to overcome their limitations. To demonstrate
how, I present below an explication of the basic tenets of
schema theory as worked out by reading researchers and
cognitive scientists, followed by a critique of schema
theory from a more humanistic standpoint.
The key question for cognitive scientists — and by
extension, for theorists in reading who recognize the
70
importance of cognitive research to a theory of language
comprehension — is the nature of the knowledge base that
serves any intelligent entity, human or computer, capable
of understanding discourse. According to Rumelhart, there
is substantial agreement among cognitive scientists about
the broad outlines of this knowledge base and the processes
of understanding written texts based upon it
("Understanding” 1).^ Although terminology varies — for
example, Marvin Minsky uses the term "frame" (211-212)
which analyzes roughly into what Schank and Ableson term
"scripts" (36-68) and "plans" (69-98) — I will follow
Rumelhart and use the term schema for two principal
reasons: 1) it is the term used most generally by reading
researchers and cognitive scientists to refer to a general
theory of knowledge representation, and 2) it has
historical parallels with a similar use of the term by
philosophers and psychologists ("Building Blocks" 33).
For example, Kant wrote that schemata are products of
the imagination that represent to consciousness generic
concepts underlying particular images:
The concept "dog" signifies a rule according to which
my imagination can delineate the figure of a four-
footed animal in a general manner, without limitation
to any single determinate figure such as experience,
or any possible image that I can represent in
concreto, actually presents. This schematism of our
understanding, in its application to appearances and
their mere form, is an art concealed in the depths of
the human sould, whose real modes of activity nature
71
is hardly likely ever to allow us to discover, and
to have open to our gaze,(B180-81).
Such schemata are not imagistic, but represent instead the
results of pure synthesis whose purpose is to relate
experience to pure concepts. Kant’s notion of schema is
deeply humanistic, since he believed that the formation of
schemata was an art forever beyond our contemporary and
scientistic passion for empirical explanations.
Nevertheless, his concept is a precursor to the work of
early Gestalt psychologists and more particularly to the
work of F.C, Bartlett whose monograph Remembering opposed
the belief that memory involved separate, immutable traces
that represent exact copies of original experience (Durkin
25) .
Bartlett had advanced the view that memories are
constructions of the human mind based upon conceptual
schemata (206). Bartlett was never clear about how the
schemata that.guided memory construction worked other than
to emphasize the importance of concept-driven processing
(Anderson and Pearson 257); however, Rand Spiro points out
that Bartlett’s theory anticipated the major tenets of
contemporary constructivist theories of meaning, including
the belief that language provides only skeletal
representations that must be "enriched and embellished so
that they conform with the understander's preexisting world
views and the operative purposes of understanding at a
72
given time" (245). Thus the schema concept rests on a broad
base of agreement, and researchers in cognitive science are
now attempting to lend to it the precision necessary for a
better model of the structure of general concepts and the
5
procedures for their use.
To understand how, consider the following sentence:
I was brought into a large white room and my eyes
began to blink because the bright light hurt them.
According to Rumelhart, almost without fail people to whom
he showed this sentence believed that it was part of either
an interrogation situation in which the protagonist is
being held prisoner, or it was part of a hospital scene in
which the protagonist was a patient. Further, almost all
of the subjects reported that it was the ’'bright lights” or
the "large white room" which had guided their
interpretation ("Understanding" 5). Since there is no
specific reference to either of these scenarios in the
example sentence, Rumelhart suggests subjects were
consulting their own private theory about the nature of
reality to construct an explanatory setting of the
situation consistent with information that they have
understood from the sentence and subsequently deemed to be
valuable ("Understanding" 3). Experimental results like
these challenge reading models like Gough’s that describe
reading as the sequential processing of letters, words, and
73
sentences, for they indicate that reading also includes an
evaluative element that focuses attention on particular
discourse elements regardless of the stage of processing.
Such results also lend credence to some aspects of models
like Goodman's that describe reading as, in part, a
guessing game, for readers would seem to be required to
guess at or to infer pivotal aspects of the scenario in
order to understand it. According to Rumelhart and Ortony,
schemata have at least four essential characteristics which
lend them the explanatory power to describe such a process:
1) they have variables; 2) they embed one within another;
3) they represent generic concepts which vary in their
level of abstraction; and 4) they represent knowledge,
rather than definitions (101).
Perhaps the most important of these characteristics is
that schemata have variables that may become associated
with, or bound by, different aspects of the environment on
different occasions. For example, any number of people
could be associated with the act of breaking a window —
perhaps a youngster, a homeowner, or a burgler — for
different reasons — to vandalize property belonging to a
cross neighbor, because of forgetting the key to the front
door, or for gaining a quick and stealthy entry into an
unoccupied house, respectively. Moreover, these people
might all use different methods (that is, perform different
74
acts) that require different tools — instruments — for
the breaking: throwing a convenient stone, rapping sharply
with a fist covered in a protective rag, or striking with a
ballpein hammer from the tool box in the trunk of the car.
But no matter what the instantiation of the schema, the
relationships internal to the schema will remain relatively
constant: for a particular cause, some entity will still
break an object by performing a particular act with an
appropriate instrumentality,^
In addition, schemata specify information about the
types of variables that may be bound to them. In the BREAK
schema, possible variables for the instrumentality might
include not only rocks, protected fists, and hammers, but
bricks, axes, and possibly even two-by-fours, but not such
things as automobiles and wicker wastepaper baskets, since
driving an automobile at an object would surely lead to a
wider area of destruction than is necessary or usual to the
controlled act of "breaking," and wielding a wicker
wastepaper basket to break an object, even a pane of glass,
would be extraordinarily difficult for the obvious
reasons. Such constraints on the type of variables that
can attach to schemata serve two important functions: 1)
they tell what sorts of objects might be realistically
bound to each variable; and 2) when there is insufficient
information they can allow good guesses to be made about at
75
least some of the unspecified variables (Rumelhart and
Ortony 103).
Guesses about the nature of unspecified variables,
however, are also constrained by the quality of other
variables specified by the environment. For example, if a
homeowner, unpopular with the youngsters in his
neighborhood, is summoned to his living room by a loud
crash to find a broken pane in a casement window and a
muddy rock lying among shards of glass in his carpet, he
might very well make sense of the scene by recognizing that
the broken glass and the rock satisfy some of the variable
constraints for the BREAK schema. Moreover, in order to
evaluate the goodness of fit between the situation and the
BREAK schema selected to explain it, the homeowner might
feel compelled to guess at (or to infer) the identity of
the culprit, possibly concluding that the culprit was a
neighborhood youngster, since youngsters typically use
rocks to destroy the windows of their offending neighbors,
Marvin Minsky calls the assignment of such inferred
variables the assignment of "default values" (212); these
may be fixed independently of specified variables, but more
likely they are assigned contingently, depending upon the
value of the variables already assigned,
Rumelhart and Ortony make one additional point about
variables and variable constraints: it is rarely the case
76
that a schema never will accept particular variables, but
rather it is useful to think of variable constraints as
representing distributions among possible values (105).
Thus the BREAK schema would prefer bricks and hammers to
baskets and automobiles to fulfill the role of
instrumentality. However, in the absence of other choices
closer to the "average,” a variable constraint may accept a
deviant variable: a panicky person fleeing a bedroom fire
might very well effect an escape by breaking a window with
a wicker basket if it were the nearest, most available
object at hand. Consequently,
the set of variable constraints for a given schema
should be considered to form a multivarient distri
bution with correlations among the several vari
ables....The extent to which a particular schema
fits a particular state will roughly depend on how
probable that particular configuration of vari
ables is for that particular schema. (Rumelhart and
Ortony 105)
Schema theory, then, is a prototype theory of meaning;
that is, inasmuch as a schema underlying a concept (like
BREAK) instantiates our understanding of that concept, our
understanding is in terms of the typical or normal
situations or events which correspond to that concept
(Rumelhart, "Understanding" 3). Thus schemata do not
specify so much a set of necessary and sufficient
variables, but rather they function to specify a set of
standards by which intepreters judge the goodness of fit
77
between variables specified by the environment and
candidate schemata. In short, schemata establish criteria
of sufficiency whereby they may be instantiated by any
number of variable combinations among which no one variable
or set of variables acts as a necessary or essential
component. Such a theory of understanding — of meaning —
is strengthened by parallels in philosophy and
linguistics.
For example, Wittgenstein has demonstrated the
difficulty of trying to find a common essential element
among different activities we term "games.” Chess, noughts
and crosses, tennis, ring-a-ring-a-roses have common
qualities that crop up among some groupings of games and
disapear when other groupings are considered. Thus, ”We
see a complicated network of similarities overlapping and
criss-crossing: sometimes overall similarities, sometimes
similarities of detail” (32e), However, since no single
similarity or group of similarities apply to all of the the
games, Wittgenstein is forced to admit:
I can think of no better expression to characterize
these similarities than "family resemblances"; for the
various resemblances between members of a family:
build, features, colour of eyes, gait, temperment,
etc. etc. overlap and criss-cross in the same way,
— And I shall say "games" form a family. (32e)
Furthermore, Bernard Comrie, in the context of
defining the notion of grammatical subject across
78
languages, has noted that the notion correlates highly with
the idea of the agent in a sentence (the entity with the
most reponsibility for the action described) and the topic
of the discourse (that to which the discourse most
generally refers); but he shows that not all instances of
grammatical subject instantiate either of these concepts
(either singly or together), concluding that "definitions
[of grammatical subject] based on prototypes must be
allowed as a possibility" (104). William Labov — in
contradistinction to the work of Katz and Fodor who
attempted to show that the meanings of words could be
described by universal and conjunctive semantic markers —
has shown that the boundaries of concepts to which words
refer is fluid, made up of an invarient core of properties
that is itself variable in extent and controlled by a
number of different factors ("Boundaries" 366-67).
The prototypical, variable nature of schemata lends to
them another descriptive advantage; they are efficient for
describing any number of contexts, so that it is
unnecessary for human beings to be burdened with a separate
schema to describe each separate situation. Bobrow and
Norman suggest that schemata are normally unambiguous in
the specific context in which they are first formed, but
because they are only minimally specified (that is, they
typically instantiate relationships among variables instead
of constants) they can apply in many different contexts
(133); for example, the BREAK schema does not specify a
particular object nor a particular instrument, so it may
apply to any number of contexts which contribute variables
that serve to instantiate that schema, for example broken
glass and a muddy rock. Moreover, the efficiency of
schemata as vehicles for understanding also emerges because
the variables themselves may be schemata; that is, schemata
embed one within another.
The overall organization that results from such
embedding is hierarchical, so that a model of any
particular schema might include a number of subschemata;
for example, the BREAK schema might require a sub-schema
for CAUSE to represent the motivation for the entity that
breaks the window, presumably because the cause of such an
event links variables in a relationship complex enough to
require a separate schema with a number of its own
variables and subschemata. Such an organization provides a
number of theoretical advantages, foremost of which is a
trend toward economy. Since schemata may take as
constituents other schemata, a smaller number than would
otherwise be necessary is sufficient to combine in the
infinite variety of permutations necessary for making sense
of the world. Moreover, schemata embedding may be
recursive, processing the same variables and subschemata in
80
various roles to instantiate the greatest number of
candidate schemata in order to explain the environmental
input.
A second advantage of the property of embedding is
that a situation or object can be comprehended in terms of
major constituents or parts without necessary reference to
the internal structure of the constituents themselves.
Thus, for example, a face can be thought of as a
certain configuration of eyes, nose, and mouth, etc.,
rather than an enormously complex configuration of
those elementary perceptual attributes upon which the
perception of a face must be presumed to ultimately
depend. Similarly, the structure of a schema allows
us to distinguish between the subschemata and those
that exist between the constituents within any of
those subschemata. (Rumelhart and Ortony 106-07)
Thus, if reference to the internal structure of some
schemata is, theoretically at least, unnecessary for
comprehension, then the process of embedding constrains its
own potential for infinite regress without depending upon
some finite set of "primitive" schemata beyond which
analysis is impossible. At the same time, such a structure
preserves such a possibility through the potential to
accomodate any depth of analysis, depending upon the goals
and needs of the entity attempting to understand its
environment.
The characteristic of embedding introduces the third
characteristic of schemata, that they represent all levels
81
of abstraction. Thus schema theory represents an
improvement upon semantic theories, like that of Katz and
Fodor, that depend wholly upon describing concepts
represented by primitive lexical items. Schemata can
describe conceptualizations at any level of discourse from
the constitution of alphabetic letters to the structures of
narratives and beyond.^ As a result, they can
simultaneously accomodate both data driven processing —
like that of some information processing models of reading
— and concept driven processing — like some
psycholinguistic models of reading. According to
Rumelhart, schema-directed processing procedes in the
following way: some event in the sensory system
"automatically" activates certain "low level" schemata
which in turn, in a data-driven fashion, activate "higher
level" schemata (the most probable ones) of which they are
constituents. The higher-level schemata would in turn
initiate conceptually-driven processing by activating
subschemata not already activated in an attempt to evaluate
the goodness of fit between the original event and the
schema. At some point, when one of the higher level
schemata has further positive results about its goodness of
fit (that is, it has found evidence to accomodate other of
its constituents) it would activate still higher level
schemata that would look for larger constituents. Whenever
82
a schema initiates a search for sensory data that are not
present, this may count as evidence — especially if the
missing data cannot be plausibly inferred — against that
schema and schemata that require the presence of that
schema as a constituent subschema ("Building Blocks" 42).
The fourth characteristic of schemata — that they
represent knowledge rather than simple definitions —
derives from the other three. Since schemata represent
concepts according to prototypical relationships among
variables, since they embed one within another, and since
they describe concepts at varying levels of abstraction,
they are theoretical opposites to definitions that attempt
to describe concepts in terms of essentials or in terms of
inflexible and necessary boundaries.
Consequently, [schemata].,.are not linguistic entit
ies, but abstract symbolic representations of know
ledge which we express and describe in language, and
which may be used for understanding language, but
which are nevertheless not themselves linguistic.
(Rumelhart and Ortony 111)
Additionally, two other characteristics of schemata
emerge from the preceding examples: they are active
processes (a characteristic necessary to accomodate their
variable, prototypical nature), and they are recognition
devices whose processing is aimed at their goodness of fit
to the data being processed (Rumelhart, "Building Blocks"
41). Consequently, schemata "represent knowledge in the
83
kind of flexible way which reflects human tolerance for
vagueness, imprecision, and quasi-inconsistencies"
(Rumelhart and Ortony 111). Such flexibility is essential
to a model of reading that itself tolerates differences
among readers’ interpretations of texts. But just as
important is that the prototypical nature of schemata
guarantees a good deal of agreement among readers as well.
Again consider an experiment reported by Rumelhart
("Understanding" 6—7) to illustrate how readers process
text, specifically the following fragment (I have numbered
the sentences in the following passage to make subsequent
references to the text easier) :
[1] Business had been slow since the oil crisis.
[2] Nobody seemed to want anything really elegant
anymore. [3] Suddenly the door opened and a well-
dressed man entered the showroom floor, [4] John
put on his friendliest and most sincere expression
and walked toward the man.
Despite the brevity of the fragment, Rumelhart’s subjects
constructed a relatively clear and consistent
interpretation of the story, John is a car salesperson,
fallen on hard times, who probably sells large, elegant
cars. A good sales prospect enters the showroom where John
works, and, in order to make a good impression and
subsequently sell the man a car, John tries to appear
friendly and sincere. He walks over to the man to begin
his sales pitch.
g
Rumelhart infers from the reading protocols of his
subjects that such an interpretation is not arrived at all
at once; instead, subjects construct their interpretations
as they read. The first sentence is usually interpreted to
mean that business is slow because of the oil crises, which
frequently leads to the formation of hypotheses involving
either the selling of gasoline or the selling of cars. The
second sentence tends to disconfirm the gasoline-selling
hypothesis since the quality of "elegance” rarely applies
to gasoline. On the other hand, the car-selling hypothesis
fits well since people tend to prefer smaller, less
elegant, and more economical cars in oil-scarce times. The
third sentence generally supports this hypothesis since the
phrase "showroom floor" strongly suggests a typical scene
wherein cars are often sold. Expectations generated by the
car-selling hypothesis are confirmed by the appearance of a
potential buyer in sentence (3) and of "John" in sentence
(4) who is a candidate for the role of seller; his behavior
is, moreover, stereotypical of a car salesperson readying
himself to make his pitch, Rumelhart asserts that this
interpretive process is qualitatively consistent with a
general theoretical account of comprehension described by
schema theory:
The process of comprehension is very much like the
process of constructing a theory, testing it against
the data currently available, and as more data becomes
85
available, specifying the theory further — i.e.,
refining the default values (as perhaps was the case
when those holding the "car hypothesis" from the
beginning encountered the sentence [2] about nobody
wanting anything elegant anymore). If the account
becomes sufficiently strained, it is given up and a
new one constructed or, alternatively, if a new theory
presents itself which obviously gives a more cogent
account, the old one can be dropped and the new one
accepted.... These theories are, of course, schemata.
("Understanding" 7)
Additionally, a schema-theoretic model organizes the
protocol results more precisely: the first sentence
suggests the notion that business was slow which, in a
data-driven fashion, activates a schema about business and
the economy; the business schema includes a variable for
the type of business, which is constrained by the
information about the oil crises (perhaps part of a cause
variable of the economy schema). At this point, the
business schema would activate a concept-driven search for
possible oil related businesses, suggesting as possible
candidates gasoline selling and automobile selling.
Information in the second sentence matches well only the
latter choice, which is futher confirmed by the later
appearance of likely candidates to fulfill the variables of
car buyer and car salesman.
The results of experiments like this one reported by
Rumelhart are useful to confirm general hypotheses about
how readers make use of knowledge during the process of
constructing meaning from written discourse; however, these
86
results are also useful because they tend to substantiate
more formal schema-theoretic models of reading
comprehension such as Kintsch and van Dijk’s wherein
different levels of textual representation are more
precisely specified. Their model takes as its input
propositions (setting aside the problem of their
derivation), which they describe as concepts composed of a
predicate and one or more arguments, which may be concepts
or other related propositions fulfilling different semantic
9
functions such as agent, object, and goal (367). Thus,
their desription of "proposition" is consistent with that
of "schema," including the properties of taking variables,
embedding one within another, representing different levels
of abstraction, and representing knowledge and not
linguistic definition. Moreover, not only are the input
propositions essentially schematic in structure, but as a
result, so is everything derived from them: the
microstructure, the structure of individual propositions
and their relations; the macrostructure, the synthesized,
global characterization of the discourse as a whole (365);
and the over-arching schema (Kintsch and van Dijk reserve
the term for this special use), a high order conventional
structure such as a story (narrative), an argument, or a
psychological report (366). Moreover, schemata control the
derivation of the macrostructure from the microstructure by
87
means of macrorules which take the general form of
deletion, generalization, and construction rules
(consolidating the microstructure by means of deleting
unnecessary information or compacting information through
inductive and deductive processes respectively). Despite
the unique terminology, Kintsch and van Dijk*s model is
nevertheless a schema-based model of reading, albeit a more
precise model than the general account offered by
Rumelhart, and like Rumelhart's account, theirs depends
heavily on the process of inference to construct a theory
to account for the textual data, for example, to solidify
the microstructure by adding the propositions necessary to
make it referentially coherent (367).^^
According to Anderson and Pearson, the inferences that
J schema-theoretic descriptions of reading like Kintsch and
jvan Dijk's (or Ruraelhart’s for that matter) depend upon are
ithe product of four different processes. The first is the
I
process of deciding which schema to instantiate; the second
is the process of infering which variables remain
uninstantiated within a schema (to direct attention to the
constraints that control those variables); the third is
assigning a default variable; and the fourth is drawing
conclusions based upon a lack of knowledge (that is,
instantiating an explanatory schema without specific data
to activate it) (269). Moreover, as more inferences are
88
made, choices become more and more constrained according to
the requirements of the emerging model of the text (Collins
et al. 387) .
However, schema-theoretic descriptions of reading
often characterize the most local levels of inferencing in
terms of a priori abstractions, for example input
propositions (Kintsch and van Dijk) or sentences
(Rumelhart), leaving aside the problem of how these
abstract structures come to be constituted by readers. The
result is that such descriptions of reading are
incomplete.^ But such a problem is the consequence of
experimental interest rather than theoretical limitation.
Rumelhart has been most concerned with the description of
high-level schema such as stories, and Kintsch and van Dijk
are concerned with overall strategies of comprehension.
Additionally, since reading researchers generally believe
that readers abstract the "gist" of texts by constructing
highly conceptualized theories to account for textual data
(Sanford and Garrod 67), the trend is naturally toward
describing these global conceptualizations. Furthermore,
because the higher level schemata that different readers
instantiate from the same text are likely to display the
most disagreement, since these levels presuppose a number
of different opportunities to make inference choices that
will lead different readers, more or less, down a number of
89
theoretical, and perhaps exclusive, garden paths. Thus as
a focus for experimental inquiry, higher level schemata
will likely display more interesting features by
simultaneously agreeing and disagreeing in substantial
ways.
Nevertheless, schema theory can accomodate much more
local structures than sentences and propositions.
Rumelhart and Ortony have proposed that schemata may
incorporate primitive elements beyond which processing is
impossible (106-7); and Adams and Collins have suggested
more specific schemata for more focal aspects of written
discourse. Specifically, they suggest that data-driven
(they use the term bottom-up) processing insures that
readers will notice novel information in the text as they
look for data to fill out potential schemata at the same
level or further up. Examples are letters that suggest
neighboring letters and words and determiners like "the"
that suggest an accompanying noun (20). At these levels,
the inference processes seem much more constrained because
the schemata are less abstract and thus more specific about
variable constraints; however, at these levels there are
potentially a greater quantity of schemata to be
assimilated, and potentially a greater number of choices
that can lead readers away from agreement, choices that
jinclude not only which candidate schemata to instantiate
90
and which variables to attach, but also which schemata to
ignore as unimportant within the context of the reading
situation. Choices at these levels are important, for they
may lead, eventually, to wildly different interpretations;
that they often don't is a function, however, of how narrow
the choices actually may be at any given time.
However, the disagreements that inferencing processes
seem to guarantee only partly explain why readers differ in
their interpretations of texts. A more basic question for
schema theory is what explanations it can provide to
account for the fact that some readers fail to understand
some texts or parts of texts at all. Rumelhart sees three
reasons:
1. The reader may not have the appropriate schemata.
In that case he or she simply cannot understand
the concept being communicated.
2. The reader may have the appropriate schemata, but
the clues provided by the author may be insuffi
cient to suggest them. Here...the reader may not
understand the text but, with appropriate addi
tional clues, may come to understand it.
3. The reader may find a consistent interpretation of
the text but may not find the one intended by the
author. In this case, the reader will "under
stand" the text but will misunderstand the author.
("Building Blocks" 48)
The third reason implies not so much a failure of
understanding but a failure at achieving a normative
understanding constrained by the will of the author, This
reason also reveals Rumelhart's definition of what
91
constitutes sufficient understanding of a text, and I leave
discussion of this matter for later. The second of these
reasons — that readers may fail to use schemata that they
do possess — is cited by Betty Holmes in her study of how
prior knowledge contributes to the ability of readers to
answer questions about passages they have read. She
concludes that "Poor readers with more prior knowledge
[than other poor readers] did not use their stored
information to help them modify incorrect information or
learn new information as well as good readers with more
prior knowledge" (17),
As it is characterized here, schema theory has several
advantages as a model of conceptual knowledge, the most
important of which is its flexibility, the quality to
conform to any number of specific situations and by doing
so to offer a concept to organize them for the
understanding. But, as Andrew Ortony points out, this
flexibility is an empirical stumbling block; that is,
schema theory is so general that it readily lends itself to
ad hoc accounts of almost any phenomenon, so that schema
theory may be too general and too vague to have any
interesting testable consequences. Ortony concludes,
therefore, that the best that can be hoped for is that the
conjunction of such a_d hoc accounts is consistent if not
compelling, so that schema theory at least remains capable
92
of accounting for empirical phenomena (359). Spiro agrees,
asserting that most work on schemata is conjectural without
attempts at experimental verification, though some
empirical studies argue for their conformity to the
prevailing paradigm, for example, the work of Holmes (256).
The consistency of schema theory, despite its broad
generality, is imposing. Sanford and Garrod, writing
specifically about Minsky's concept of "frame," a
progenitor of contemporary schema theory, summarize the
appeal of such descriptions of conceptual knowledge:
They provide a good mechanism for the phenomeno
logical experience of recognizing something as
familiar....Whatever the shortcomings of partic
ular implementations, they provide a connection
between knowledge and the procedures for using
knowledge which is bound to lead to useful
structural formulations, rather than scholastic
descriptions. (37)
However, the general flexibility of schema theory may
lead one to ignore the fact that schema theories of reading
are derived from machine models. Witness the terminology
that proliferates: slots, default values, frame systems,
terminals (Minsky 212), data-driven processes,
concept-driven processes (Bobrow and Norman 148), etc. The
bottom line of this sometimes explicit sometimes implicit
human/machine comparison is that schema theoretic accounts
of reading are nevertheless information processing accounts
of reading wherein data is input into the reading "machine"
93
which processes it and gives back some sort of summary
output, certain and invariable in character. The
implication of such a metaphor is that no matter how much
the machine has to accomplish to achieve its output, that
output is necessarily the result of the interaction between
a specific and stable set of input data and an independent
set of schemata instantiated to match it.
The machine metaphor implicit in schema-theoretic
accounts of reading has several advantages, first among
these is the precision with which it describes the
structure of knowledge. Because AI researchers interested
in discourse are forced by the requirements of their
machines to specify knowledge in terms of specific
structures that include information and instructions for
its use, they have been able to offer conjectures about the
requirements of such structures for human comprehension
based upon the recognition that whatever the differences
between humans and machines, they both share some
requirements for storing and using conceptual knowledge,
requirements that chiefly include the need for economy of
variables and processing strategies. Such an advantage is
prerequisite to a second: the machine metaphor describes
readers as actively participating in the construction of
meaning through inferencing constrained by candidate
schemata, contradicting older reading models that depict
94
the process of reading as a linear process of extracting
meaning from text. And finally, this second advantage
seems to contain within it a third: that since meaning is
held to be a construction of readers instead of an a priori
given, it resists the compelling and seductive appeal of
what Richard Bernstein calls the "Cartesian anxiety” —
which presents the perhaps implausible choice between some
fixed foundation for our knowledge or intellectual and
moral chaos (18-19).
But this last advantage is a deceptive one, for schema
theoretic accounts of reading don’t so much deny the power
of the Cartesian dichotomy so much as shift its ground.
While meaning is held to be a construction of active
readers, the input text itself, in most schema theoretic
accounts, appears curiously stable, whether it be in the
form of propositions derived from a discourse or in the
form of sentences that comprise a discourse. As noted
previously, such stability is partly the result of an
experimental bias, for the machine metaphor requires a
stable input. Consequently, the question Stanley Fish
refuses to answer — the question of what an interpretation
is an interpretation of — is assumed to be answered.
But schema theory itself undermines this answer.
Since schemata are constitutive of concepts at all
different levels, even down to and beyond the local level
95
of perceiving letters and their relations, the assumption
that a text provides a stable input is counter-intuitive,
Sullivan and Rabinow assert that "interpretation begins
from the postulate that the web of meaning constitutes
human existence to such an extent that it can not ever be
meaningfully reduced to constitutively prior speech acts,
dyadic relations or any pre-defined elements" (31), A
consequence of this position is that meaning is a rational
product of rational processes that are, at least partly,
constrained by human motivation. As a result, the "input"
for any model of human understanding most likely should not
be assumed to take a form that is the consequence of prior
acts of interpretation. Jonathan Culler, of course, has
argued for the practical value of deeming some such
structure to be the object of interpretation
(Deconstruction 74); but whatever the practical value of
such a need, theoretically a model of reading needs to
explain the constitutive activity of readers who may make
of physical documents any number of texts according to
their idiosyncratic goals in particular environments.
Fish's demonstration that readers may make over a simple
reading assignment into an aesthetic object like a poem is
a convincing demonstration of the power of readers to
redesign the supposedly a priori form of textual "input."
96
Moreover, the machine aspect of schema theory tends to
downplay the environmental aspects of understanding in
another way: it tends to depict the understander as
separate and apart from the context within which
interpretation takes place. As Rosenblatt points out, such
a model implies that the text and the reader interact, but
then remain isolated ("Viewpoints" 99) so that the text
itself has no particular impact upon the reader other than
as a field of discourse to be processed, and the reader has
no impact upon the text other than as an entity that will
do the processing. To criticize the machine metaphor on
these grounds is actually to criticize the concept of
memory that underlies it, specifically the concept that
human memory is a container of discreet and stable traces
of past experience, or as Roger Schank as characterized it,
"a repository of information" that provides knowledge to
the inference making process (10),
Bransford and his colleagues argue that such a
characterization is not particulary fruitful; they assert
that the static nature of "spatial" or "container"
metaphors fail to reveal the flexibility of memory.
Instead, they propose a more ecological view, reminiscent
of Bartlett, by asserting that remembering is partly a
*
reconstructive process that occurs within an impinging
environment (462-63). Rand Spiro agrees, writing that
97
rather than stored traces of past experience as the heart
of remembering, the past is inferentially reconstructed
from stored fragments of the past and knowledge of the
world; thus he concludes that reconstructed memories —
based upon the assumption that the past is orderly — may
be inaccurate, but when errors and distortions make a
coherent whole, individuals are unable to distinguish them
from accurate memories.
In sum, Spiro and Bransford et al. assume that
remembering is set in a specific context of human
motivation and human ability. Such a position implies that
the nature of schemata that are "stored" in memory are
themselves fluid, so that many of these — perhaps the most
abstract and complex — are reconstructed according to the
impinging environments that readers find themselves in and
according to their own motivations within these
environments. And such a position implies that the process
of reading, requiring the instantiation of schemata and the
formation of inferences to accomodate the textual input,
will itself affect the those schemata and vice versa; that
is, schemata may be partially instantiated, fully
instantiated, or radically changed at any given time, not
only because some texts seem to defy interpretation without
such changes, but also because of the motivations and
abilities of readers who may reconstruct high level
98
schemata in a variety of different ways. Such an advantage
— or some may say such a disadvantage — is not implied by
a machine based model of reading, since machines process
input wholly because of the motivations of others who also
see to the stability of the schemata that machines store.
A corollary disadvantage to the machine metaphor is
more closely related to its AI genesis and stems from the
reliance of AI researchers on conceptual dependency
theory. This theory holds that for any two sentences with
identical meaning, there should be only one representation,
and that any information in a sentence that is implicit
must be made explicit in the representation of the meaning
of that sentence (Schank and Abelson 11); thus schema
theory implicitly assumes that non-identical sentences may
have identical meanings since they are organized by a
single representation. This position is consistent with
Noam Chomsky's early models of syntax and even the case
grammar of Charles Fillmore, since all of these models
assumed that a single "deep structure" may underlie
sentences which seem to represent the same sequence of
action and refer to the same arguments, active and passive
12
forms of sentences, for example. But such a position,
like the linguistic models it parallels, ignores the
reasons why sentences have passive and active forms in the
first place, among which are to preserve the topic
99
structure of a discourse by moving the object of a sentence
into the subject position, thus attempting to adjust the
structure of the discourse to the perceived needs of the
addressee (Chafe 54). The differences between active and
passive sentences are not merely formal; they are
meaningful in that they serve to bind the structure of a
discourse more tightly and convey, at the very least and in
the case of written language, information to a reader about
the cooperativeness and the intentions of the author.
Thus, insofar as schema-theoretic accounts of language seem
to be dependent upon the representation of discreet
concepts that underlie specific input, they are inadequate
to address the pragmatic issues of how discourse is used;
as Wittgenstein has observed, language is an activity and
meaning is best understood in light of what words — and by
extension all levels of linguistic organization — do
(18e).
Finally, the machine metaphor leads to a view of
language use that is particularly strict in its assumptions
about correctness. Goodman attempted to ameliorate this
problem in his psycholinguistic model of reading by using
the term "iniscue" instead of error, since he felt that such
a term avoided unfortunate value identifications. Yet even
a term like "miscue" implies some sort of normative
standard and thus assumes that deviations from this
100
standard are undesirable. With Rumelhart’s
characterization of schema theory, however, the norm
becomes more explicit: he asserts the possibility that a
reader may find a consistent interpretation of a text, but
not the one intended by the author (see 91 above). Such a
position partly derives from the machine metaphor because
it assumes that the text intended by an author takes some
stable a priori form as input which will necessarily
require the reader to use some particular set of schemata.
Yet as literary critics like Stanley Fish and Roland
Barthes have argued, the power of authorial intention to
constrain readers is not a settled issue nor is it clear
that authorial intention should be the primary factor in
gauging the validity of an interpretation. Rumelhart's
bias for authorial intention as an interpretive norm seems
to me to be the product of the general focus in reading
research on "efferent" reading (to borrow Rosenblatt’s
term), reading most concerned with the construction and
interpretation of information qua information. But such a
bias is problematic in that it ignores other potentially
legitimate means of characterizing the differences in
understanding among readers; it does not account for the
goals and motivations of readers who are not reading merely
for information and may thus be legitimately engaged in an
interpretive activity wholly at odds with an author’s
101
intentions. If understanding information is the main goal
of a reader, then reconstructing the intentions of an
author may play a pivotal role in the reading process. But
if some readers, instead, are more concerned with immersing
themselves in the ambiguities of language, with enjoying
the freeplay of signifiers, and with using the text to
construct and inhabit "a world that doesn't have to conform
to the accepted real world" (de Beaugrande "Critical
Theories" 537), then the specific intentions of an author
may be of less moment. In any case, the legitimacy of any
textual interpretation is a function of several factors —
including but not limited to — the intentions of authors.
To assert otherwise is to give importance to an intepretive
norm by fiat instead of by arguing for the necessary
primacy of such a norm.
In sum, the machine metaphor has several advantages
for the characterization of conceptual knowledge and its
role in the reading process; chief among them is that it
recognizes the need for precision and for accomodating the
temporal aspects of meaning by emphasizing the process of
its construction. But as Shaw and Bransford recognize,.the
machine metaphor responsible for the information processing
approach is limited because it typically gives short shrift
to questions concerning the nature and origin of the
information processed, how significant information is
102
selected, and how its value might be determined (4). I
hasten to add, however, that such a limitation does not so
much imply that the machine metaphor is wrong, or even that
it hasn't been useful, for clearly it has been a pivotal
instrument for the development and understanding of the
requirements for a model of conceptual knowledge and for
reading. Nevertheless, I wish to suggest that it falls
short of being able to describe some important aspects that
intuitively seem necessary to augment such a model. Thus,
the problem is not so much a problem of schema theory
itself as a problem of transcending the stubborn
explanatory metaphors that aided the genesis of schema
theory and continue to be used in its explication. For
schema theory to transcend its machine metaphor, however,
entails substituting a more congenial metaphor than that
which currently prevails, not only to insure that the
limitations of the old analogy are forsaken but to provide
a new, positive way to see schema in the context of human
understanding. The substitution of a new metaphor would
not be for the purpose of revising or contradicting the
basic tenets of schema theory, but instead for the purpose
of putting those tenets to work in new ways. The principal
drawback of the machine metaphor was its inability to
accomodate the goals and motivations of readers who may be
said to construct the text itself in terms of the schemata
103
they use to organize the textual variables presented in a
particular environment. I propose as a substitute a
metaphor of design: readers are thus not seen as processing
machines of independent texts but as designers of texts
built from the raw materials of the environment and guided
by the conceptual schemata that they deem to be appropriate
to their personal reading goals.
I believe that the positive implications of the design
metaphor for schema theory remedy all of the major
drawbacks of the machine metaphor. The "input" of the old
model would become a field of raw material for the new
model, so that readers could be seen as evaluators of their
environment who are motivated to select those materials
they deem to be necessary for their own understanding; the
schemata that they possess act as a priori evaluative
guides in terms of presenting concepts that organize
material for the understanding and providing criteria by
which readers can select and arrange that material. The
schemata themselves may be seen to be fluid, changeable
according to readers’ goals and to the material to be
organized and interpreted. Moreover, the Cartesian duality
is effectively collapsed, since the materials of the text
are intricately connected to the activities of readers who
participate in its design and hence the construction of its
meaning. And finally, the issue of interpretive
104
correctness is better addressed as an issue of validity,
partly judged in terms of what a community of readers deem
to be useful in the design of the text and its meaning and
which may include criteria according to the perceived
intentions of authors. Such a characterization in no way
contradicts the major tenets of schema theory as outlined
by Rumelhart; in fact, it does them justice by conforming
them to a more ecological perspective that relates human
understanding to the environment — social and otherwise —
in which it takes place. Host importantly, however, such a
characterization promises above all to preserve the
descriptive precision of schema theory, which now promises
to become a useful adjunct to theories of literary
reader-centered criticism that lack such descriptive
precision but nevertheless allow the flexibility to
accomodate the vagaries of aesthetic response.
105
Chapter IV
INTERPRETIVE SCHEMATA AND LITERARY RESPONSE
In an article entitled "Transaction Versus Interaction
— A Terminological Rescue Operation," Louise Rosenblatt
recently took issue with an ERIC/RCS Report by Karl Koenke
that compares Rosenblatt's transactional view of reading
with information processing approaches. Koenke divides
information processing into bottom-up theories, top-down
theories, and interactional theories wherein "comprehension
is an interaction between the processing of the text and
the use of the reader’s experiences and expectancies," and
he asserts that "It is at this point that the information
processing theory and the literary criticism theory of at
least one recognized authority, Louise Rosenblatt, seem to
find common ground," because, according to Koenke,
Rosenblatt's theory describes the relationship between
reader and text as one in which the elements of a total
situation are conditioning others and being conditioned by
others within a framework of a transaction (116).
Rosenblatt writes of this comparison:
...I must admit that I began to feel somewhat uneasy
when information-processing and literary
criticism were presented as cognate theoretical
106
sources. And the impulse to write the present paper
was triggered when I found my transactional theory
being said to share "common ground" with information
processing, and "top-down" and "bottom-up" processing
being shown as squatters on that common ground under
the rubric of "interactive processing"...."Inter
action" and "transaction" seem to be accepted as
interchangeable. (96)
Rosenblatt argues that the difference between the two
terms is basic: transaction designates an ongoing process
in which all of the elements of the reading process are
seen as aspects or phases of a total situation (98);
interaction designates a mechanical concept in which the
text and the personality of the reader can be separately
analyzed (100). In effect, Rosenblatt differentiates the
terms by means of showing the differences between the
metaphors upon which they are based, the organic,
ecological metaphor on the one hand and the machine
metaphor on the other. However, by insisting on the
differences, Rosenblatt obscures what is in fact common
between some theories of literary criticism and some
information processing approaches — the constitutive role
of the reader — and consequently implies that the field of
literary criticism, because of its more humanistic roots,
has little to gain from information processing theory.
Rosenblatt also observes that the separation between
literary criticism and the field of "’reading' per se"
reflects the deficiencies in the theoretical underpinnings
of both groups: "The reading experts ignore what is usually
107
called 'literature,1 and the literary folk, starting with
an agreed-upon canon of 'literary works,' usually ignore
the problems of the reading of 'ordinary' prose and how it
differs from 'literary' reading" (101). Thus the
differences that separate the two fields of inquiry are not
defined necessarily by theory and practice so much as by
their respective objects of study, a point which takes much
of the sting out of Rosenblatt's earlier complaint that
literary criticism and information-processing should not be
regarded as cognate theoretical sources. In contrast, Teun
van Dijk assumes the necessity of taking into account
cognitive processes like those described by some
information processing theories in order to understand
literary interpretation; he holds (at least for his own
model of discourse processing) that the processes and
principles of discourse comprehension have a general nature
so that in principle they hold for literary as well as
extra-literary discourse ("Cognitive Processing" 151).
The real problem of conflating the theoretical
terminology of "transactional" literary criticism and
"interactive" or "information processing" reading research,
then, is not that these fields don't share common
theoretical ground, nor even that thay address mutually
exclusive objects of study — since both assert the
constitutive powers of readers and thus both make of
108
literature a potentially open category. The problem, as
Rosenblatt shows, is that they are developed in light of
competing and mutually exclusive metaphors. However,
despite the alien mechanistic genesis of information
processing approaches to reading — here I am speaking
specifically of schema theoretic approaches — there is no
good reason why such approaches need be limited by their
founding metaphor; in fact, as shown in the previous
chapter, that metaphor can be tapped for its precision and
grafted upon a more organic and more congenial metaphor of
design, consequently ameliorating Rosenblatt's real
concern. I believe that schema theory thus provides
literary criticism with a new means by which to describe
the nature of a priori constraints to which readers appeal
for guidance during the process of interpretation. But, to
understand how requires a glance back at the strengths and
weaknesses of reader-centered literary criticism to see how
at least one critic, Steven Mailloux, has offered a more
precise description of the social constraints operative in
literary interpretation and to see how schema-theoretic
accounts of reading might be subsequently used to extend
such a description.
Recall that the four reader-centered critics whose
theories have made productive use of the concept of a
priori inter-subjective constraints — Hirsch, Barthes,
109
Fish, and Rosenblatt — have nevertheless failed to specify
these constraints convincingly. Hirsch's concept of
intrinsic genre, designed to explain why meaning is an
objective and stable construct, becomes much too rigid when
the phenomenon it is designed to explain — the objectivity
of meaning — is denied by critics, like Barthes, who
convincingly argue for the power of readers to construct
meaning idiosyncratically; but Barthes’ explanation for
this phenomenon, that readers use a priori "codes” in
idiosyncratic ways to construct texts, is flawed by the ad
hoc nature of the codes upon which he believes readers
depend. Rosenblatt’s description of how readers evoke
literary works is, in a sense, intuitively satisfying since
it champions the role of the reader while avoiding the
problems of specifying inter-subjective structures that are
over-determined on the one hand or ad hoc on the other.
Yet her theory achieves this distinction largely by
refusing to grapple with the specific nature of
inter-subjective reading constraints, a problem shared by
Stanley Fish’s concept of interpretive community.
Fish’s theories of reading literature have received a
good deal of critical attention, in part because they
explicitly focus attention on the importance of elusive,
temporal aspects of form and divert it from what had
hitherto been conceived as stable elements of spatial
110
form. However, since Fish strives to show how readers are
constrained by the interpretive communities to which they
belong, his theory is more conservative than it would seem,
for readers will largely read in ways already specified by
their communities. But as Mailloux has shown, Fish's
theory is finally inconsequential since he has refused to
specify the exact nature of the a priori constraints shared
by interpretive communities. Consequently, Fish is unable
to assert exactly how different interpretive communities
select and use evidence to justify their approaches to
literature, thereby making it impossible, within a Fishian
paradigm at least, to argue for, or even to understand, the
regulative qualities that underlie any particular critical
practice.
Fish had avoided such a problem in his earlier theory
of "affective stylistics" by characterizing the constraints
that act upon readers more explicitly. He developed the
concept of the "informed reader," someone who
(1) is a competent speaker of the language out of
which the text is built up; (2) is in full pos
session of "the semantic knowledge that a mature
...listener brings to his task of comprehension,"
including the knowledge (that is, the experience,
both as a producer and comprehender) of lexical
sets, collocation probabilities, idioms, profess
ional and other dialects, and so on; and (3) has
literary competence. That is, he is suffi
ciently experienced as a reader to have internal
ized the properties of literary discourses, includ
111
ing everything from the most local of devices (fig
ures of speech, and so on) to whole genres.
(Text 48)
The informed reader is a much more normative concept than
interpretive community, since it implies much more strongly
what readers of literature should be and what they should
know. But as Fish later recognized, the informed reader is
a value laden concept designed to "take account of, by
stigmatizing, all those readers whose experiences were not
as I described them" (Text 22).
The two positions Stanley Fish has taken describe two
extremes available to reader-centered critics striving to
understand inter-subjective reading constraints: one can
either describe them as normative structures that a reader
must use to productively interpret literature and thereby
assume their objective value, or one can give up the
project and explain differences in terms of theoretically
equal reading communities, each of which values different
properties of discourse. The former is little more than an
assertion that some reading standard should exist (and it
does little to define what exactly that standard must be,
other than to imply that an informed reader will adhere to
the same standard as Stanley Fish); and the latter is
little more than a recognition that different standards do
exist and that the groups that adhere to these standards
require different combinations of reading activities.
112
Jonathan Culler, perhaps a more cautious theorist than
Stanley Fish, has defined a middle ground in Structuralist
Poetics through his own characterization of literary
competence. If Culler is no more precise than Fish, who
uses the same term to characterize one aspect of the
informed reader, he is nevertheless more hopeful about the
ability of theories of criticism to be so. In short,
Culler is hopeful that structuralism will "formulate the
rules of particular systems of convention rather than
simply affirm their existence" (265). He follows Barthes in
asserting that to read a text as literature is not to
approach it without preconceptions, but instead, to bring
to it an implicit understanding of the operations of
literary discourse, "the conventions by which fictions are
read" (113-14). Literary competence, therefore, is a set of
conventions of which readers must be in command if they are
to read and understand literary texts (118).
Culler’s concept has much in common with Barthes’
concept of codes — since these may be seen to be literary
conventions — and with Hirsch’s concept of intrinsic genre
— since Culler sees genres as conventional functions of
language (136). But more importantly, the concept looks
forward to the specification of inter-subjective structures
shared by members of Fish's interpretive communities
without committing itself to the role of defining what
113
those structures necessarily should be. Culler's concept
of literary competence, therefore, avoids the normative
force of Fish's informed reader and preserves the
possibility of argument, negotiation, and agreement
concerning the nature of literary conventions among both
members of interpretive communities and among interpretive
communities themselves. Nevertheless, Culler's concept is
unspecified and so in practice not yet more useful than
Fish's unspecified interpretive communities. However, the
idea that a priori reading constraints can be defined as a
system of conventions has inspired Mailloux to speculate
about the nature of literary conventions and to attempt a
more precise description of reading literature.
Mailloux does not uncritically accept Culler's
assertion that the institution of literature is defined
solely in terms of reading conventions. Instead, he claims
"that it is not just reading conventions that constitute
literature; it is the conventional framing of a discourse
as literature plus the reading conventions (used vigorously
and together) that the framing brings with it" (JIC 136).
This important critical move allows Mailloux to preserve
the notion that literature is somehow convention bound
while avoiding a problem implicit in Culler's view of
literature: that his definition does not distinguish
literature from the institution of writing in general.^
114
Instead, Mailloux has approached the same problem in terms
of how discourse is "framed"; thus his approach
accommodates the objections of critics like Mary Louise
Pratt, who reject the distinction of literature from
non-literature in terms of essential characteristics by
showing the role of social and professional institutions in
"defining and bringing into being the institution we call
literature" (117-118); and his approach is consistent with
Rosenblatt's assertion that the differences between
literature and non-literature are best understood in terms
of the stance of the reader, aesthetic or efferent. Thus,
for Mailloux, literature is an elastic category, defined by
conventional framing activities of readers within
historical communities, a fact with two important
consequences:
First, what is framed may change from culture to
culture, from period to period, and the hierarchy
of reading conventions brought along with the
frame may change as certain conventions achieve
priority and others lose status. Second, the
successful act of framing is communal, not
individual. (IC. 136)
Mailloux calls the hermeneutic constraints that
control this framing "interpretive conventions": shared
ways of making sense of reality, communal procedures for
making intelligible the world, behavior, communication, and
literary texts. These conventions can be as simple,
115
according to Mailloux, as instructions to a baby sitter —
"When Mary Nell cries, it means she wants her bottle" — or
as complex as Augustine's rule of faith — "if a scriptural
passage 'seems to commend either vice or crime or to
condemn either utility or beneficence,' it should be taken
as 'figurative' and 'subjected to diligent scrutiny until
an interpretation contributing to the reign of charity is
produced'" (IC 149),
However, this account of convention — derived from
Culler and fueled by Fish's concern for a communal,
intersubjective base — does not yet add to Culler's
virtual account. To accomplish this, Mailloux proposes a
typology that is designed to "particularize the dynamics"
of interpretive conventions (IC 150); he thus asserts that
there are three kinds of conventions that make the
production of meaning possible in different ways (IC 151).
Traditional conventions are "habits of art" that include
modal and generic conventions; "for instance, symbolism,
realism, and romanticism are modes that can be manifested
in the lyric poem, the historical drama, and the war novel;
genres are conventional categories of literary works" (IC
130). Traditional conventions can sometimes function as the
second kind of convention, regulative conventions' which
stipulate what should and should not be written or read;
classical and neoclassical insistence on purity of kinds
116
and decorum within genres are examples, as are rules of
propriety, for example prohibitions against explicit sex
(IC 133). Finally, constitutive conventions include those
that make a text count as literary, an open set of
conventions defined by historical communities (IC 135-36).
Mailloux asserts that the hermeneutic function of all three
types of convention is the same, but they make the
production of meaning possible in different ways:
Traditional conventions allow interpreters of an
action to fit that act into the context of a
mutually believed regularity described by the
convention. Interpreters use regulative conven
tions either to infer the reflexive intention of
someone who obeys or disobeys a binding rule or
to make sense of the motivation for an action
that accords with the behavior sanctioned by the
convention. Constitutive conventions define (and
therefore describe and regulate) institutional
actions that would not be possible (or identifi
able) apart from the conditions specified by
the convention. (IC 151)
Mailloux thus has pushed beyond the virtual category left
open by Culler's definition of literary competence to
specify the nature of interpretive conventions and to
assert something about their use for understanding
literature. Yet his contribution is limited by the
vagueness of his definition of convention.
Mailloux asserts that the term convention, in its
widest sense, refers to "shared practices." But the most
interesting members of his typology, regulative and
117
constitutive conventions parallel John Searle's distinction
between two sorts of rules: Searle defines constitutive
rules as those which constitute an activity the existence
of which is logically dependent on the rules (34), while
Mailloux defines constitutive conventions as descriptions
determining (present) meaning (jCC 127); moreover, both
Searle and Mailloux use the same example to pin down their
definitions, rules that define "touchdown'1 in American
football. Although Mailloux is clear that conventions are
not ipso facto rules (though they seem to function the same
way), the fact that he includes rules in his examples
obscures the distinction between conventions and rules for
which he has strived. Mailloux cites the definition of
convention developed by the philospher David Lewis:
Conventions are regularities in action, or in
action and belief, which are arbitrary but perpe
tuate themselves because they serve some common
interest. Past conformity breeds future conform
ity because it gives one a reason to go on conform
ing; but there is some alternative regularity which
could have served instead, and would have perpetuated
itself in the same way if only it got stared.
("Languages" 4)
Elsewhere, Lewis distinguishes convention from other
social, regulative concepts such as social contract, norm,
and rule by emphasizing that conventions are arbitrary,
remaining in force only because members of a population
expect that other members of that population will abide by
118
them (Convention 88-107). Not all rules are conventional,
says Lewis, because, unlike conventions, rules may be
enforced by sanctions so strong that one would have a
decisive reason to obey them even if others do not
(Convention 103). Nevertheless, Mailloux is primarily
interested in those conventions that "can loosely be
translated into descriptive ’rules' that account for
conventional behavior and belief" (JX 128). Since Lewis has
demonstrated the overlap between the concepts of convention
and rule (Convention 100), technically Mailloux's apparent
conflation of the terms is permissible. But conflating the
terms in this way tends to obscure those social aspects of
interpretive regularities originally highlighted by the
notion of convention: that conventions are powerful because
everyone in a population wishes to adhere to a regularity
because others do. My complaint is not that there are no
interpretive rules, but that the notions of rule and
convention need to be kept distinct to accommodate the fact
that interpretive regularities that are created by and that
form interpretive communities vary in force and character.
More importantly, however, a convention based theory
of literary criticism must necessarily accept the
2
principles upon which the notion of convention depends.
Drawing eclectically from different fields, Lewis defines
the concept of convention by relying specifically on
119
principles of game logic and coordination problems to
demonstrate how conventions might arise independent of
specific agreements among members of a community. Richard
Grandy summarizes Lewis' approach:
Consider a game in which two players, Row chooser and
Column chooser each choose an action, A, B, C, or D,
and where the relative payoffs depend only on the two
actions chosen according to the following diagram,
where Row's payoff is shown in the lower left corner
and Column's in the upper right:
A B C D
1 4 1 1
1 4 0 0
4 1 0 3
4 1 1 3
1 1 2 1
1 1 2 1
1 3 1 3
1 0 0 3
If Row chooses A and Column chooses B, then neither
player could have obtained a better outcome by a
change in either one's choice. These points
are called coordination equilibria.... A game
with two or more proper coordination equilibria
is a coordination problem. (130-31)
Lewis considers these problems pivotal in explaining
the nature of conventions, for he asserts that players will
prefer, once they recognize the mutual advantages of
choosing the particular rows and columns that lead to
120
coordination equilibria, to persist in choosing these same
rows and columns. "This requires no communication — only
the recognition that neither of them can unilaterally
improve his position (Convention 131). Thus arise
conventions, regularities in action, or in action and
belief that are arbitrary and serve some sort of common
interest. Propp's morphology of Russian folktales, for
example, is an analysis of the specific generic conventions
that readers use to understand different tales — such as
that a tsar giving the hero an eagle, an old man giving the
hero a horse, or a sorcerer giving the hero a boat (19-20)
coordinate with a single function, to specify the means by
which the hero is transferred to a designated place where
the object of the hero's search is located (50).
But if Lewis' analysis highlights the function of
conventions and explains their genesis, it also reveals a
principle of conventions that is often overlooked: their
relative rigidity and their intolerance of variation. By
this analysis, the generic conventions of Russian folktales
not only allow a reader to coordinate particular elements
with particular explanatory functions, but they also guide
a reader to reject all other functions, however useful and
productive they may seem, that are not conventional
because, as Lewis' game matrix reveals, conventions specify
categories with precise boundaries. Moreover, Lewis notes
121
that conventions are arbitrary and thus open to change,
accounting for this phenomenon by showing how members of a
community may come to recognize that some regularity other
than the one historically accepted as a convention could
provide members of that community with a better mutual
"payoff" and become a convention. Thus once a regularity
becomes conventional, it takes both time and community
recognition to change its regulative force. Lewis* game
matrix is a prescient metaphor: the coordination of column
and row leads to a topocentric set of inelastic
regularities whose precisely specified boundaries are
beyond the reach of single individuals to emend.
The point is important, for a theory of reading
literature needs to explain how readers individually
evaluate whether or not specific works adhere to culturally
accepted regularities. Culler points out that such a
process is a necessary means of "naturalizing the text," of
giving it a specific place in the world which our culture
defines, of leaving no chaff, of making everything wheat
(Structuralist Poetics 137). But if we accept the point
that conventions are intolerant of variation — and I
believe that Lewis' demonstration compells us to do so —
then conventions are insufficient to explain the process of
"naturalization" Culler speaks of for two reasons.
122
First, as Lewis has shown, conventions define exact
boundaries that easily accomodate, to extend Culler's
metaphor, the wheat but cannot stretch to accomodate the
chaff. Propp's analysis of Russian folktales is a good
example, for he asserts that a folktale specifies a number
of necessary functions that form the conventional
categories by which folktales are recognized. Thus his
analysis is sensitive to the similarities he discerns
within the genre, but it does little to explain how some
elements are fit into the genre when readers discern that
they cannot reduce them into any conventional function —
there is no place to put the chaff within his scheme, no
way to accomodate difference unless that way is to assert
that such elements are unconventional and meaningful in
that way. Such an explanation is useful as far as it goes
(and it goes a long way toward explaining how, for example,
readers recognize the unconventional nature of a novel like
John Hawkes' The Cannibal which seems to violate
conventional expectations that narrative fiction be clear
about its chronological form), but it nevertheless leaves
open crucial questions: for instance, how difficult must a
difficult task be before it is no longer difficult enough
to define a folktale? Thus, ultimately, many of the
categories that define conventions of literature may be
considerably less clear-cut than they at first seem.
123
Second, in order for conventions to explain these more
variable categories, they would need to proliferate to
accomodate any number of possible variations of generic
regularities, since they themselves tolerate little
variation. And by such proliferations they would become ad
hoc and lose explanatory force. Should a reader encounter
a work that in all respects seemed to be a folktale with
the exception of not specifying, for example, a ’’villain”
or a "difficult task," then that reader would be forced,
since conventions are too rigid to accomodate the anomoly,
to create a new category for the work. Since conventions
are topocentric — a folktale either specifies the function
of "villain" or "difficult task," or by convention it is no
folktale — they seem to define a set of unsatisfying
binary categories: in or out; conventional or
unconventional; good or bad. I make this last point not to
imply that literary critics like Mailloux — and by
extension other critics like Culler who use notions of
convention to ground their interpretive theories — don't
appreciate that genres are elastic categories with fuzzy
boundaries, but rather to illustrate the difficulty
inherent in depending chiefly upon the notion of convention
— often a useful and elegant means by which to explain
some aspects of linguistic regularity — to pin down the
social constraints that control interpretation. As Marilyn
124
Cooper argues, Mailloux's solution to the problem of what
grounds interpretation — a constitutive hermeneutics that
sees texts, readers, and authors as the products of reading
conventions — is incomplete, because although conventions
certainly play a role in interpretation, they simply define
texts in terms of interpretive practices, which they leave
unexplained (13).
Mailloux's achievement is substantial; he has taken
important steps to fill out Culler's unspecified category
of literary conventions. But the limitations of Mailloux's
approach are equally important, pointing out what yet needs
to be done to describe the interpretive regularities shared
by communities. A more satisfying description will thus
fulfill three requirements: first it will preserve
distinctions among conventions, norms, rules, and even
explicit agreement so that there is no necessity to
conflate terms in order to appropriate qualities inherent
in one term to fill out another. Second, it will explain
how individual readers can vary the boundaries of
interpretive regularities in order to achieve their own
goals and accomodate interpretive products that don't
easily "fit" within conventional descriptions that specify
a reductive set of binary choices, for instance, in or
out. And third, it will be consistent with other aspects
of reader-centered criticism -- for example the collapse of
125
the Cartesian dichotomy and the emphasis on the temporal
aspects of form — about which there is broad agreement
among reader-centered literary theorists.
Schema theory — subject to the criticism of the
previous chapter — fills this need by providing a
ready-at-hand description of socially derived regulative
structures consistent with the major characteristics of
reader-centered literary criticism. Despite the fact that
some theorists now recognize that literary and
extra-literary texts share requirements for reading
strategies, as I have pointed out earlier, literary critics
have been slow to adapt reading models derived from work on
extra-literary texts to their own models of reading
canonical literary works. Nevertheless, some critics have
used schema theory to explain how readers understand some
aspects of difficult literary texts.
For example, George Dillon borrows the notion of
schemata to show how order is imposed by readers upon both
literary and non-literary works, recognizing that recent
work in language processing and cognitive science suggests
a high-level model of comprehension — schema theory —
that is useful to explain "certain experiences of texts"
(51). He divides schemata into those that give form or
structure to experience and those that interpret its
content (51), consequently implying the existence of a
126
form/content split such as that implicit in Hirsch's theory
of objective meaning where different linguistic signifiers
are held to instantiate the same meaning. But Dillon,
aware of this problem, qualifies the distinction:
The distinction between formal and content-oriented
schemata probably cannot be drawn in an absolutely
categorical fashion, and for our purposes it need
not be. For one thing content schemata may have
internal structures that can be described in formal
terms, as for example the general/particular,
permanent/fleeting, universal/particular
employed by Johnson in the ... Preface to
Shakespeare. (52)
Thus Dillon recognizes that content may impose its own
order, and, inversely, that formal schemata may be imbued
with content. His overall point is that these "content”
schemata are useful for understanding and organizing
literature, pointing to the example of how a schema for the
concept "golf" will enable a reader to comprehend the
beginning of William Faulkner's The Sound and the Fury.
Although Dillon doesn't develop this example, I wish to
consider the opening of Faulkner's novel to clarify how
schemata work in this context and to develop an
enlightening comparison with another reader-centered model
of criticism. Faulkner writes:
Through the fence, between the curling flower spaces,
I could see them hitting. They were coming toward
where the flag was and I went along the fence. Luster
was sitting in the grass by the flower tree. They
took the flag out, and they were hitting. Then they
127
put Fhe flag back and they went to the table, and
he hit and the other hit. Then they went on, and I
went along the fence. (1)
Faulkner presents this scene through the eyes of
Benjy, the idiot son of Jason and Caroline Compson. Benjy
presents information in terms of tangible observations left
unorganized by any explicit concept; as a result, the
activities that Benjy describes appear engimatic and
confused — as if they were expressed by a naive child.
Moreover, the sentences depend upon coordinate structures,
so that the reader discovers no clues to the relative
importance of information by searching for patterns of
subordination. Confusing as this passage seems, especially
since Faulkner places it at the beginning of the novel
where the character of Benjy is not yet developed, it
immediately becomes comprehensible when the reader realizes
that Benjy is describing golfers. The golf schema — which
includes the flag marking the hole, the action of hitting
the ball, and the removal and replacement of the flag — is
later filled out by the appearance of a caddy (a word Benjy
characteristically mistakes for his sister’s name, Caddy)
who searches for a lost ball (18). The interpretation of
this passage, then, depends upon a reader who can bring to
it the content schema necessary to organize and understand
aspects that seem unorganized by Faulkner’s presentation.
128
In Barthes’ analysis of "Sarrasine" in S/Z, as with
the analysis of Faulkner’s passage above, aspects of
Balzac’s story are seen to be enigmatic, challenging the
reader to bring them together, to organize them, and
finally to thematize them to form some interpretation.
Barthes demonstrates this process using five codes that he
believes penetrate both the text and the experience of the
reader. But Barthes' codes can just as well be seen as
high level content schemata; that is, once the reader
realizes the castrato singer La Zambinella and the old man
— who seems to hold the key to the mystery of the Lanty’s
wealth — are the same person, the reader can organize the
story according to a schema defining the operatic castrati,
a hermeneutic schema, or code, that includes wealth in
exchange for a peculiar sacrifice.
Furthermore, Barthes’ other four codes seem to
function the same way. For example, in one lexia (Barthes*
shorthand for a unit of reading) Barthes discerns two
examples of the action code, to narrate and to hesitate to
tell (S/Z 88), both of which might be just as well
understood as schemata for linguistic activities that share
a common characteristic — that they define some human
action. The seiic code, which provides cultural
stereotypes, the symbolic code, which governs symbolic
interpretations, and the referential code, which provides
129
more general cultural information (Culler, Barthes, 84) can
be regarded the same way. Moreover, in this sense, the
golf schema used by Faulkner may be just another instance
of a hermeneutic code, since it provides the means by which
readers may organize and understand the opening passage of
The Sound and the Fury. But the point I wish to emphasize
here is that these schemata perform a heuristic function
for the interpretation of literature in much the same way
as they do for the interpretation of extra-literary
discourse — they bring together and organize seemingly
disparate elements and by so doing create and maintain
criteria for recognizing how aspects of text not yet
encountered may fit into an overall interpretation. Dillon
says that the type of discourse one is engaged in bears
with it conventions of the kind of schemata that should
prove relevant in constructing a particular text (53); a
productive rereading of Mailloux's analysis that describes
literature as the conventional framing of a discourse as
literature plus the reading conventions that the framing
brings with it, for schemata are more responsive than
conventions to variation. According to Dillon, "We may
think of schematizing as aiming for a 'best fit' of schema
to details of text, but in practice we may settle for
relatively good or less good fits" (53).
130
However, this analysis of the role of "content"
schemata in the interpretation of sections of The Sound and
the Fury and "Sarrasine" only begins to show the potential
role of schema-theoretic models of reading in the
performance of literary interpretation. The analysis is
incomplete in several ways that limit its usefulness as a
model for criticism: it does not address, except naively,
the question of what constitutes a text and its meaning; it
leaves unexamined the precise role of authorial intention
in the rhetorical transaction; and it ignores the question
of the reader’s role and motivation in the constitution of
these elements within a particular environment. In short,
these content schemata have been presented almost as ad hoc
constraints. What is thus needed, then, is a more careful
merger between schema theory and reader-centered literary
criticism in order to show how schemata may offer a
principled means by which interpretive constraints may be
described and the activities of interpretive communities
may be understood.
Though a merger of theories across disciplines
requires a great deal of common ground, fortunately
schema—theoretic models of reading and reader-centered
models of literary criticism share a great deal. The most
obvious — and for reader-centered criticism the most
necessary — aspect that both models share is the
131
assumption that readers constitute meaning, that meaning is
not contained within or conveyed by a text, or more
3
precisely, by a document. This position is perhaps most
clearly expressed in Fish’s work, especially in his
demonstration that readers can imbue even an assignment on
a chalkboard with literary qualities and thus with special
meanings. The constitutive role of readers is central, as
well, to Rosenblatt's concept of the "evoked" poem, to
Barthes’ demonstration of reading Balzac, and even to
Hirsch's assumptions about intrinsic genres and how readers
use these to understand meaning. And the position is just
as important to schema-theoretic models of reading.
According to Adams and Collins:
A fundamental assumption of schema-theoretic approaches
to language comprehension is that spoken or written
text [read document] does not in itself carry meaning.
Rather a text only provides directions for listeners
or readers as to how they should retrieve or construct
the intended meaning from their own, previously
acquired knowledge. (3)
Following from this fundamental assumption is a second
that unites both schema theory and reader-centered literary
criticism: that as a consequence of the constitutive
activity of readers, the Cartesion dichotomy is collapsed
since there can be no stable subject/object relationship in
which to ground reading activities. Reader-centered
literary theory explicitly denies the objectivity of
132
meaning and the text, going so far as to assert that some
formal textual qualities hitherto believed to be objective
and stable characteristics are at least partially the
result of reader's activities. Hence the text of an
"evoked" poem is partly a psychological construct and
consequently temporal in nature. Moreover, this position
is emphasized by reader-centered theories of literary
interpretation such as Fish's that hold that the act of
perception itself is constructivist, so that the text
itself is given form by the act of reading, a position
stressed by Barthes in his analysis of "Sarrasine," wherein
he claims that the reader is "no longer a consumer but a
producer of the text" (4). Schema theory accomplishes the
same thing, albeit implicitly, by insisting on the
contribution of readers to the psychological construction
of meaning, even though most researchers who work in schema
theory, for example Rumelhart and Adams and Collins, don't
go so far as some literary critics who explicitly assert
that the text is partly a psychological construct, making
the distinction between document and text of less moment.
However, as I have shown previously, schema theory easily
adapts to the more radical constructivist position when the
mechanical metaphor which founds it is forsaken for a more
human metaphor of design, for since schemata can be seen to
be fluid, changeable according to a reader's idiosyncratic
133
goals within a particular environment, the materials of a
document are better described as the raw material that is
designed by a reader into an interpretable text.
Finally, reader-centered literary criticism and schema
theory are united by their assumption that form has a
temporal dimension. As Iser makes clear, interpretations
of literary works take place over time so that readers form
images from the process of reading texts against a
background of prior images, thus forming contradictions and
contrasts (Act 148). Iser believes that interpreters of
literature adjust their interpretations as they read in
much the same way that reading researchers believe that
readers eventually form the "gist" of a text by filling
out, revising, and rejecting schemata as they seek higher
level representations of the meaning of a text.
In sum, reader-centered literary criticism and
schema-theoretic models of reading seem to share three
theoretical assumptions: that readers constitute meaning
and texts, that there can be no subject/object dichotomy in
which to ground descriptions of reading (since readers are
actively engaged in constitutive acitivities that take
place during, not prior to, the act of reading), and that
form — as a necessary result of these two previous
assumptions — has a temporal quality. The fact that some
literary critics and some reading researchers share common
134
assumptions is, however, an insufficient foundation for a
merger of their two outlooks. To complete the necessary
groundwork, I turn to the problems of language and
convention as explored by philosopher H.P. Grice.
In his most recent discussion of the notion of
meaning, Grice hazards the supposition that meaning is not
essentially connected with convention. ’’What it is
essentially connected with is some way of fixing what
sentences mean: convention is indeed one of these ways, but
it is not the only one” (238). Grice — adopting the
standpoint of the speaker of a language — consequently
proposes an alternative way to fix the meaning of
sentences, a notion he calls the "optimal”:
The general suggestion would ... be to say what a
word means in a language is to say what it is in
general optimal for speakers of that language to
do with that word, or what they are to make of it;
what particular intentions on particular occasions
it is proper for them to have, or optimal to
have, (239)
Grice's notion of the optimal defines meaning, therefore,
as a rational construct based upon the activities of
speakers — and by extension readers — faced with the task
of evaluating language in their own environment according
to currently prevailing circumstances and according to
their own particular goals. However, any regulative entity
that would fulfill the function of Grice’s
135
context-sensitive optimal would need to fix what language
means without recourse to context-independent standards.
Schemata provide a concise description of such an entity
since they respond to individual circumstances.
Unlike conventions which define a binary set of values
that ignore differences among environmental contexts,
schemata instead define criteria for sufficiency that are
sensitive to context; that is, schemata, which specify
variables and variable constraints, may be instantiated by
a reader according to whatever number of variables that
seem to her to be necessary within a particular context.
For example, even though the opening section of The Sound
and the Fury specifies a limited number of variables — the
flag, the action of hitting, and the removal and
replacement of the flag — these may nevertheless be
sufficient for a reader within a particular reading
environment and motivated by idiosyncratic goals to
instantiate the golf schema, which in turn would specify
other variables that a reader may look for to confirm the
choice, such as the caddy that appears later. Notice that
no one variable is essential to the instantiation of the
schema, but a certain number, itself variable, combines to
establish a sufficient reason, a sufficient threshold by
which to evaluate the relative "goodness of fit.11 In a
sense, the golf schema defines, for this example, the
136
optimal means for most readers by which to organize the
evidence of the text, itself a function of lower order
schemata interpretable as rules, conventions, and norms
shared by speakers of English. At the same time, these
schemata aid the reader in making sense of the work in
different ways.
Pratt draws upon Grice’s earlier work on the logic of
conversation to show how Faulkner establishes the narrative
point of view for this section. In ’’Logic and
Conversation,” Grice postulates that participants in a
discourse depend upon a ’’Cooperative Principle" (45) which
holds that participants in discourse expect others to make
their contributions neither more nor less informative than
is necessary, make contributions that are believed to be
true, make contributions that are relevant, and make
contributions which avoid obscurity or ambiguity. Grice
calls these expectations maxims of Quantity, Quality,
Relation, and Manner, respectively (46-47). If one or more
of these maxims is flouted by a speaker, and if other
participants expect that that speaker is nevertheless still
attempting to be cooperative, then the flouting of a maxim
may be meaningful in that participants will assume that the
speaker is attempting to implicate something (Grice calls
this a conversational implicature) by intentionally failing
137
to observe a maxim. One of Grice's examples will clarify
the procedure:
A is writing a testimonial about a pupil who is a
candidate for a philosophy job, and his letter reads
as follows: "Dear Sir, Mr. X's command of English
is excellent, and his attendance at tutorials has
been regular. Yours, etc." (Gloss: A cannot be
be opting out [of the cooperative principle], since
if he wished to be uncooperative, why write at all?
He cannot be unable, through ingnorance, to say
more, since the man was his pupil; moreover, he
knows that more information than this is wanted
[i.e., he has failed to observe the maxim of
quantity]. He must, therefore, be wishing to
impart information that he is reluctant to write
down. This supposition is tenable only on the
assumption that he thinks Mr. X is no good at
philosophy. This, then, is what he is impli
cating. (52)
Pratt uses these principles to assert that Faulkner is
flouting, in the opening passage of The Sound and the Fury,
the maxims of quantity and manner, since the character
Benjy, who serves as narrator for this section, does not
provide words like "golf," "putt," "tee," or "course"
(184). Pratt concludes that the contrast between what Benjy
says and what we might have expected him to say leads to an
implicature: "Faulkner...is implicating the speaker of the
story has some cognitive or perceptual impediment, that
this fact is relevent to our understanding of what follows,
and that he intends us to share, contemplate, and evaluate
Benjy's view of the world and contrast it with our own"
(184). Thus it is not Benjy who intentionally flouts the
138
maxims (although he unintentionally fails to fulfill them),
but Faulkner who flouts them and thereby invites the reader
to speculate on the reason. Pratt pins down her analysis
by comparing Benjy’s narrative style to data of William
Labov that shows that pre-adolescents relate personal
experience more completely than vicarious experience,
concluding that one would need two different grammars to
describe the differences (185-86). Thus she asserts that
the key to our response to Benjy’s speech is in our own
cognitive past (186), presumably because we could conclude
that Benjy’s narrative technique is that of a child.
Grice points out that the calculation of the presence
of an implicature presupposes an initial knowledge of the
conventional force of the expression of the utterance which
carries the implicature (57); but we have seen, Grice now
prefers his notion of the optimal to that of convention,
which in no way lessens the importance of his point that
language users depend upon some regularity to fix meaning
and by extension calculate implicatures. Pratt ties her
analysis to the notion of grammars, which have the same
intolerance to variation as convention. A more satisfying
account might be to accept Grice’s later intuition, and
subsitute the more flexible and context-sensitive criteria
specified by the optimal. Thus in place of two separate
grammars, we could substitute one narrative schema — with
139
its variables of point of view, chronology, agency, etc.
— more or less successfully instantiated according to the
abilities and cognitive and cultural development of the
narrator. The advantage of such a substitution is that it
observes the necessity of characterizing the regulative
contraints upon language (which lends language users a
foundation from which to calculate implicatures), while at
the same showing that foundation as flexible and
context-sensitive.
Nevertheless, conventions remain, as Grice
acknowledges, one way to fix what language means, and a
theory of reading literature must still accomodate those
aspects of literature that are conventional, that are
arbitrary yet compelling in their regulative force. One
way to accomplish such a task while remaining consistent to
the schema-theoretic model of reading is to use terms like
convention — as well as other terms that define socially
based regularities such as rule, norm, and explicit
agreement — as modifiers that describe qualities of
schemata. Such a critical move allows for the consistent
description of regulative, a priori concepts in terms of
proto-typical schemata while making explicit the
differences among different schemata in terms of regulative
force. Consequently, there is no need to conflate terms
140
like convention and rule in order to capture the nuances of
both types of social regularity.
For example, the detective story may be defined in
terms of a conventional schema that includes variables that
specify a victim, a detective (or someone, who along with
the reader fulfills such a role by showing interest in
and/or solving a crime), clues, a weapon, a scene, and
often suspects. The crime itself serves as the central
focus that organizes these variables and motivates the
actions of the main characters. Raymond Chandler's The
High Window is easily understood in terms of such a schema,
even while the crime that initially motivates the actions
of the detective character Phillip Marlowe — the theft of
the Brasher Doubloon — leads him to uncover another sort
of crime wherein the victim has become the perpetrator.
Vladimir Nabokov's Pale Fire, on the other hand,
instantiates the conventional schema of detective fiction
in a different way. The gunshot death of victim-poet John
Shade haunts the narration throughout the novel, yet it is
not, as in most detective stories, seemingly the center of
attention, nor the motivation for the actions of the main
character, the narrator Charles Kinbote, who tells the
story of himself and consequently of John Shade by means of
the footnotes that masquarade as critical commentary on
Shade's final poem, ''Pale Fire." The seeming adherence of
141
both novels to a schema that defines detective fiction
seems to link two disparate works to a common regularity.
Yet the conventional quality of the detective novel schema
is fulfilled much better by the Chandler novel, since the
force of convention sets a high threshold for the goodness
of fit between the variables noticed by the reader and the
generic schema that might organize them. On the other
hand, even though readers of Nabokov’s novel may notice
many variables fulfilling the same function, the detective
novel schema may seem inadequate to organize all of the
high level variables and sub-schemata that readers will
notice when reading Pale Fire, for example its seeming
adherence to a schema for ’’critical edition," including an
"authorized" text, footnotes, and commentary. As a result,
readers may be faced with a number of alternatives, among
which are to continue to read the story as a detective
novel (a possible alternative since schemata may be emended
to conform to standards dictated by individual readers) or
to consider that Nabokov is flouting the conventional
detective novel schema. Thus, the conventional nature the
of the detective story schema says something about its
regulative force, but its proto-typical nature lends it an
attractive — and playful — flexibility.
However, some schemata, unlike the detective story
schema assert a more powerful regulative force. These
142
schemata might be better described as "rule" schemata
wherein the variables and variable constraints are more
precisely specified and wherein the threshold for goodness
of fit is consequently higher. Examples of these schemata
might include a syntax schema that specifies that a noun
follows a determiner or an orthograhic schema that
specifies the letters that form a particular word (Adams
and Collins 20). These sorts of schemata are harder for
readers to adjust because they function at relatatively
local levels, as compared to the more global levels where
some content and genre schemata function. Nevertheless,
readers may still overcome their regulative force and
recognize, for example, that a deviant spelling of a
particular word still may serve to instantiate the word and
the concept that goes with it; that readers will
undoubtably feel more disquiet over such cases than they
might, for example, by reading Pale Fire as a parody of a
detective novel is evidence for the different quality of
regulative force behind the two schemata — the former has
the powerful force of rule and the latter the less
compelling regulative force of convention.
Schemata, then, act as a priori reading constraints by
lending readers criteria for evaluation that they may use
to organize documents into interpretable and interpreted
texts. They therefore fill out Stanley Fish’s notion of
143
interpretive community because they function to describe
the inter-subjective structures shared by members of those
communities; moreover, they fill out Jonathan Culler’s
concept of literary competence which can now be seen to be
a set of schemata with varying degrees of force shared by
readers of literature. However, if schemata describe the
regularities that readers use to understand literature,
they also describe the regularities that guide authors in
the design of documents which they intend to be read as
literary works, and therefore they theoretically describe a
means by which readers may recognize authorial intention.
In this sense, schemata function in much the same way
as E. D. Hirsch’s concept of ’’type," which he defines as
the willed meaning of an author, conveyed through a system
of conventions shared by readers and writers (Validity 92).
But whereas Hirsch’s concept is rigid — because he insists
that meaning is objective and stable — schemata are not,
so that they not only function to link readers and writers
by means of a priori regulative constraints that each may
reasonably expect the other to share (guaranteeing that
authors are afforded membership in interpretive communities
and thereby giving readers another criteria by which to
judge interpretations), but by virtue of their flexibility
they also accommodate differences among readers — and
writers — since individuals will construct texts according
144
to idiosyncratic goals that define evaluative thresholds
within different environments. However, to extend a point
made by Fish about interpretations of texts, an analysis of
authorial intent based upon such a model enters the realm
of argument rather than demonstration, since the degree to
which readers and writers share regulative constraints is
always necessarily an open question. Nevertheless, as
Husserl observed, the recognition of any sort of physical
phenomenon as significant — such as marks on a page —
intimates something about the thoughts of whomever is
responsible for those marks (LI^ 277); consequently the
nature of the motivations that compel authors to design
documents is a powerful and seductive issue that readers
must recognize at some level, even if to dismiss as
tangential to their preferred concerns.
For example, readers of Raymond Chandler’s novel The
High Window are likely to decide that part of Chandler’s
intention was to write a novel that closely conformed to
the schema for detective stories; this conclusion is
largely the rational result of the expectation of readers
that Chandler would share, more or less, the same criteria
for detective stories as do they, and that since they
construct a detective story from Chandler’s work, then that
must have been Chandler's intention. The train of mutual
expectations is endless, but here it suffices to say that
145
Chandler and his readers all have good reason — by virtue
of a common linguistic and literary competence and similar
(but by no means identical) social and cultural backgrounds
— to use the detective schemata. Thus schemata may serve
the same function as Hirsch’s concept of type while
avoiding the rigidity that limits that concept; moreover,
it provides a critical theory within which critics can
argue about intention and about reader-response while
maintaining a reader-centered focus. Perhaps most
importantly, however, such a theory comprehensively
describes the relative roles of readers, writers, and
social constraints.
To realize how a schema-theoretic version of
reader-centered criticism achieves such a description,
however, requires the recognition that any critical theory
implies a model of communication within which it
functions. Often, such models have been described in terms
of the communication triangle of Charles Morris, and
perhaps one of the most influentual of these is James
Kinneavy's model (see Fig. 1) which specifies an encoder, a
decoder, and reality (the world) at the three points of the
triangle while the signal (the document) is contained
within it (96). Kinneavy asserts that whenever one of these
elements dominates a discourse, then the discourse
accomplishes a function consistent with that domination;
146
for example, "when the language product is dominated by the
clear design of the writer or speaker to discharge his
emotions or achieve his own individuality or embody his
personal or group aspirations in a discourse, then the
discourse tends to be expressive" (95). Kinneavy's model,
however, is limited by the spatial scheme he adapts to his
purpose, which seems to isolate each of the discourse
elements from one another and to imply that the three
elements at the points of the triangle are connected only
through the medium of the signal.
In contrast, a communication model grounded in schema
theory posits two triangles (see Fig. 2)j the first one
would define relationships among the historical author, the
psychological text that that author intends, and the
document, or form, which she physically creates. A social
environment penetrates and surrounds this triangle within
which is specified those schemata common to the author and
her interpretive community of probable or imagined
readers. Additionally, a second triangle, distanced from
the first in space and time, would define relationships
among the document created by the historical author
(usually if not always represented in a new form by way of
editors, reviewers, printers, etc.), the actual reader, and
the psychological text which he constitutes by means of the
reading process. Another social environment, not
147
Expressive — Encoder - - - - - - - - Decoder — Persuasive
\ /
\ /
\ Signal ---/------- Literary
\ /
\ /
\ /
\ /
Reality -------- Referential
Figure 1. Kinneavy's model of discourse; terms at the
points of the triangle and within it designate different
aspects of the discourse situation; terms connected to
these designate the functions of discourse when it is
focused on one aspect of the triangle.
Historical Author _ ___ _ _ Intended Text
\ /
Author’s Environment
\ /
\ /
\ /
Document
1
1
Document
/ \
/ \ Reader’s Environment
/ \
/ \
Actual Reader - - - — - - Psychological Text
Figure 2. Representation of a temporal model of dis
course illustrating the distance between writers and read
ers and the probable differences in the form of the docu
ment, the environmental constraints, and the intended and
psychological texts.
148
isomorphic with the first, penetrates and surrounds this
triangle; because of the distance in space and time that
separates the two triangles, some, but probably not most,
of the schemata shared by the historical author and her
community will have changed with the appearance of an
actual readership.
Such a model easily fulfills the two central criteria
for a model of reading literature set up by Fish in
"Interpreting the Variorum"s the ability to explain how a
reader reads differently when reading different "texts" and
how different readers perform similarly when reading the
same "text" (Text 167) since it makes clear the nature of
the environments and the set of schemata — the
interpretive constraints — known to writers and their
individual readers. That these usually overlap explains
how communication is possible, because authors and readers
have in common very similar documents and, more
importantly, a great deal of socially derived, a priori
schemata for the constitution of texts from those documents
and their subsequent interpretation. That the two
environments, and the goals of writers and readers who
inhabit them, vary to greater and lesser degrees explains
why author's intentions will never be entirely known to
readers. Thus authors, and their intentions, are not so
much extensions of texts, as neo-Aristotelean critics
149
claim, but are more like temporal, psychological constructs
of readers who depend upon a priori constraints as
evaluative guides as they fill out the role of writer left
vacant by the absence of the historical author.
This model illustrates a more ecological — and
arguably a more comprehensive — perspective upon literary
criticism than that yielded by considering reader-centered
literary criticism or schema-theoretic descriptions of
reading by themselves. On one hand, it improves upon the
models of reader-centered critics like Fish and Rosenblatt
by specifying the nature of interpretive constraints that
lend readers criteria by which to constitute and understand
literary texts, and it improves upon models of
reader-centered critics like Barthes and Hirsch by
accomodating, respectively, their concerns about the
absence of historical authors and the means by which
readers assign intentions to them despite their absence,
this without specifying either ajd hoc constraints or rigid
structures that are unable to conform to the idiosyncracies
of readers, the environments which they inhabit, and their
particular goals. On the other hand, this model improves
upon schema theory by rejecting its founding metaphor —
that of the machine — for the metaphor of human design,
and by so doing it underscores the activity of readers
motivated by idiosyncratic goals within particular
150
environments; moreover, consistent with the collapse of
Cartesian dualism and its subject / object dichotomy, the
substitution of the new, organic metaphor lends this model
its ability to describe texts as psychological, temporal
structures dependent upon the activities of readers who use
and emend the a priori constraints described by theoretical
schemata.
Finally, and most importantly, the model presented
here easily accomodates the agreements and disagreements
among readers of literary works. Agreement is explained by
the hypothesis that readers of literary works will share a
good deal of similar schemata that they will use to specify
similar criteria by which to constitute texts and their
meanings. Disagreement is explained by the hypothesis that
not all readers will share isomorphic sets of schemata, and
that even if they did, readers are nevertheless bound to
disagree — sometimes in trivial, sometimes in substantial,
ways — since their own motivations will affect the way in
which they use schemata within particular circumstances.
Such an explanation of the variety of interpretive
responses also has the advantage of ameliorating at least
one of the theoretical problems of Fish’s notion of
interpretive community, that such a notion implies that all
members of an interpretive community will agree in all ways
about the interpretations of texts, or else they will, in
151
effect, be members of different communities. If, however,
interpretive communities are defined in terms of shared, a
priori schemata for constituting texts, the members are
still free, since they individually will evaluate the
thresholds of "goodness of fit" between schemata and
different arrays of discourse, to disagree (within some
variable tolerance) with other members within the same
community. Consequently, the memberships of interpretive
communities, while always in flux, will be somewhat more
stable in a schema-theoretic model of literary criticism,
making Fish's concept theoretically more compelling.
However, if Fish’s explanation of how critics proceed
is, as Mailloux asserts, inconsequential because it changes
nothing in critical practice, how then can a
schema-theoretic theory of literature — in some ways
equally esoteric — offer means by which to evaluate the
critical problems of reading literature? The answer lies
in the recognition that any critical theory is an argument
for the use of evidence; if nothing else, schema theory,
thanks to its genesis in the world of cognitive science and
reading research — is so far the most comprehensive theory
developed by these disciplines to explain the way we read
extra—literary texts, and consequently it provides a good
reason for us to assume it is the way we read literary
texts as well. Therefore, if schemata represent the
152
structures by which critics themselves evaluate literary
works, then schemata should provide a useful heuristic by
which to evaluate the standards by which critics respond to
works and judge their literary qualities.
Of course in recognizing the heuristic potential of
schema theory, we must also recognize its potential for
specifying an extremely top-heavy model of critical
response; that is, it is impossible for readers to focus
upon and evaluate their use of every possible schema that
aids them in their response to any particular document
since schemata are operative at all different levels of
discourse processing, some of which are probably beyond
access to even the most careful and discriminating of
phenomenological investigators. Fortunately, there is no
need to be overly specific about the totality of response,
since the theory specifies that for most discourses there
will be large areas of agreement that pose little or no
critical problems and as a result hold little or no
critical interest. Instead, schemata offer a powerful
heuristic by which to investigate the questions of
disagreement among critical interpretations, which by their
very nature do hold critical interest. And, fortunately
again, since schemata do, theoretically, operate at all
levels of response, they can be used as heuristics at
153
whatever level seems constitutive of important interpretive
disagreements.
Fish holds that the decisions about the relative
validity among competing interpretations are decisions that
depend upon persuasion rather than demonstration. The
advantage of interpretive schemata is that they provide a
common framework — based upon a comprehensive theory of
reading — within which arguments can take place, so that
important critical issues can be traced to differences
among critics in the schemata they have used to constitute
the text and its meaning, in the thresholds they are
motivated to set for the use of particular schemata, in the
environment within which they read and the schemata that
are valued by the community of reader/critics within that
environment, and in the assumptions that they make about
the schemata shared by readers and writers. Even these
issues can be broken down — for example, to the variables
that constitute any individual schemata — as necessary to
develop an understanding of the disagreements among
critics, an understanding which itself may serve as a
heuristic for adjudicating among responses.
Nevertheless, the promise of such a heuristic remains
to be tested, and the adequacy of my theory of interpretive
schemata partly depends upon its practical value. I
propose, then, to evaluate the theory I have detailed by
154
using it as a heuristic by which to evaluate the critical
disagreements that have surrounded the poetry of Robinson
Jeffers, and specifically, his narrative poem "Thurso's
Landing,11 the occasion of some of the most wide-ranging
disagreements among respected critics of American
literature.
155
Chapter V
CRITICISM AND THE POETRY OF ROBINSON JEFFERS
The poetry of Robinson Jeffers has drawn from critics
some of the most vicious — and arguably some of the most
entertaining — condemnations accorded to any modern body
of literature. At mid-century R.P, Blackmur attempted to
anticipate future critics by deciding which poets of the
twentieth century would enjoy lasting reputations; he finds
fault with a wide variety of writers, among them Auden,
Empson, and Houseman, only to conclude that "all of them
are better than the flannel-mouthed inflation in the metric
of Robinson Jeffers with his rugged rock-garden violence"
(11). Not many years later, Kenneth Rexroth added his voice
to the vociferous anti-Jeffers crowd in reviewing Radcliffe
Squires’ source study of Jeffers, and again the
condemnation is nothing if not extreme:
In my opinion Jeffers’s verse is shoddy and pretentious
and the philosophizing is nothing but posturing.
His reworking of Greek tragic plots make me shudder
at their vulgarity, the coarsening of sensibility,
the cheapening of the language, and the tawdriness
of the paltry insight into the great ancient
meanings, (30)
156
But as all Jeffers scholars would agree, these are
mild expressons of dislike, comparatively weak heirs to a
tradition established two decades earlier by Yvor Winters,
whose dislike for Jeffers' verse was as constant and
abiding as the coastal headlands in which Jeffers set his
narrative poetry. As Rexroth himself has noticed, it was
one of the most devastating attacks in modern criticism,
helping to push Jeffers' reputation — then at its height
— into a steady decline (30).
Winters first took aim at Jeffers' work in an essay he
reprinted in In Defense of Reason:
The Women at Point Sur [1927] is a perfect
laboratory of Mr. Jeffers' philosophy and a perfect
example of his narrative method. Barclay, an insane
divine, preaches Mr. Jeffers' religion, and his
disciples, acting upon it, become emotional mech
anisms, lewd and twitching conglomerations of
plexuses, their humanity annulled. (33)
As Squires wryly points out, it is impossible to improve
upon Winters' diction when he is inspired by indignation
(159). And Winters himself didn't try; instead he dismissed
a later Jeffers volume, Thurso's Landing and Other Poems
(1932), claiming that "for clumsiness and emptiness"
Jeffers' writing here can hardly be equalled, and
concluding, "The book is composed almost wholly of trash"
("Review" 684),
157
Still, the force of these condemnations is in part a
response to the high praise that was accorded to Jeffers’
poetry by other critics. Jeffers’ first two volumes of
poems — Flagons and Apples (1912) and Californians (1916)
— attracted little notice; but, with the publication of
his first mature narrative, ’’Tamar” (1924), Jeffers was
praised by some of the most influential reviewers of the
time: James Rorty, Babette Deutsch, and Mark Van Doren, who
said of the work, ’’few [volumes] are as rich with the
beauty and strength which belongs to genius alone” (’’Tamar”
268) .
What seems to have moved many of Jeffers’ early
critics was his compelling imagery and the depth of his
philosophic conviction, intangible qualities that gave
Jeffers his ’’power.” Harriet Monroe, though by no means an
enthusiastic admirer of Jeffers’ work, claimed that he was
nevertheless a "poet of extraordinary power" (164),
Granville Hicks, reviewing Thurso’s Landing, named the same
quality: "There is, one cannot deny, a kind of validity in
this and all of Jeffers’s poems; such power was not born of
self-deceit" (433). Deutsch found it in Solstice and Other
Poems (1935): "The poem which opens Jeffers' latest volume
restates his familiar themes, with no loss of power, and
with the additional interest of a greater technical
variety" ("Love" 8). And Van Doren, while offering only
qualified praise of Point Sur, is careful to report: "I
have read it with thrills of pleasure at its power and
beauty” (Point Sur 88).
Whatever the qualities of Jeffers’ verse, one must
agree with Rolfe Humphries' claim that Jeffers "either
knocks you or leaves you cold” (154). This gulf in the
evaluative responses of Jeffers’ critics offers a
remarkable opportunity to compare different critical
approaches to literature, for in a debate over whether an
author is to be included in the canon of important literary
writers, the fundamental assumptions of the different
approaches will be highlighted. Although it is possible to
compare the relative usefulness of critical approaches as
they are applied to more limited literary problems — for
instance the efficacy of New Criticism versus archetypal
criticism to explain Ernest Hemingway's intuition that In
Our Time has "a pretty good unity" (qtd. in Burhans 314) —
such comparisons may be deemed complementary rather than
incompatible since the objects of investigation often enjoy
a relatively stable position within the canon. Since
Jeffers holds a much more tenuous place in the canon, the
different critical approaches to understanding Jeffers’
work more completely ultimately become — explicitly or
implicitly — arguments for whether or not Jeffers' work
holds any literary interest at all, arguments that
159
highlight the criteria for literaryness assumed by the
various approaches.
Frederic Carpenter has observed that most critics —
even those sympathetic to Jeffers' work — have simply
emphasized the qualities which his poetry lacks. The long
poems have usually been called tragic, but condemned for
lacking the virtues of classic tragedy; or they have been
called narrative, and have been denounced for lacking the
fully developed characters which a novelist should create
(Jeffers 55). Even though these critics have applied their
respective — and as Carpenter shows, inflexible —
criteria vigorously to this question and correlative
issues, they have generated such strongly divergent
evaluations that the field of Jeffers criticism still
stands to benefit from a critical perspective with the
explanatory power to evaluate these contributions within a
common frame. Consequently, I propose to examine some of
the critical approaches to Jeffers’ work from a
schema-theoretic, reader-centered perspective (as defined
by the critical criteria for which I have previously
argued), using as a focus the Jeffers poem "Thurso’s
Landing." But first, in order to understand the critical
problems particular to "Thurso’s Landing," I wish to
briefly scketch how contemporary reviewers
160
received the work and what qualities they found to be the
most significant.
The reception of Thurso's Landing and Other Poems was
generally favorable. After Jeffers published Tamar and
Other Poems, the response was so positive that his
publisher reissued that narrative with new material in
another generally successful volume, Roan Stallion, Tamar
and Other Poems (1925). The Women at Point Sur followed,
attracting the unfavorable notice of many critics and
leading some early Jeffers admirers to cool a bit; Rorty,
for example, complained that Jeffers "had not yet
demonstrated that he could create people" ("Satirist" 26).
But the volumes that followed, Cawdor and Other Poems
(1928) and Dear Judas and Other Poems (1929) did so well
with reviewers that by 1932 — the year of publication for
Thurso — Jeffers' reputation was at a remarkable critical
and popular high:
...for in that year...his picture appeared on the
cover of Time; [Benjamin] DeCasseres returned
to predict that "in fifty years only two living
Americans will be read, Robinson Jeffers and James
Branch Cabell"; and Mabel Dodge Luhan dedicated to
Jeffers her book on D.H. Lawrence's sometimes
ludicrous adventures with her in New Mexico,
Lorenzo in Taos. (Vardamis 15)
Of course, Jeffers' enviable reputation at this point
was no barrier to Winters' condemnation; recognizing that
Jeffers might justify himself to his publishers by selling
161
("Review11 685), Winters nevertheless implies that such a
criterion should not be the standard for art. Dismissing
Thurso as "trash," he touches upon what will prove to be
one of the most persistent concerns in Jeffers criticism,
the rhythms of Jeffers’ verse:
Mr. Jeffers' verse continues to miss the virtues of
prose and verse alike: it is capable neither of the
fullness and modulation of fine prose nor of the con
centration and modulation of fine verse. There is
an endless, violent monotony of movement, wholly
uninteresting and insensitive. (684)
More favorably, William Rose Benet claims that despite
a good deal of prosaic statement, Jeffers’ technique
"gathers great rhythms into itself," explaining further
that
Frequently it seems to me to be prose writing
shortened into arbitrary lengths. It has, however,
a fine prose quality and he achieves through his
particular method a dramatic condensation not
possible to prose written as such. I do not know
that there are any rules for this sort of writing,
but I feel on finishing a poem of his that the means
is justified by the total effect. (20)
Other critics describe the rhythms variously,
Granville Hicks claims that "Thurso’s Landing" moves
"swiftly, in lines terser and firmer than those the poet
had hitherto composed" (433); Deutsch, in some contrast,
sees the lines as "deep-breathed, billowing" ("Hunger" 7);
and Humphries claims that the lines express "a rhythm that
162
induces an almost hypnotic response" (154). The following
descriptive fragment illustrates the unorthodox lines with
which these reviews contend:
A curl of sea-cloud stood on' the head of
the hill
Like a wave breaking against the wind; but when they
reached it, windows of clearness in it were
passing
From the northwest, through which the mountain sea
wall looked abrupt as dreams, from Lobos like
a hand on the sea
To the offshore giant at Point Sur southward.
(Selected Poems 274-75)
Connected with the question of whether or not Jeffers'
work qualifies as prosaic verse or verse-like prose is the
question of its overall form, the genre and its attendant
means of organization through which the poem is unified.
The question is important, as Hirsch has shown, because
genres provide readers with expectations about the works
they read. Winters adopts the terminology of most critics
and focuses on the narrative qualities (or lack thereof) of
"Thurso's Landing": "There is an attempt at some coherent
narrative, but the result is merely dogged and soggy
melodrama" ("Review" 684). But other qualities noticed by
reviewers implicitly question the adequacy of such
categories. Louis Untermeyer finds a mixture of terms
necessary to describe the poem, calling it a "transplanted
Greek tragedy" wherein the dramatis personae are symbols of
tortured humanity; and he continues to adopt the
163
terminology of Greek drama to describe the denouement of
the poem, "Only the old mother, a grim chorus, remains"
(815). Yet despite his reliance on dramatic terminology,
Untermeyer casually introduces the term "narrative" as
descriptive of the work (815), That both sorts of terms are
necessary to adequately account for Jeffers’ work points to
a critical dilemma, one that Humphries attempts to solve by
claiming that "Thurso's Landing" is a drama with amplified
descriptions of scenery and stage directions on whose
writing the dramatist has lavishly expended his "finest
art" (154). To complicate the matter, Benet sees one
passage in the poem (where Reave Thurso severs a cable that
spans a nearby canyon [SJ? 321]) as "one of the most vivid
and powerful bits of writing that I know of in modern
fiction" (20), thereby questioning, at least implicitly,
whether the poem might better be considered as a versified
novel,
A second crux is equally prevalent in the reception of
"Thurso’s Landing," the problem of Jeffers' philosophy,
"inhumanism." Jeffers was eventually able to clarify the
main tenets of inhumanism into a convenient schema:
First: Man is also a part of nature, not a miraculous
intrusion. And he is a very small part of a very
big universe, that was here before he appeared, and
will be long after he has totally ceased to exist.
Second: Man would be better, more sane and more happy,
if he devoted less attention and less passion (love,
hate, etc.) to his own species and more to non-human
164
nature. Extreme introversion in any single person is
a kind of insanity; so it is in a race; and race has
always and increasingly spent too much thought on
itself and too little on the world outside.
Third: It is easy to see a tree, a rock, a star are
beautiful; it is hard to see that people are beautiful
unless you consider them as part of the universe —
the divine whole. You cannot judge or value any
part except in relation to the whole it is part
of. (Letters 291)
Jeffers added one more component to his philosophy, a
natural deism, or pantheism as Jeffers himself reluctantly
named it (Themes 23-24), giving his philosophy an
essentially four-tenet structure,^
When Milton wrote Paradise Lost and invoked the aid of
his heavenly muse to accomplish his avowed intention to
"justify the ways of God to men" (1.26), he could at least
count upon a readership familiar with the biblical story of
Genesis, and furthermore he could count on his audience to
place his retelling of the story in relationship to
Protestant dogma. The case was different with Jeffers; his
philosophy of inhumanism was not a part of the contemporary
culture within which he wrote, and to the extent that
reviewers understood his intentions were to express that
philosophy, it often became an aesthetic issue. For
example, Randall Jarrell blames much of the faults he sees
in Jeffers’ poetry on his philosophy:
He celebrates the survival of the fittest, the
war of all against all, but his heart goes out
to animals rather than to human beings, to
minerals rather than to animals, since he despises
___________________________ . _ .... 1.6.5
the bonds and qualifications of existence. Because
of this, his poems do not have the exactness and
concision of the best poetry; his style and temper-
ment, his whole world-view, are to a surprising
extent a matter of simple exaggeration. (322-23)
Untermeyer boiled down Jeffers' creed into a formula
— life is horrible, love is inverted and hence incestuous,
death is the savior, and civilization is a transient
sickness — and then comments that Jeffers' philosophy is
an idee fixe which burdens characters with "states of
unremitting tension" (816). In conclusion, Untermeyer
contrasts Jeffers’ achievement as a philosopher with his
achievement as a poet: "The philosophy is negative,
repetitious, dismal; the poetry, even when bitterest, is
positive as any creative expression must be" (817).
Almost all of Jeffers’ critics seem to accept this
dichotomy. Hicks, more sympathetic than Untermeyer, sees
Jeffers' philosophy not so much as an idee fixe but as a
"vision of annihilation, power and beauty." But even he
implies that the philosophy is a burden upon the poetry:
"If that passion had been, at the outset, directed into
other channels, what might it not have accomplished?"
(434). In general, critics are drawn by the "power" of
Jeffers' expression but repelled by his message.
Nevertheless, the consistency of almost all of Jeffers'
mature work with his inhumanist creed makes his philosophy
an important issue for critics, underscoring the question
166
of how much of Jeffers' intention in writing his verses is
rhetorical, communicative, persuasive; that the expression
of this philosophy is also an aesthetic issue is born out
by the demands that inhumanism makes upon the selection of
form, characters, imagery, and especially diction, for
Jeffers — steeped in science and driven by a terrific
2
curiosity — often used words from medical science and
astronomy, among other scientific disciplines, lending his
3
work a peculiar idiom.
"Thurso's Landing," though fully consistent with
Jeffers' inhumanist tenets, is not quite typical of his
work. On the flyleaf of the 1932 volume, Jeffers wrote:
It ... seems to me to be the best thing I have yet
written.... The time is perhaps more distinctly near
the present than is usual in my verses; the persons
seem to me to be a little more conscious of the moral
implications of what they do. (Quoted in Alberts 72)
In their respective reviews, both Hicks and Benet took
notice of and agreed with Jeffers’ evaluation of his work.
At a later date, Carpenter too would remark upon the
difference, but see it in a less favorable light: for him
the poem is a relative failure because Jeffers had
abandoned his more usual mythic mode for a more
contemporary and realistic story (Jeffers 83). And John R.
Alexander rated it not one of the best of Jeffers'
narratives, though it still "exhibits the virtues of human
167|
strengths and endurance against the inevitability of death
(93). In terms of the Jeffers canon, "Thurso’s Landing"
presents most of the problems of philosophy and form
presented by the other long works, with the additional
advantage that Jeffers and some of his contemporary critics
regarded it as one of his best; that other contemporary
critics and many later scholars disagree simply heightens
the interest.
The poem, according to Jeffers, is about "courage."
Prior to the time in which the poem is set, Reave Thurso's
father had committed suicide after the failure of a
lime-kiln venture. Reave still resents his fathers
"cowardly" act, and struggles to make the ranch — a small
patch of land compared to what his father had previously
owned — once again productive, in part because he is
responding to a felt challenge to prove that there are
alternatives to his father's failure and in part to provide
for his family: his brother Mark, lamed in France during
the First World War and mentally unstable; his wife Helen,
dissatisfied with the bleak ranch life and impatient with
her husband's dogged determination; and finally his mother,
a tough old women suspicious that Helen may not be devoted
to Reave. When Helen runs off with a friend of Reave's, he
stubbornly pursues her for a year, bringing her back to
renewed unhappiness.
168
Subsequently, Reave attempts to make a fresh start by
planning to cut down the great steel cable by which Old
Thurso had delivered his lime across a gorge to waiting
ships. By removing the cable Reave would remove the last
physical reminder of his father’s failure and thereby
exorcise the influence that Old Thurso still holds over
him. Reave files through nearly all of the great cable's
wire strands, and as Helen watches, he chops through the
last few strands with an ax, releasing the energy stored in
the cable all at once: with a tremendous force, the cable
coils back at him, breaking his back and rendering him a
permanent and almost helpless cripple. Afterward, Reave
suffers unremitting pain, but even though life has become a
constant agony, he refuses to surrender in the way of his
father. Helen begins to admire his courage, but even as
she grows more tender toward Reave she wishes for an end to
his agony. Mark, slowly driven more insane by portentious
visions of Old Thurso and by a repressed desire for Helen,
hangs himself rather than live with his torment. Finally
Helen cuts Reave's throat, then poisons herself, leaving
the old mother to bury their bodies.
The plot of "Thurso's Landing," then, would seem to be
one of unrelieved tensions, violence, and death —
requiring if nothing else, a strong sensibility to see it
through. However, the questions of language, form, and the
169
aesthetic influence of inhumanism are central for an
examination of how the poem works, for it is through these
elements that Jeffers has sought to affect his readers.
Perhaps the most unusual of these aspects — and probably
the most obvious to new readers of Jeffers’ work — is the
nature of Jeffers’ language, the question of its qualities
as verse and/or prose, and it is to this question that I
turn first.
Considering how unusual Jeffers’ lines are, it is
surprising that critics have for the most part neglected to
analyze how they are structured. As Powell points out, it
was not until 1930 that any critic had noticed and studied
the metrical practices underlying what was ’’apparently free
verse”^ (118). Instead, critics compared Jeffers' lines to
those of Walt Whitman; Everson obliquely makes this
comparison by asserting that only two poets — Jeffers and
D.H, Lawrence — were able to resist the ’’Wave of the
Future” represented by Ezra Pound’s attempts to purify the
language (Fragments 25-26).
Such comparisons tend, however, to conceal the more
idiosyncratic qualities of Jeffers' verse, Powell (drawing
upon the work of Herbert Klein) asserts that the narratives
are not nearly so free-form as comparisons to Whitman might
seem to imply. He finds that Jeffers' shorter poems are
fairly regular"* but that
170
the narratives are more irregular, though an exam
ination will reveal that Jeffers carries his
prosodical principle into their making. The pattern
for the narratives is the ten stress line alternating
with the five, and varied by lines of uneven length.”
(120-21)
Moreover, according to Klein,
Jeffers uses stress to define and limit the line;
quantity to regulate it. That is to say, the
sheer possible syllabic length of the line is set
(within exceptionally broad limits) by the number
of stresses which the beat pattern of the poem
permits. But the tempo of the lines, the contrast
of breathless in one line (or joyful skipping) with
sonorous deliberation (or hard wrenching plodding) in
another is due to quantity, (qtd. in Powell 121)
Everson makes one additional point about Jeffers’
prosody: it has a typographical looseness which lends it a
particular appearance on the printed page, a looseness,
which, had Jeffers wished, could be somewhat overcome by
arranging lines to correspond to shorter phrases (Fragments
42). Everson justifies the decision by speculating that
Jeffers was attempting to render experience as it was,
ineffable and imprecise, rather than as something concrete
capable of being contained and structured. But even so,
the immense length of a typical Jeffers line can make a
narrative like "Thurso’s Landing" hard to read as verse.
Jeffers begins his narrative with a short section
describing the work of a construction crew repairing the
coast road. The section soon focuses on how the dynamite
171
man Rick Armstrong — the man with whom Helen will later
flee — sets an explosive charge
As if he shared a secret joke with the dynamite;
It waited until he had passed back of a boulder,
Then split its rock cage; a yellowish torrent
Of fragments rose up the air and the echoes bumped
From mountain to mountain. The men returned slowly
And took up their dropped tools, while a banner of dust
Waved over the gorge on the northwest wind, very high
Above the heads of the forest.
Some distance west of the road,
On the promontory above the triangle
Of glittering ocean that fills the gorge-mouth,
A woman and a lame man from the farm below
Had been watching, and turned to go down the hill.
The young woman looked back,
Widening her violet eyes under the shade of her
hand. "I think they'll blast again in
a minute."
And the man: "I wish they'd let the poor old road
be. I don't like improvements." "Why not?"
"They bring in the world;
We're well without it." (SJP 266)
Thus are readers introduced to three of the characters of
the work, but more importantly to the linguistic structure
of Jeffers' verse and the problems that it presents to the
reader, the most striking of which is the elongation of the
line that coincides with the dialogue between the
characters of the woman, Helen, and the lame man, Mark.
Prior to this elongation, readers will notice features
of a clearly poetic schema:^ assonance, alliteration,
metaphor, inverted syntax, personification, and the
breaking of the discourse into lines The qualities of these
initial lines are easily organized by the reader by means
172
of a schema defining verse and its constitutive
sub-schemata, and such an organization is supported by the
likelihood that readers will follow an additional
sub-schema for reading poetry by recognizing the line
breaks that define the poetic line and thereby constructing
the verse according to the relatively short five beat lines
noticed by Klein and Powell. Moreover, the typographical
layout of the lines will confirm this reading, since the
poetic line begins and ends with each typographical line, a
schema that the reader may readily interpret as a rule or
convention of Jeffers* verse. At this point, Jeffers'
prosody may be interesting, but well within the bounds of
the a priori schemata that many readers will use to
organize their perceptions of poetic form. To the extent
that most readers are likely to agree on such a reading the
poetic schema and constitutive sub-schemata are
conventional.
However, such schemata are at first glance inadequate
for understanding the lines that follow, for with the
introduction of the characters of Mark and Helen, the line
lengths seem much more indeterminate. Part of this is no
doubt due to the typography, as Everson has noticed, for as
the lines lengthen, the width of the page can no longer
accomodate them without turning, sometimes more than once.
Moreover, the problem is intensified by the fact that
173
Jeffers indents some of his lines, so that indentations
cannot serve as easy and conventional clues for a continued
poetic line. As a result, reading these lines as verse
becomes that much more difficult, and readers are forced to
change their expectations of how the lines work,
consequently giving up the schema that defines the poetic
line in terms of typographical units. Readers are faced
with a choice every time they come to the typographical end
of a line: either read that as the end of Jeffers' long
poetic line and look for a confirming upper-case letter at
the beginning of the next, or read it as a typographical
break dictated by the page size and read the next line down
as a continuation of the poetic line, confirming this by
looking for an indentation and lower-case letter.
Furthermore, the irregularity of these longer lines
may tempt readers to read them as ordinary prose, or at
least as something closely akin to writing with irregular
units of rhythm. Thus readers may be likely to shift from
a relatively conventional poetic schema to one that
foregrounds itself by means of less regular and thereby
more prosaic rhythms.
However, despite the likelihood that readers will read
the lengthened lines differently, the longer lines are not
totally without rhythm; recall that Powell and Klein found
within their boundaries approximate rhythms of ten beats to
174
a line. Scanning a line from the dialogue confirms a more
or less rising meter throughout that, to my ear at least,
sets up a loose, poetic cadence:
/ / /
And the man: I "I wish | they’d let | the poor
II / /
old road | be. |[ I don’t like | improvements." |
"Why not?" | "They bring in | the world,"
The shift in line length is not so much indicative of
readers changing from a poetry schema to a prose schema,
but of a shift in emphasis, a relaxation of the conventions
of the poetic schema in order to let other qualities of the
language emerge. In a sense, readers construct a
super-ordinate schema as they read Jeffers' verse, relaxing
and emphasizing the relative force of organizational
schemata, expecting to shift their attention to different
elements of the verse at different times.
Jeffers’ long lines are not limited to sections of
dialogue. For example, at the beginning of Section X when
Reave returns Helen to the ranch, the lines lengthen as
Hester Clark, Reave's mistress, is introduced:
They drove through the two
deserts and arrived home. Helen went in
With whetted nerves for the war with Reave's
mother, resolving
Not to be humble at least; but instead of the
sharp old woman a little creature
With yellow hair and plated excess of clothing
stood up in the room; and blushed and
175
whitened anxiously
Gazing, clasping thin hands together. (SJ? 298)
And shortly thereafter the lines lengthen once again for
one of the infrequent commentaries by the poem’s narrator:
No life
Ought to be thought important in the weave of the
world, whatever it may show of courage or
endured pain;
It owns no other manner of shining, in the broad gray
eye of the ocean, at the foot of the beauty
of the mountains
And skies, but to bear pain; for pleasure is too
little, our inhuman God is too great,
thought is too lost. (SJP 325)
In short, the lines lengthen in "Thurso’s Landing"
simultaneous with some specifically narrative function,
either the introduction of dialogue, the shifting of
setting and the introduction of character, and the
intrusion of the narrative voice. Thus readers may give up
— within some tolerance — paying attention to more poetic
elements and focus instead on the conversational qualities
of the dialogue, the exposition within the narrative,
and/or the narrative commentary, and to the extent that
these schemata become the focus of attention, the meter and
other poetic elements are less foregrounded in the
attention of readers.
However, in those sections where the line length
returns to the tighter, five beat pattern, readers may once
again shift their attention to more conventional, poetic
176
features of the discourse. For example, consider that
section where the servant Olvidia discovers Mark Thurso’s
lifeless body hanging from a sycamore tree:
A broad brown face peered from a window, a woman whose
blood had known this coast for ten thousand
years
Perceived the strangeness of shape in that moony
sycamore. She had work in hand, but an hour
later she poured
Coffee for Mark Thurso, who1d not come down yet, and
set it with his egg back of the stove and went
To empty the grounds into the willows. She hooded her
head against the rain with an apron, but
returning
Saw the sycamore framed in the apron fold
And it looked dreadful. She approached and found Mark
hanging
Long-necked, very wet with rain. (SP 338)
Readers who by now are used to shifting the schemata they
use to understand the poem and its prosody will expect that
the long lines, here functioning as exposition for the
narrative aspects of the poem, are giving way to more
compact rhythms to allow a different kind of reading
wherein readers slow down, take greater notice of the more
frequent line breaks, and appreciate the poetic qualities
— the alliteration, the imagery, and the caesura — that
thereby emerge. This shift focuses attention on the
rain-soaked body and the subject of suicide. Thus the
shifting line lengths seem motivated to the extent that
they help the reader to focus attention on aspects either
poetic and evocative of some emotional response — as in
177
the above section wherein the reader constructs the image
of Mark's suicide — or less poetic and evocative of a
different kind of experience — as in a previous section
where the cadences seem suited to the imitation of natural
dialogue. Readers learn to effect this shift by creating
schemata in the course of their reading by which they can
productively account for the apparent differences among the
lines and consequently overcome the disjunction they feel
when they are forced to read different sections in
different ways.
This schema-theoretic account of Jeffers' prosody also
explains the critical dilemma of Jeffers' reviewers and
critics. That readers may productively read "Thurso's
Landing" by means of organizational schemata by which
readers shift attention to manifest different qualities of
language explains contradictory comments like those, for
example, of Benet, who saw in the lines a "fine prose
quality," and those of Hicks, who saw the lines as "terser
and firmer" than those of previous Jeffers works.
Presumably Hicks paid greater attention to those aspects of
the verse that corresponded to schemata for poetic rhythms,
and that consequently those schemata provided him with
expectations about continuing rhythms that were
substantially born out by his reading. Benet, in contrast,
noticed more those schemata which tended to divert
178
attention from the rhythms of the language and thereby to
emphasize a certain prose-like quality of sections of the
poem.
In both cases, the relatively positive responses
result from a willingness to entertain new forms, a
willingness not shared by a critic like Winters who
condemned Jeffers partly because he could find in "Thurso's
Landing" neither the "fullness and modulation" of fine
prose nor the "concentration and modulation" of fine
verse. To me, such a comment betrays the strength of
Winters' adherence to his own conceptions of the proper
forms of prose and verse. As a result, verse such as
Jeffers wrote could never be successful according to
Winters' criteria because it would seem to question the
boundaries by which literary form is defined.
But it is important to realize that what Jeffers'
critics have usually done is to argue for a single schema
by which the text is or should be organized. Winters
argues from the perspective of traditional and formal
literary categories, almost infusing these with the force
of rule; to the extent that a work does not "fit" these
pre-existing categories it is, for Winters, a failure. In
contrast, Robert Brophy argues for the adoption of
archetypal categories on the basis of authorial intention:
"...unless critics have fully absorbed Jeffers' mythic
179
intent, ritual structures, and allied imagery patterns,
they have a very uncertain basis for many of their
judgments" (286). In effect, these critics share an impulse
to emphasize in some way the force of a single conventional
schema of some kind, and thereby they betray a predilection
to remake Jeffers’ individual texts into a pre-existing
pattern, to make them rejoin some original ideal of which
these are then copies. Such attempts are successful in
only limited ways. To the extent that Brophy, for
instance, is successful in his reading of Jeffers' works in
terms of archetypal schemata, he can support his assertion
that Jeffers' intent in writing his long works was
"mythic"; but such a position does little to explain a work
like "Thurso's Landing" that most critics — and Jeffers
himself — agree is not mythic. Robert Zaller's particular
frame of reference, the Oedipal conflict, is more helpful
in understanding "Thurso" since there is general critical
agreement that the psychological conflict between Reave and
his dead father is central to the poem. But insofar as
Zaller projects this conflict as an aspect of the text and
extrapolates beyond to speculate about the relationship of
Jeffers to his own father, the critical schema is weakened
in explanatory force because it tends to pay little
attention to the roles of characters outside the immediate
Oedipal drama and more importantly to the transcendent,
180
positive assertions of religious awe that critics find in
the longer works. My point is not that Jeffers did not —
in some of his longer works — intend to unify them by
means of archetypal or Oedipal schemata, but that these
schemata are of limited use to explain the organization of
all of his longer works that whatever their differences
seem of a peculiar type.
But in addition to the fact that these critics have
adopted some more or less psychological criterion by which
to explain Jeffers' mode of organization, they also share
another critical strategy that limits their
interpretations: they all implicitly assume that the text
that they recreate in reading is the only possible text.
For example, Everson writes of Point Sur:
...when Barclay enjoys the gross Indian woman Maruca
in Chapter VI this act evokes his daughter, who
symbolizes virginity and purity, in the next chapter.
They are not unrelated anecdotes, therefore; they are
related by archetypal accomodation, each evoking, and
subsequently yielding to, another. (120)
Everson assumes that the relationship between episodes
he sees is a property of the text rather than the product
of his application of an archetypal schemata in reading.
The result in this case is that Everson assumes simply that
Winters fails to see in the text a property Everson finds
obvious. But Everson and Winters are not using different
181
approaches on the same text; they are creating different
texts by reading Jeffers1 works using different schemata.
These critics have done Jeffers criticism a valuable
service by offering complex, sensitive readings of Jeffers’
works and by revealing their potential as evocative,
aesthetic documents. In particular, they have been
collectively valuable for restoring some measure of
scholarly respectablity to a body of poetry that, as
Everson most particularly shows, has been treated unfairly
by some of the New Critics. However, I do question the
universal utility of the schemata they adopt in terms of
their ability to explain the mode of organization for works
like ’’Thurso’s Landing” that rely less explicitly upon
mythic or Oedipal modes for their organization, and
certainly I question the assumption implicit in their
criticism that each schema that they individually prefer as
an interpretive heuristic for understanding Jeffers’ texts
is actually a quality of the text itself. In contrast, a
schema-theroetic criticism acknowledges a role present in
all critical enterprises: the role of the reader. By
showing that critics, like all readers, constitute the
texts which they seek to interpret, a schema-theoretic
criticism underscores and explains the basis of the
differences of critical assumptions among these critics.
182
The sheer variety of schemata by which critics have
organized Jeffers’ works suggests that no one schema is
essential; instead it suggests that these schemata can be
used by readers in a variety of different ways, so that
readers have the task of creating some super-ordinate
schema wherein schemata interpenetrate one another. In
this way, Winters’ claim that Jeffers’ longer poems are
organized by repetition — the restatement in successive
stanzas of a single theme, the terms or images being
altered with each restatement (Defense 30) — can help
readers appreciate some aspects of "Thurso’s Landing." For
example, the second section of the poem ends:
The group dissolved apart, having made for
a moment its unconscious beauty
In the vast landscape above the ocean in the colored
evening; the naked bodies of the young bathers
Polished with light, against the brown and blue denim
core of the rest; and the ponies, one brown,
one piebald,
Compacted into the group, the Spanish-Indian horseman
dark bronze above them, under broad red
Heavens leaning to the lonely mountain. (SP 268)
Aside from the remarkable color imagery, readers aware of
Jeffers’ inhumanist tenets are likely to notice the
contrast between the human and natural worlds, the way in
which the natural world towers over, surrounds, and dwarfs
the human events just concluded. Such elements easily
become variables within a developing "inhumanist" schema
since they suggest the place of the human within — as
183
opposed to beside — the natural world. Compare this
tableau with Helen’s characterization of her husband as
from a distance she and her lover watch Reave carry away a
deer he has just killed during a hunt:
The two men were seen, one burdened, like mites in a
bowl; and Helen with a kind of triumph: ’’Look
down there:
What size Reave Thurso is really: one of those dirty
black ants that come to dead things could
carry him
With the deer added.” (SP 271)
This excerpt falls within the first third of the third
section, and like the tableau above calls into question by
means of comparison of the human and natural worlds the
importance of human events within a greater whole. To the
extent that both of these scenes are repetitions in
different terms of tenets of Jeffers' inhumanism, Winters'
schema of repetition serves as a neat means by which to
characterize the narrative mode of "Thurso's Landing," But
it is note the only schema that readers may use to tie
together the narrative; another important schema is the
symbolic, for example the great cable that Old Thurso slung
across the gorge.
Readers first notice this cable at the beginning of
the third section when Rick Armstrong passes under it on
his way to the Thurso farm:
184
He passed under a lonely noise in the sky
And wondered at it, and remembered the great cable
That spanned the gorge from the hill, with a
rusted iron skip
Hanging from it like a stuck black moon. (SJP 268)
The lines in this section are tight, recalling the poetic
schemata that readers will have developed and requiring,
therefore, special attention. The simile comparing the
skip to the "stuck black moon" focuses that attention on
the skip and the cable that supports it, so that readers —
struck by the lifeless blackness of the comparison — are
likely to read the skip and the cable as symbolic, perhaps
as an evil portent; readers may eventually see the
lifeless, hanging "black moon" as an omen that forshadows
the lifeless hanging body of Mark Thurso in "that moony
sycamore" (SP 338). More functionally, however, the cable
itself serves to augment the structure of the poem by
psychological means which, according to Kenneth Burke,
create an "appetite in the mind of the auditor, and
[demand] the adequate satisfying of that appetite" (31).
Thus the cable is recalled by readers after Reave returns
to the farm with the runaway Helen, but here with a new
association: according to Mark, Old Thurso returns every
night to stare at the cable he once strung (SJP 303). Thus
when Reave proclaims his intention to cut down the "old
advertisement of failure" to start a new life (SP 303),
185
readers understand that Reave’s motivation is once and for
all to remove the last physical reminder of his father's
failure. When Reave does the deed and cripples himself
beside — "He was bent at the loins backward / And flung on
the face of the hill" (SP 321) — readers realize a more or
less complete symbolic schema that is both semantic and
structural: the cable serves not only as the symbol of Old
Thurso's continuing influence on Reave, but its
introduction early in the poem forshadows subsequent
appearances later. The point is that in narrative terms,
readers need to use not only a schema that recognizes
repetition of an idea in new forms, but also the recurrence
of a unique symbol that serves to bind the work.
Readers are likely to lend Jeffers' work form in at
least these two ways. That these schemata may also
function productively to augment, dominate, or fill out a
schema defining, for example, Zaller's preferred category
of Oedipal conflict, points not so much to a flaw in the
work, but to the limits of any theory of criticism that
seeks to force the individual work to rejoin its "ideal"
form. In fact, like Jeffers' prosody, the mode of
organization of his longer works serves to question the
boundaries of conventional schemata as they are habitually
used.
186
I have already shown how readers may use a developing
schema of inhumanist tenets to organize the poem by means
of repetition. I must also add, however, that the success
of such a process depends a good deal upon the knowledge
that readers of Jeffers’ work bring to a poem in their
first or subsequent readings. In other words, the
experience of reading Jeffers' other works provides a
background of schemata — including a set that defines
Jeffers’ philosophy — by which readers evaluate the
elements that they recognize in later readings. That
critics do this is evident in the comparisons that they
inevitably make among different works from within and
without the Jeffers’ canon. Nevertheless, readers lacking
a good deal of background in Jeffers' work will still
recognize in different ways that an individual work like
"Thurso's Landing" resists easy organization by means of
conventional philosophical schemata and that some less
conventional schema needs to be formed and used. Thus
critics who recognize, for example, the trans-human point
of view adopted by the persona of "Roan Stallion" —
"Humanity is / the start of the race; I say / Humanity is
the mould to break away from, the crust to break through,
the coal to break into fire, / the atom to be split" (SP
149) — will be at some advantage to understanding later
works like "Thurso’s Landing." In short, they will have an
187
a priori content schemata that takes on an almost
conventional force within the Jeffers canon; readers
without such a schema will have a more difficult time
insofar as they perceive Jeffers1 philosophic point of view
as unconventional within their experience, even though such
elements as characterization, imagery, and diction may
serve these readers as variables as they fill out and form
such a schema.
This philosophic point of view gives Jeffers room to
speculate upon the nature of human consciousness. Powell
believes that Jeffers is drawn to states of the body as
well as the mind, claiming that he is perhaps unique among
poets in the extent to which has distilled poetry from
physiology, biology, and chemistry (72). This excerpt from
"Thurso's Landing" represents the moment of Rick and
Helen's love-making, a moment that at once is within and
beyond the level of Helen's consciousness;
...under the thick brown hair and under the
cunning sutures of the hollow bone the nerve-
cells
With locking fibrils made their own world and light,
the multitude of small rayed animals of one
descent.
(SP 276)
To the extent that readers notice that the diction is
scientific, they may also associate it with a scientific,
theoretically disinterested description of the workings of
188
human consciousness. Rather than triggering romantic
schema of humanity, the diction tends to build up a schema
in which humanity — in the form of its constitutive form
of consciousness — becomes an "automatism of life" (SP 77)
and thereby to question its priority within the natural
order. Critics familiar with Jeffers' work will likely see
the unusual diction as a variable within the inhumanist
schema because it emphasizes the place of humanity within
the whole, as an aspect of the influence of science which,
according to Everson, provided Jeffers with the objectivity
and hence the authority to effect his "religious mission"
to destroy human complacency by revealing humanity's
insignificance ("Forward" viii). Readers who do not bring
to the work a more or less complete inhumanist schema may
relate the diction, which fits uneasily into conventional
poetic schemata, to the inhumanist schema they have been
developing from their observations of the poem's
philosophic unconventionality.
Furthermore, readers may also use an inhumanist schema
to understand elements of Jeffers' characters and their
allied imagery patterns. As Carpenter has noted, the
characters of Jeffers' works are often redolent with mythic
overtones. However, they are also associated through
images and metaphors with symbols by which readers learn to
recognize inhumanist qualities. Jeffers' lyric poem "Rock
189
and Hawk" makes a useful gloss for two of the most
prevalent symbols in Jeffers' work:
Here is a symbol in which
Many high tragic thoughts
Watch their own eyes.
This gray rock, standing tall
On the headland, where the seawind
Lets no tree grow,
Earthquake proved, and signatured
By ages of storms: on its peak
A falcon has perched.
I think here is your emblem
To hang in the future sky;
Not the cross, not the hive,
But this; bright power, dark peace;
Fierce consciousness joined with final
Disinterestedness;
Life with calm death; the falcon's
Realist eyes and act
Married to the massive
Mysticism of stone,
Which failure cannot cast down
Nor success make proud. (S^P 563)
Although "Rock and Hawk" orginally appeared in a volume
published three years after "Thurso's Landing," Solstice
and Other Poems, it is nevertheless a useful shorthand by
which to understand the place of these variables within an
inhumanist schema.
In this poem, readers will notice that the rock is
steadfast against the seawinds that scrape across it and
the earthquakes that occasionally cause the stress
190
fractures that signature its mass. The hawk perches
indifferently upon it, an arrogant and savage creature that
with the rock forms a single image: a natural scene in
which natural elements are allied and balanced through
strength and fierce consciousness. Contrasted to this is
an element of a more conventional symbolic schema: the
cross, an aspect of traditional Christian orthodoxy and its
concern with human morality. Readers may thus see in this
poem a contrast of philosophic/religious schemata:
traditional Christianity evoked by the symbol of the cross
replaced by the natural, pagan symbolism of the rock and
hawk, which together function to turn the reader’s
attention away from human values toward the sometimes
savage but indomitable and transcendent natural order. The
poem is thus easily organized and explained within an
inhumanist schema.
In "Thurso's Landing," the hawk and the rock are
images that lend characters in the poem their attributes.
The indomitable arrogance of the hawk is given to Reave
Thurso’s mother first by the third-person narrator of the
poem and then by her daughter-in-law Helen in a scene that
follows immediately upon Helen's illicit lovemaking:
Helen said, "Reave went after a wounded deer
And sent me home. He hasn't come home yet?"
Reave's mother said "We've not seen him," steadily
watching her
Across the lamplight with eyes like an old hawk's,
191
Red-brown and indomitable, and tired. But if she
was hawk-like
As Helen fancied, it was not in the snatching look
But the alienation and tamelessness and sullied
splendor
Of a crippled hawk in a cage. (_SP 277)
Readers familiar with Jeffers’ work will notice the
hawk image and the characteristics that are associated with
it, all as part of Jeffers' inhumanist design. Moreover,
these readers may associate the ’’crippled hawk in a cage"
with the Jeffers' lyric "Hurt Hawks," originally published
previous to "Thurso's Landing" in Cawdor and Other Poems,
wherein the natural and inhumanist spirit of a crippled
hawk serves as the central theme. However, even readers
unfamiliar with the frequency of this image in Jeffers'
work and the inhumanist message that it may evoke, may use
it as a variable as they construct a high-level thematic
schema and look for confirmation in the form of additional
occurances, as happens, for example, later in the poem when
Helen compares Reave's mother with "an old hawk" when she
complains of the lack of humanity in the Thurso family (SP
281), The hawk imagery thus prepares readers for the close
of the poem, when the old woman is left with nothing of her
family save their dead remains, yet bears her loss and her
unhappiness with a kind of indomitable spirit; her words to
her servant Johnny Luna close the poem: "Help me: you'll
have to carry all the weight. I am the last / And worst of
192
four: and at last the unhappiest: but that's nothing” (SP
357) .
Untermeyer had characterized the role of the old woman
as that of a ”grim chorus" (815). In one sense, readers may
find that that is an appropriate description of her role,
since through the hawk image she is associated with the
prevailing values of the inhumanist world. But readers
also notice how her son Reave is associated with the image
of the rock, and consequently use that association to
further fill out an inhumanist schema. Helen sees in her
husband these qualities: "How can I live / Where nothing
except poor Mark is even half human, you like a stone, hard
and joyless, dark inside" (SP 281), Such qualities — like
those of the rock in "Rock and Hawk" — help readers to
form a sub-schema, highly conventional within Jeffers'
works, by which to organize Reave's actions througout the
poem, by which to understand Reave's force of will in
hunting down Helen after she flees with Armstrong and his
detestation of his father's lack of fortitude. Moreover,
such a sub-schema fills out the higher level inhumanist
schema by means of contrast: Reave's stone-like
determination is obsessive and unhealthy, but better by
contrast than Helen's lust for human comforts and her
self-love.
193
Some critics — for example Carpenter — have remarked
upon the relative flatness of Jeffers* characters.
Carpenter explains this flatness by asserting that the
characters are actually mythic, representations of forces
of nature. In "Thurso’s Landing" this explanation is
inadequate partly because the characters, despite their
association with natural images, are nevertheless clearly
human in their motivations. Still, it is possible —
within a schema-theoretic frame — to see the flatness of
the characters in functional terms. Jane Tompkins, arguing
for a new perspective for neglected works of American
fiction like Uncle Tom’s Cabin, makes the point that
when.one sets aside modernist demands — for psycho
logical complexity, moral ambiguity, epistimological
sophistication, stylistic density, formal economy —
and attends to the way a text offers a blueprint for
survival under a specific set of political, economic,
social, or religious conditions, an entirely new story
begins to unfold, and one's sense of the formal exi
gencies of narrative alters accordingly, producing a
different conception of what constitutes successful
characters and plots. (Design xvii-xviii)
Thus in the case of "Thurso’s Landing," we need only
ask if these characters — are part of a different,
unconventional design, ultimately rhetorical, perhaps
persuasive or expository. Thus within an expository
schema, readers who do not find complexity in the
characters will focus on how the characters function as
variables within an inhumanist schema which in turn gives
194
readers the leisure to shift their attention to the
not-human world. In a sense, the events and characters
become examples of inhumanist tenets, inviting readers to
organize them less by conventional narrative schemata and
more by some form of flexible expository schema. Thus
Jeffers’ work, once again, seems to push against the
boundaries of conventional schemata; what appears to be an
aesthetic work, evocative of poetic and narrative elements,
is an efferent one as well, the aggregate organized by
means of a high-level schema that incorporates all of these
elements.
Earlier I claimed that the work of Robinson Jeffers
was particularly appropriate as a subject by which to test
a theory of criticism because previous critical approaches
were actually arguments according to their own criteria for
whether or not Jeffers' work holds any literary interest at
all. Formalist critics like Winters believe it does not;
psychological critics like Brophy and Zaller believe that
it does — each critic judging the value of Jeffers' work
in terms of their own critical schema. As a final argument
for the practical advantage of a schema-theoretic,
reader-centered criticism, I wish to adjudicate the issue
by arguing for a different standard by which the validity
of interpretations and critical judgments may be decided,
195
The previous analysis of the criticism of "Thurso's
Landing" has accomplished at least two things. First, it
makes clear that the responsible critics were themselves
responding to the work according to their own sometimes
unexamined assumptions about what constitutes good
literature. The schema-theoretic approach thereby argues
that the judgments of value directed to particular works
are formed within a particular ecology that is itself in
constant flux; as a result, the place of Jeffers' works
within the canon will change with the critical environment
— as has the work of most other writers. As long as
formalist approaches that value tight and logical
structures are prevalent, poems like "Thurso's Landing" can
expect little critical sympathy; but with the ascension of
structuralist approaches like Brophy's which argue for the
adherence of the text to a pre-valued set of schemata (like
archetypes), poems like Jeffers' may be seen as more
valuable insofar as they conform to those schemata. The
schema-theoretic approach is a heuristic by which to
compare the difference in critical assumptions, to make
them explicit and examine them within the current critical
environment. Contemporary critical opinion with its
celebration of principled differences in the reading of
literature would probably reject the formalism inherent in
both approaches, Winters' and Brophy's.
196
Second, the schema-theoretic approach is a useful
method to further test the adequacy of previous criticism
since it provides a heuristic by which to look with a new
perspective at the critical problems already identified in
literary texts as well as problems not so historically
defined, I believe that the insights into the prosody,
mode of organization, and the philosophy of "Thurso's
Landing" achieved by means of the schema-theoretic approach
have together formed a coherent reading, interesting and
useful in and of itself. But more importantly, I believe
that these insights and the method by which they were
developed serves the work more completely because the
method is applicable to the analysis of criticism and
literature together, recognizing that critics extend the
works they read by means of their own writing, making it
possible to examine within a common frame the work and its
criticism and to treat this aggregate as a macro-text.
But finally these advantages are as nothing without
some claim for the validity of the approach. I believe
that a schema-theoretic criticism can make that claim —
not on grounds of the need for apprehending the intention
of the writer or for the demands of the text — but on the
basis of utility, since the schema-theoretic approach to
"Thurso’s Landing" best accounts for the complex texture of
the work, its potential as a document by which the
197
prevailing conventional schemata for reading are
questioned, emended, and used in different ways, as a
document that ultimately deserves inclusion in the canon of
important literary works because of this potential. The
formalist approach was unable to appreciate much of this
potential, and the psychological, religious, and
philosophic critics have each argued for a portion of it.
But the schema-theoretic approach, by means of recognizing
the power of readers and the social environments in which
they read, is a means by which to begin to appreciate its
enormous complexity.
Jeffers once wrote, "Poetry is less bound by
circumstances than any other of the arts; it does not need
tangible materials; good poetry comes almost directly from
a man’s mind and senses and bloodstream, and no one can
predict the man" ("Gongerism" 16). I will concede that no
one can predict the man, but as for his other point,
Jeffers was dead wrong: poetry, like any mode of discourse,
is certainly bound by circumstances. The reputations of
few other poets demonstrate this so clearly as does that of
Robinson Jeffers, a writer of unusual forms in an age of
increasing critical formalism, of trans-human philosophic
concerns in an age of psychoanalysis and technology. The
advantage of the schema-theoretic approach to literary
criticism is that it also examines circumstances, making
198
these the subject of critical judgment and thereby making
it possible to set them aside and enjoy works like
"Thurso's Landing" that have been so indifferently served
by prevailing critical fashions.
199
NOTES
Chapter I
For example, Steven Mailloux points out that three of
the most influential reader-response critics — Fish, Iser,
and Culler — all prefer a temporal model of the reading
experience, describing a succession of actions in the
reading process to show that form has a temporal dimension
(96).
2
Dewey’s anti-objectivist position is clear in a
chapter of Art as Experience entitled significantly "The
Human Contribution": "Experience is a matter of the
interaction of organism with its environment, an
environment that is human as well as physical, that
includes the materials of tradition and institutions as
well as local surroundings. The organism brings with it
through its own structure, native and acquired, forces that
play a part in the interaction. The self acts as well as
undergoes, and its undergoings are not impressions stamped
upon an inert wax but depend upon the way the organism
reacts and responds" (246).
3
Husserl’s phenomenology is based upon the
relationship between consciousness and the objects to which
it is directed. Husserl calls this relationship
intentionality, and he maintains that the quality of the
intentional relationship will affect how an object is
perceived by consciousness. See Logical Investigations and
Ideas Pertaining to a Pure Phenomenology.
4 t
Holland discounts Bleich s charge in a College
English exchange: "... 5 Readers Reading says^ * Id and
superego, drive and defense, look less like psychic
entities in themselves, more like functions, and still more
like ways of looking at one central relation, between the
individual ... and the other' (127)" (Bleich and Holland
298) .
^For example Wayne Booth’s Rhetoric of Fiction argues
for elements iji a work that are "friends of the reader"
(106) and that minister to the needs of that reader (108);
200
and Walker Gibson’s Tough, Sweet and Stuffy argues that
particular stylistic elements in a literary work require a
certain type of ’’assumed writer” which speaks to a
particular sort of reader (7-13). Both critics thereby
resist objectivist claims that the work is independent of a
larger rhetorical situation, but both critics focus on the
role of the work in controlling that situation.
Fish says of affective stylistics: ’’More than any
other way of teaching I know, it breaks down the barriors
between students and the knowledge they must acquire, first
by identifying that knowledge with something that they are
already doing, and then by asking them to become
self-conscious about what they do in the hope that they can
learn to do it better (22).
^See Saussure’s Course in General Linguistics: ’’The
bond between the signifier and the signified is arbitrary"
(67); "...every means of expression used in society is
based, in principle, on collective behaviour or — what
amounts to the same thing — on convention" (68).
Chapter II
^"Although the main text of Validity in Interpretation
is not clear about the nature of this reliance, a 1960
discussion "Objective Interpretation" is included as an
appendix (209-244), making this point explicit.
2
See Elements of Semiology, which extends Saussurean
linguistics to a theory of semiology powerful enough to
describe linguistic and non-linguistic human signifying
systems. Although the volume is inspired by Saussure,
Barthes includes analysis of relevent work by others,
including Peirce.
3
Barthes may have been referring to the work of Propp,
whose analysis of Russian folktales isolates 31 separate
elements that perform the same function in any tale. Some,
according to the analysis, are essential, some optional,
some mutually exclusive (e.g. the struggle with the villain
and the difficult task). Propp feels that any tale can be
explained by isolating its functional morphemes, making it
possible to compare tales on the basis of stable functions
instead of variable content or specific characters. His
analysis, besides emphasizing the similarity of tales, is
structural and thus atemporal, which leads him open to
201
critics such as Jameson who decry his lack of a diachronic
element that accounts for the sequence of events (69),
Chapter III
"''For example, see Beck (60); Gibson and Levin
(445-49); Samuels and Kamil; and Smith (231). Beck's review
of Pearson's Handbook of Reading Research shows that
Gough’s model is still actively discussed in graduate
courses in reading research, and the inclusion of such a
model in the discussion of reading models by Samuels and
Kamil confirm its contemporary importance. Samuels and
Kamil assert that all models of the reading process have
holes since they focus on different aspects of reading;
since there is no one comprehensive model, they point out
the necessity of considering a range of alternatives in
order to appreciate the required scope of such a model
(220) .
2
The term "data driven" is opposed to "concept driven"
by Bobrow and Norman; "Any given process is either data- or
resource- limited,...We believe the system to be driven
both by the data (in a bottom-up fashion) and conceptually
(in a top-down fashion)" (148). The terms are also used by
Sanford and Garrod (101). I prefer these terms to the more
prevalent terms (in reading research) of bottom—up and
top-down processing, because they avoid the unfortunate
value connotations that attach to the directional metaphor
of top vs,bottom, and they are more congenial to the
description of lateral movement which may be instigated by
data or concept driven processing; e.g. Adams and Collins
assert that "bottom-up" processing occurs when schemata
suggest other schemata at the same level or further up
(20), an assertion that seems at odds with the terminology
they use.
3
Schank and Abelson outline four example programs
(175-221), of which two are SAM (Script Applier Mechanism)
and PAM (Plan Applier Mechanism) which they feel are
prototypes for a general text understanding program (176).
Also see Dehn (82) for a discussion of PAM's capabilities
in particular and of the perspective of AI on comprehension
in general.
^Rumelhart cites Fillmore ("Alternative"), Minsky,
Rumelhart and Ortony, and Schank and Abelson as important
proponants of similar theories of understanding based upon
the representation of knowledge. To this list I would add
202
Adams and Collins, as well as Anderson and Pearson, for
their articles that present schema theory in the specific
context of reading research.
“*The description of schema theory I will offer below
follows closely and depends principally upon descriptions
by Rumelhart ("Building Blocks" and "Understanding") and
Rumelhart and Ortony, except where supplemented by other
sources noted in the text.
£
Rumelhart and Ortony assert that Charles Fillmore's
case grammar is a parallel in linguistics to schema
theory. Fillmore's case grammar (see "The Case for Case")
is in a sense a schema theory of verbs and their
arguments. Propositions typically take a verb and a number
of arguments, labelled by means of semantic roles.
^For example, Rumelhart (Bobrow and Norman) has
suggested a type of story grammar that is based upon
schemata, but that takes its real genesis from the
transformational-generative grammars and from Propp:
"...the approach developed here is, to some degree,
designed to be a systematization of Propp's analysis of
Russian folk tales" (235). The problem, as Rumelhart
realizes elsewhere ("Understanding" 4) is that the analysis
ignores much of what readers bring to the text, though it
captures something of how readers organize what they find
there. More specifically, Collins et al. assert that what
is missing from story grammars like Rumelhart's is a notion
of planning knowledge (see Schank and Abelson).
8
"The basic experimental paradigm involved presenting
subjects a series of stories a sentence at a time and,
after each sentnce, asking them WHO they thought the
characters under discussion were, WHAT they felt was going
on in the story, WHY the characters behaved as they did,
WHEN they think the event took place, and WHERE they think
the story is set" (5).
g
Kintsch and van Dijk's characterization of
propositions parallels closely Fillmore's characterization
of verbs and their arguments, which as already noted (see
note 4 above) is a description of syntax consistent with
schema theory.
See Witte and Faigley for a discussion of textual
coherence (based on work by Halliday and Hasan) that
characterizes more specifically types of cohesive ties
across sentence boundaries.
203
Kintsch and van Dijk justify bracketing the problem
of how propositions are derived by citing other research
that describes the process (367),
12
See especially Chomsky (135-136) for the semantic
importance of deep structure.
Chapter IV
■^Mailloux (IC 136) attributes this point to John
Reichert (Making Sense of Literature [Chicago: Univ. of
Chicago Press, 1977], 155-65).
2
I rely here on Lewis' definition of convention as
developed in Convention: A Philosophical Study and
"Languages and Language" for two principal reasons. First,
Lewis' precise description of the term has held up to
criticism and review (see Tyler Burge's criticism of Lewis'
definition and Richard Grandy’s review of Convention).
Second, Mailloux extends the definition to cover his
category of "traditional" conventions (IC 127),
3
To preserve precision and to avoid confusion
throughout this discussion, I wish to reintroduce a
distinction drawn earlier between the term document — the
physical object with marks, pages, etc. — and the term
text — the theoretical object constituted by a reader and
interpreted by her.
Chapter V
For a more complete analysis of Jeffers' Inhumanism, see
Arthur Coffin’s Robinson Jeffers: Poet of Inhumanism and
Squires' study, The Loyalties of Robinson Jeffers. Coffin's
thesis is that the proper study for the reading of Jeffers
is Nietzsche's Also Sprach Zarathustra and Lucretius’ De
Rerum Natura (257-258); he traces the emergent philosophy
of inhumanism throughout Jeffers' career, dividing it into
three periods: 1) an awareness of human shortcomings
(informed by Nietzsche and confirmed by the discoveries of
modern science and constrained by religious orthodoxy); 2)
a period of application and verification; 3) a period of
confidence in light of concurrent historical developments.
Squires' earlier study traces the additional influence of
204
Spengler in Jeffers' idea of culture cycles (57), See
Jeffers' Themes in My Poems for a succinct explication of
this idea (18),
2
The first book length study of Jeffers and his work,
Powell's Robinson Jeffers: The Man and His Work documents
Jeffers' background as a medical student at the University
of Southern California and as a forestry student at the
University of Washington, Moreover, astronomy exerted a
powerful influence on Jeffers through his brother Hamilton,
an astronomer at the Lick Observatory,
See H.H. Waggoner for a discussion of Jeffers’
scientific diction: "Although Jeffers consciously attempts
to avoid the borrowing of technical terms from science, it
is inevitable that poetry so imbued with scientific theory,
created by a mind so familiar with scientific terminology,
should contain many words of a scientific flavor, and some
that are no doubt vague or meaningless to the layman unread
in the sciences" (277).
4 . .
Powell quotes from Herbert Klein’s Master’s Thesis,
"The Prosody of Robinson Jeffers." Occidental College,
1930,
^Also see Everson's explication of a short Jeffers
poem, "Post Mortem" (21-56). Everson's purpose, in part, is
to show how well Jeffers' work stands up to the type of
analysis habitually performed by New Critics. During this
analysis, he necessarily takes into account the prosody of
this Jeffers lyric.
6
More precisely, readers may create for themselves by
means of reading the qualities in the work which form
variables within a particular schema. For purposes of
simplicity, I will often refer to elements of "Thurso's
Landing" and other Jeffers’ works without specifically
mentioning that these elements are, according to the theory
for which I have argued, perceptions of and thereby
constructions of a reader.
205
WORKS CONSULTED
Adams, Marilyn and Allan Collins. "A Schema-Theoretic View
of Reading,” Freedle 1-22.
Alberts, S.S, A Bibliography of the Works of Robinson
Jeffers. 1933, New York: Burt Franklin, 1968.
Alexander, John R. "Conflict in the Narrative Poetry of
Robinson Jeffers," Sewanee Review 30 (1972): 85-99,
Anderson, Richard C. and P. David Pearson. "A
Schema-Theoretic View of Basic Processes in Reading
Comprehension." Pearson 255-291.
Antoninus, Brother (William Oliver Everson). Robinson
Jeffers: Fragments of an Older Fury. Berkeley: Oyez,
1968.
_____________. Foreword. The Double Axe, By Robinson Jeffers.
New York: Liveright, 1977,
_____________, "A Tribute to Robinson Jeffers." Critic 20
(1962): 14-16,
Bakhtin, M. M. The Dialogic Imagination. Ed. Michael
Holquist. Austin: Univ. of Texas Press, 1981,
Barthes, Roland. "The Death of the Author." Sears and Lord
7-12.
____________ , "Introduction from Writing Degree Zero. Sears
and Lord 3-6.
_____________. Elements of Semiology, Trans. Annette Lavers
and Colin Smith, New York: Hill and Wang, 1968,
_____________. "Style and Its Image." Literary Style: A
Symposium. Ed. Seymour Chatman. New York: Oxford Univ.
Press, 1971, 3-10.
_____________. S/Z. Trans. Richard Miller, New York: Hill and
Wang, 1974,
206
_____________. "To Write: An Intransitive Verb?" Ed, Richard
Macksey and Eugene Donato. The Languages of Criticism
and the Sciences of Man. Baltimore: Johns Hopkins
Univ. Press, 1970. 134-156.
Bartlett, F.C. Remembering. Cambridge: Cambridge Univ.
Press, 1932.
Beck, Isabel. Rev. of Handbook of Reading Research, ed.
David Pearson. Journal of Reading Behaviour 17,1
(1985): 55-71.
Benet, William Rose, Rev. of Solstice and Other Poems, by
Robinson Jeffers. Saturday Review of Literature 13
(1935): 20.
_____________. "Jeffers’ Latest Work." Rev, of Thurso * s
Landing and Other Poems, by Robinson Jeffers. Saturday
Review of Literature 8 (1932): 638.
Bernstein, Richard. Beyond Objectivism and Relativism:
Science, Hermeneutics, and Praxis. Philadelphia: Univ.
of Pennsylvania Press, 1983,
Blackmur, R.P. "Lord Tennyson's Scissors: 1912-1950," The
Kenyon Review 14 (1952): 1-20,
Bleich, David, "Pedagogical Directions in Subjective
Criticism." College English 37 (1976): 454-461.
_____________. Subjective Criticism. Baltimore: Johns Hopkins
Univ. Press, 1978,
Bleich, David and Norman Holland. "Comment and Response."
College English 38 (1976): 298-301.
Bobrow, D.G. and Allan Collins, eds. Representation and
Understanding: Studies in Cognitive Science. New York:
Academic Press, 1975.
Bobrow, D.G. and Donald A. Norman, "Some Principles of
Memory Schemata." Bobrow and Collins 131-150,
Booth, Wayne. Rhetoric of Fiction. Chicago: Univ. of
Chicago Press, 1961.
Bransford, Robert, et al, "Toward Unexplaining Memory."
Shaw and Bransford 431-66.
207
Brophy, Robert. Robinson Jeffers; Myth, Ritual and Symbol
in His Narrative Poems. 1973. Hamden, CT: Archon
Books, 1976. ^
Burge, Tyler. "On Knowledge and Convention." Philosophical
Review 84 (1975): 249-255.
Burhans, Clinton S. "The Complex Unity of In Our Time."
Modern Fiction Studies 14 (1968): 313—328,
Burke, Kenneth. Counter—Statement. Berkeley: Univ. of
California Press, 1968,
Carpenter, Frederic I. "’Post Mortem': ’The Poet Is Dead.’"
Western American Literature 12 (1977): 3-10,
_____________. Robinson Jeffers. New York: Twayne, 1962,
_____________. "The Values of Robinson Jeffers." American
Literature 11 (1940): 356-366.
Chafe, Wallace, "Givenness, Contrastiveness, Definiteness,
Subjects, Topics, and Point of View." Subject and
Topic. Ed, Charles N. Li, New York: Academic Press,
1976. 27-55.
Chandler, Raymond. The High Window. New York: Vintage
Books, 1976.
Chomsky, Noam. Aspects of a Theory of Syntax. Cambridge,
MA: MIT Press, 1965.
Coffin, Arthur B. Robinson Jeffers: Poet of Inhumaninsm.
Madison: University of Wisconsin Press, 1971.
Collins, Allan, John Seely Brown, and Kathy M. Larkin.
"Inference in Text Understanding." Spiro, et al.
385-407.
Comrie, Bernard. Language Universals and Linguistic
Typology. Chicago: Chicago Univ. Press, 1981,
Cooper, Marilyn. "Grice Revisited." MLA Convention.
Washington D.C., 1984.
Culler, Jonathan. On Deconstruction. Ithaca: Cornell Univ.
Press, 1982.
_____________. Roland Barthes. New York: Oxford Univ. Press,
1983,
_____________. Structuralist Poetics. Ithaca: Cornell Univ.
Press, 1975,
de Beaugrande, Robert, "The Linearity of Reading: Fact,
Fiction, or Frontier?" Flood 45-74.
____________. "Writer, Reader, Critic: Comparing Critical
Theories as Discourse." College English 46 (1984):
533-599.
Dehn, Natalie, "An AI Perspective on Reading
Comprehension." Flood 82-100,
Deutsch, Babette. "Brains and Lyrics," Rev. of Tamar and
Other Poems, by Robinson Jeffers, New Republic 153
(1925): 23-24.
_____________. "The Hunger for Pain." Rev, of Thurso’s Landing
and Other Poems, by Robinson Jeffers, New York Herald
Tribune Books 27 March 1932: 7.
_____________, "In Love With the Universe," Rev. of Solstice
and Other Poems, by Robinson Jeffers. New York Herald
Tribune Books 27 Oct. 1935: 8.
Dewey, John. Art as Experience. 1934, New York: Capricorn
Press, 1958,
Dillon, George. Constructing Texts: Elements of a Theory of
Composition and Style. Bloomington: Indiana Univ.
Press, 1981,
Durkin, Dolores. "What Is the Value of the New Interest in
Reading Comprehension?" Language Arts 58 (1981):
23-43,
Eagleton, Terry. Literary Theory: An Introduction.
Minneapolis: Univ. of Minnesota Press, 1983,
Eco, Umberto. The Role of the Reader. Bloomington: Indiana
Univ. Press, 1979.
Encyclopedia of Philosophy. 8 vols. New York: Macmillan,
1967.
Everson, William. See Brother Antoninus.
Faulkner, William. The Sound and the Fury. New York:
Vintage, 1954.
209
Fillmore, Charles, "An Alternative to Checklist Theories of
Meaning," Procedings of the First Annual Meeting of
the Berkeley Linguistics Society 1 (1975): 123-131,
_____________, "The Case for Case," Universals in Linguistics
Theory, Ed. E. Buch and R. Harms. New York: Holt,
Rinehart, and Winston, 1968. 1-88.
Fish, Stanley, Is There a Text in This Class? Cambridge,
MA: Harvard Univ. Press, 1980.
_____________. Surprised by Sin. New York: Macmillan, 1967.
Flood, James, ed. Understanding Reading Comprehension.
Newark, DL: International Reading Association, 1984.
Freedle, R. 0,, ed. New Directions in Discourse
Processing. Norwood, NJ: Ablex Publishing Co., 1979.
Gibson, Eleanor J. and Harry Levin. The Psychology of
Reading. Cambridge, MA: MIT Press, 1975.
Gibson, Walker. Tough, Sweet, and Stuffy. Bloomington:
Indiana Univ. Press, 1966,
Goffman, Erving. Frame Analysis. Cambridge, MA: Harvard
Univ. Press, 1974,
Goodman, Kenneth. "Reading: A Psycholinguistic Guessing
Game." Singer and Ruddell 497-508,
Gough, Phillip. "One Second of Reading," Singer and Ruddell
509-535.
Grandy, Richard. Rev. of Convention: A Philosophical Study,
by David Lewis. Journal of Philosophy 74 (1977):
129-139,
Grice, H.P. "Logic and Conversation," 1967 William James
Lectures. Syntax and Semantics 3: Speech Acts. Ed.
Peter Cole and Jerry Morgan. New York: Academic Press,
1975. 41-58.
_____________, "Meaning Revisited," Mutual Knowledge. Ed. N.
V, Smith. London: Academic Press, 1982. 223-243,
Hawkes, John. The Cannibal. New York: New Directions, 1962,
Hawkes, Terence. "Taking It As Read," Yale Review 69
(1980): 560-576.
210
Hicks, Granville, "A Transient Sickness ,” Rev. of Thurso * s
Landing and Other Poems, by Robinson Jeffers. Nation
134 (1932): 433.
Hirsch, E. D, "Meaning and Significance Reinterpreted,"
Critical Inquiry 11 (1984): 202-225,
_____________, "Past Intentions and Present Meanings." Essays
in Criticism. 33 (1983): 79-98.
_____________, Validity in Interpretation, New Haven: Yale
Univ. Press, 1967.
Holland, Norman, Poems in Persons. New York: Norton, 1973.
_____________. "Unity Identity Text Self," Tompkins 118-133.
Holmes, Betty C. "The Effect of Prior Knowledge on the
Question Answering of Good and Poor Readers." Journal
of Reading Behaviour 15,4 (1983): 1—18,
Humphries, Rolfe. "Two Books by Jeffers." Rev, of Thurso1s
Landing and Other Poems and Descent to the Dead, by
Robinson Jeffers. Poetry 40 (1932): 154-58,
Hunt, Tim. Afterward. The Women at Point Sur. By Robinson
Jeffers. New York: Liveright, 1977.
Husserl, Edmund. Ideas Pertaining to a Pure Phenomenology.
Trans. F. Kersten. The Hague: Martinus Nijhoff, 1983.
_____________. Logical Investigations. Trans. J, N. Findlay,
London: Kegan Paul, 1970.
Iser, Wolfgang. The Act of Reading. Baltimore: Johns
Hopkins University Press, 1978,
_____________. "The Reading Process: A Phenomenological
Approach." Tompkins 50-69,
Jameson, Frederic. The Prison-House of Language. Princeton:
Princeton Univ. Press, 1972,
Jarrell, Randall. The Third Book of Criticism. New York:
Farrar, Straus, and Giroux, 1969.
Jeffers, Robinson. "Poetry, Gongerism, and a Thousand
Years." The New York Times 18 Jan. 1948, sec. 6: 18,
26.
211
_____________. The Selected Letters of Robinson Jeffers, Ed,
Ann N. Ridgeway. Balitimore: Johns Hopkins Univ.
Press, 1968.
_____________. The Selected Poetry of Robinson Jeffers. New
York: Random House, 1959,
_____________. Themes in My Poems. San Francisco: Book Club of
California, 1956.
Kant, I, Critique of Pure Reason, Trans. N. Kemp Smith.
London: Macmillan, 1963.
Katz, Jerrold and Jerry Fodor. "The Structure of a Semantic
Theory." Language 39 (1963): 170-210.
Kermode, Frank. "Figures in the Carpet: On Recent Theories
of Narrative Discourse." Comparative Criticism 2
(1980): 291-302.
Kinneavy, James. "The Basic Aims of Discourse," The Writing
Teacher’s Sourcebook, Ed. Gary Tate and Edward P.J,
Corbett. New York: Oxford Univ. Press, 1981. 89-99,
Kintsch, Walter and Teun van Dijk. "Toward a Model of Text
Comprehension and Production." Psychologiical Review
85 (1978): 363-394.
Koenke, Karl, "ERIC/RCS Report: An Examination of the
Construct of ’Reader-Text Relationship.’" English
Education 16 (1984): 115-120,
Labov, William. "The Boundaries of Words and Their
Meanings." Analyzing Variation in Language, Ed, R,W.
Fasold and R.W. Shuy. Washington, D.C.: Georgetown
Univ. Press, 1973. 340-373,
_____________, Language in the Inner City, Philadelphia:
University of Pennsyvania Press, 1972.
Lewis, David. Convention: A Philosophical Study. Cambridge,
MA: Harvard Univ. Press, 1969.
_____________. "Languages and Language," Minnesota Studies in
the Philosophy of Science. Ed. K. Gunderson 7 (1975):
3-35.
Mailloux, Steven, Interpretive Conventions. Ithaca: Cornell
Univ. Press, 1982,
212
______ . "Learning to Read: Interpretation and Reader-
Response Criticism." Studies in the Literary
Imagination 12 (1979): 93-108,
Mauthe, Andrew K. "Jeffers' Inhumanism and Its Poetic
Significance." Robinson Jeffers Newsletter. 26 (1970):
8-10.
McGann, Jerome, Ulysses as a Post-Modern Text: the Gabler
Edition. Criticism 27 (1985): 283-306.
Mencken, H.L. "Books of Verse." American Mercury 8 (1926):
251-54,
Milton, John. Complete Poems and Major Prose. Ed. Merritt
Y. Hughes. Indianapolis: Bobbs-Merrill, 1957.
Minsky, Marvin. "A Framework for Representing Knowledge."
The Psychology of Computer Vision. Ed. Patrick Henry
Winston. New York: McGraw-Hill, 1975, 211-277,
Monjian, Mercedes Cunningham, Robinson Jeffers: A Study in
Inhumanism. Pittsburgh: University of Pittsburgh
Press, 1958.
Monroe, Harriet. "Power and Pomp." Rev. of Roan Stallion,
Tamar and Other Poems, by Robinson Jeffers. Poetry 28
(1926): 160-64,
Mukarovsky, Jan. "Standard Language and Poetic Language,"
Linguistics and Literary Style. Ed. Donald C. Freeman.
New York: Holt, Rinehart and Winston, 1970. 40-56,
Nabokov, Vladimir. Pale Fire. New York: Perigee Books,
1962.
Nolte, William H. Rock and Hawk: Robinson Jeffers and the
Romantic Agony. Athens, GA: The University of Georgia
Press, 1978.
Ortony, Andrew, "Metaphor." Spiro, et al. 349-365,
Pearson, David, ed. Handbook of Reading Research. New
York: Longman, 1984.
Polyani, Livia. "The Nature of Meaning of Stories in
Conversation."Studies in Twentieth Century Literature
1 (1981): 51-65,
Powell, Lawrence Clark. Robinson Jeffers: The Man and His
Work. Los Angeles: Primavera Press, 1934.
213
Pratt, Mary Louise, Toward a Speech Act Theory of Literary
Discourse. Bloomington: Indiana Univ. Press, 1977.
Propp, V. The Morphology of the Folk Tale. Trans. Lawrence
Scott. Austin: Univ. of Texas Press, 1968.
Ransom, John Crowe. The New Criticism. Norfolk, CT: New
Directions, 1941,
Rexroth, Kenneth. "In Defense of Jeffers." Saturday Review
of Literature 40 (1957): 30.
Rorty, James. "In Major Mold," Rev. of Tamar and Other
Poems, by Robinson Jeffers. New York Herald Tribune
Books 1 March, 1925; 1-2.
_____________. "Symbolic Melodrama." Rev. of Thurso’s Landing
and Other Poems, by Robinson Jeffers. New Republic 71
(1932): 24-25.
Rosenblatt, Louise. "The Poem as Event." College English 26
(1964): 123-128.
_____________, The Reader, the Text, the Poem. Carbondale:
Southern Illinois University Press, 1978.
_____________. "Towards a Transactional Theory of Reading."
Journal of Reading Behaviour 1 (1969): 31-49.
_____________, "Viewpoint: Transaction versus Interaction."
Research in the Teaching of English 19 (1985): 9
6-107,
Rumelhart, David. Introduction to Human Information
Processing. New York: Wiley, 1977.
_____________. "Notes on a Schema for Stories." Bobrow and
Collins 211-236.
_____________, "Schemata: The Building Blocks of Cognition."
Spiro, et al, 33-58.
_____________, "Understanding Understanding." Flood 1-20.
Rumelhart, David and Andrew Ortony, "The Representation of
Knowledge in Memory." Schooling and the Acquisition of
Knowledge, Ed. R.C. Anderson, R.J, Spiro, and W.E.
Mantague. Hillsdale, NJ: Lawrence Erlbaum, 1977.
214
Samuels, S. Jay and Michael L. Kamil, "Models of the
Reading Process." Pearson 185-224,
Sanford, Anthony and Simon C. Garrod. Understanding Written
Language. New York: Wiley, 1981.
Saussure, Ferdinand de. Course in General Linguistics.
Trans. Wade Baskin. New York: McGraw Hill, 1966.
Schank, Roger. Reading and Understanding: Teaching From the
Perspective of Artificial Intelligence. Hillsdale, NJ:
Lawrence Erlbaum, 1982.
Schank, Roger and Robert Abelson. Scripts, Plans, Goals,
and Understanding. Hillsdale, NJ: Lawrence Erlbaum,
1977.
Scott, Robert Ian. "From Berkeley to Barclay's Delusion:
Robinson Jeffers vs. Modern Narcissism," Mosaic 15:3
(1982): 55-61,
Sears, Sallie and Georgianna W. Lord, eds. The
Discontinuous Universe, New York: Basic Books, 1972,
Shaw, Robert and John Bransford, eds. Perceiving, Acting,
and Knowing: Toward an Ecological Psychology. New
York: Lawrence Erlbaum, 1977.
_____________. "Psychological Approaches to the Problem of
Knowledge." Shaw and Bransford 1-39.
Singer, Harry and Robert B. Ruddell, eds. Theoretical
Models and Processes of Reading. 2nd ed, Newark, DL:
International Reading Association, 1976.
Smith, Frank. Understanding Reading. 3rd ed. New York:
Holt, Rinehart, and Winston, 1982.
Spiro, Rand J. "Constructive Processes in Prose
Comprehension and Recall," Spiro, et al. 245-278.
Spiro, Rand J. et al, Theoretical Issues in Discourse
Comprehension. Hillsdale, NJ: Lawrence Erlbaum, 1980.
Squires, James Radcliffe. The Loyalties of Robinson
Jeffers. Ann Arbor: University of Michigan Press,
1956.
215
Sullivan, William and Paul Rabinow, "The Interpretive Turn:
Emergence of an Approach," Philosophy Today 23 (1979):
29-40.
Tompkins, Jane P. ed, Reader-Response Criticism.
Baltimore: Johns Hopkins Press, 1980.
_____________. Sensational Designs: The Cultural Work of
American Fiction 1790-1860. Oxford: Oxford Univ.
Press, 1985.
Untermeyer, Louis. "Five Notable Poets.” Yale Review 21
(1932): 811-817.
Van Dijk, Teun. "Cognitive Processing of Literary
Discourse." Poetics Today 1-2 (1979): 143-159.
Van Doren, Mark. "First Glance." Rev. of Tamar and Other
Poems, by Robinson Jeffers. Nation 120 (1925): 268.
_____________. "First Glance." Rev. of The Women at Point Sur,
by Robinson Jeffers. Nation 125 (1927): 88,
Vardamis, Alex A. The Critical Reputation of Robinson
Jeffers. Hamden, CT: Archon Books, 1972,
Waggoner, Hyatt Howe. "Science and the Poetry of Robinson
Jeffers." American Literature 10 (1938):275-288.
Wimsatt, W.K. and Monroe C. Beardsley. "The Affective
Fallacy." 20th Century Literary Criticism. Ed. David
Lodge. London: Longman, 1972. 345-358.
_____________. "The Intentional Fallacy." Lodge 334-345.
Winters, Yvor. In Defense of Reason. Denver: Univ. of
Denver Press, 1947.
_____________. "Robinson Jeffers," Poetry 35 (1930): 279-286.
_____________. Rev. of Thurso's Landing and Other Poems, by
Robinson Jeffers. Hound and Horn 5 (1932): 681,
684-685.
Witte, Stephen P. and Lester Faigley. "Coherence, Cohesion,
and Writing Quality." CCC 32 (1981): 189-204.
Wittgenstein, Ludwig. Philosophical Investigations. Trans,
G.E.M. Anscombe. 3rd ed. New York: Macmillan, 1958,
216
Zaller, Robert. The Cliffs of Solitude. Cambridge:
Cambridge University Press, 1983,
217

Abstract (if available)

Linked assets