Cognitive Poetics

8. 1. Roots and Margins

8. 2. Beyond Reader Response Theory
8. 3. Stylistics and Rhetoric Revisited
8. 4. Cognitive Science—More or Less
8. 5. Overview

1. 1. Roots and Margins

If cognitive science has its roots in anthropology, philosophy, neuroscience and neurology,
computer science and artificial intelligence, it has also grown to be much indebted to
psychology and linguistics and also much involved in literary studies, with stylistics,
pragmatics, semantics and language studies in general as favorite fields of investigation.
Most of the contributions that have determined the development of such a discipline come
from metaphor theory, stylistics, and narratology, and owe very much to such more
specialized investigations as image schemas (Lakoff and Johnson 1980, 1987, 1990, 1991,
1999, Langacker 1987, 1990, 1991, Shank & Abelson 1977) or blending theory (Fauconnier
1985, Fauconnier and Turner 2002, with their mental space theory), to deixis, schema
poetics and frames (Stockwell 2002), the figure/ground distinction, prototypes (Rosch 1988
and Lakoff 1987) or to the description of text worlds (Paul Werth 1999).
This is, of course, much more than we can cover in this space, so our presentation—
since the domain of cognitive poetics is far from being very well defined, and much less
clearly circumscribed—we will confine ourselves to introducing two positions (that
sometimes seem to be in disagreement)—Peter Stockwell’s (who thinks himself that
“cognitive poetics, the application of cognitive science to illuminate the study of literary
reading, is maturing as a discipline,” “Cognitive Poetics and Literary Theory,” web article)
and Reuven Tsur’s, with David Miall and others in-between.

8. 2. Beyond Reader Response Theory

The above statement from Stockwell—“to illuminate the study of literary reading”—
already points out that cognitive poetics is basically—and for the time being—a poetics of
the reading processes and the way audiences (Fish’s “interpretive communities”) respond
to literary texts, something that Jauss, Fish again, and Iser had explored in their reader
response investigations. Thus we are still at the end of the reception stage in the process of
literary communication, with creativity still in the background and at some distance too.
Reuven Tsur—who may be the first to have used the concept of “cognitive poetics,” as
far back as the nineteen-sixties—admits from the very beginning that the disciplines of
literary criticism, literary history, linguistics and aesthetics are still very much to be taken
account of; one of the problems that we will encounter again and again is where we draw
the line between cognitive poetics and interpretation, and Tsur does not prove to be of
much help here: cognitive poetics, he announces, “attempts to find out how poetic language
and form, or the critic’s decisions, are constrained and shaped by human information
processing,” or how one can account for the relationship between the structure of literary
texts and their perceived effects. (web page) His book Toward a Theory of Cognitive
Poetics (1992) attempts, as a matter of fact, to illuminate the aspects of poetic structure on
a wide variety of strata and from multiple angles: the sound stratum of poetry, the units-of-
meaning stratum, and the world stratum, with literary theory, once again, period style,
stylistic typology, archetypal patterns, genre, etc. in the background. Large sections are
devoted to poetry and altered states of consciousness and to another new concept, that of
the implied critic (side by side with the older implied author and implied reader) and his
mental dictionary.
One of Tsur’s main assumptions is neuropsychological in nature and origin: language is
a predominantly sequential activity, of a logical character, and as such is known to be
associated with the left hemisphere of the brain; on the other hand, the emotional processes
that poetry is supposed to and does stir are reputedly placed in the right hemisphere. Thus
the question is—and he proposes several case studies—how emotional qualities can be
conveyed by poetry; so, the effects of poetry once again, and the job of the cognitive
poetician is to relate these effect to particular features and regularities that occur in the
literary texts. And so the interest falls on the cognitive correlates of poetic processes which
include the normal cognitive processes (that were “initially”—difficult concept here—
evolved in poetry for non-aesthetic purposes, but rather historical, cultural, ethical, etc.), a
modification of these processes by feelings and emotions aroused during reading (or
creating?—poetry as “emotion recollected in tranquility”), and their reorganization
according to new principles. If one pays careful attention, the point here is still dependent
on the interpretability of poetic language, with a particular emphasis upon the diffuse
nature of emotions and the fact that they are associated with some deviation from the
normal energy level and Tsur comes back again and again to the idea that a major
assumption of cognitive poetics is that poetry exploits for aesthetic purposes cognitive
processes that were evolved for non-aesthetic purposes. This probably has to do with the
representative nature of initial forms of poetry and literature in general, or simply with
imitation.
Next he takes into account the process of categorization, which may be rapid and
delayed, depending on how the verbal labels underload or overload the items in one’s
cognitive system; delayed categorization may involve a period of unpleasant uncertainty
about what is going on in a specific passage of a poem, while rapid categorization may
involve a loss of important information. This way, different categorization strategies may
generate, at both ends, different categorization strategies. These two types of
categorization are related to the way poetic metaphors are understood, to the implied
critic’s decision style and the above-mentioned altered states of consciousness. On the other
hand, readers and critics may differ from one another in their tolerance of delayed
categorization, of various types of metaphors or various aesthetic categories (like the
grotesque, for instance).
The main difference we see here is between the stable, well-organized categories (in
expository or scientific discourse) that convey straightforward loads of information on
one’s cognitive system, without any sensory (or very little), emotional information; and
poetic categorization (mostly delayed, but also sometimes rapid)that allows an overload of
sensory information which results (in the reader) in altered states of consciousness
(generated, most likely, by similar ones in the author). Tsur is highly quotable here when
he notes that this is
 “an element of suspension of boundaries between self and not-self, of immersion in a
thing-free and gestalt-free quality. Altered states of consciousness are states in which one is
exposed for extended periods of time to pre-categorial, or low-categorized information of
varying sorts. These would include a wide range of states in which the actively organizing
mind is not in full control, ranging from hypnagogic states (when one is half-awake, half-
asleep), through hypnotic state, to varieties of religious experience, most notably mystic
and ecstatic experiences. In the creative process, moments of ‘inspiration’ or of ‘insight’
too may involve such altered states of consciousness, though less readily recognized as such.
“ “Aspects of Cognitive Poetics,” p.11)

The hypnagogic state, or the hypnotic, or the other types of non-rational (“the mind is
not in full control”) states may refer primarily to those of the author while creating, though
they might very well characterize some moments of reading, so that what the effort here
seems to be is that of distinguishing between intention (“the intentional fallacy”) and
consciousness: nonconceptual experiences can be conveyed—one way or another—by the
use of language, which (back at the beginning of this section) is conceptual in nature.
Quoting psychologist Robert Orenstein, Tsur re-emphasizes that logical and rational
consciousness is related to the left hemisphere of the brain, while meditative consciousness
is related to the right hemisphere; the information is processed sequentially in the left
hemisphere (of the language) and it comes out as compact and logical, while the right
hemisphere processes information simultaneously and its output is experienced as diffuse,
integrating input from many senses (orientation—i.e. deixis in literature--, emotions, mystic
experiences). And thus what we have in poetry is the “transfer of a significant part of
language processing from the left to the right hemisphere, thus rendering the related
precepts more diffuse,” (p.12)
What cognitive poetics in this view seems to be doing is what (cognitive) criticism has
been doing for the past 2500 years, i.e. attempt to find a better, clearer, more satisfactory
way of saying or pointing out or suggesting what poetry has always said or communicated;
and Reuven Tsur manages to be unambiguous, as, obviously, an ambiguity (of criticism) on
top of another ambiguity (that of poetry) can only result in more ambiguity, so the possible
cognitive stance gets to be even more distanced from the quality of meaning a poem
contains:

“I claim that the right hemisphere’s output is ‘ineffable’ not because no semantic
features are involved, but because those features are diffuse and simultaneous. It is not the
information that is not unparaphrasable, but its integration and diffuseness. Diffuseness
and integration are not semantic information added, but the structure of information as it
appears in consciousness. Whereas semantic information can be paraphrased, the
impression that arises from its structure can only be described.” (p.13)

Therefore, paraphrase and description, which, together, may easily be called

interpretation; all in all, what Tsur seems to be investigating in the kind of language
(grammatical structures, elliptic sentences, deixis—the generation of a coherent scene or
‘world’—and orientation, prosodic structures, point of view and irony…) used in poetry to
produce certain effects and what other devices are used (figurative language in general and
metaphor in particular, distance, self and ego, perception of space an time…) to add up
these effects to an overall one; as we can see, there still are many gaps to be filled and many
questions to be asked.

8. 3. Stylistics and Rhetoric Revisited

Peter Stockwell’s “Cognitive Poetics and Literary Theory” argues that “cognitive poetics is
best seen as the latest development in the progressive evolution of stylistics,” while “the
endpoint of the process represents the return of rhetoric to the centre of literary
scholarship”(p.1). So our concepts are cognitive poetics, literary theory, literary
scholarship, stylistics and rhetoric, with cognitive poetics providing a descriptive account of
“how readers construct propositional content from literary reading;” and thus, once again,
reception theory and reader response. Like tsur, whom he often quotes, Stockwell adds
aesthetic analysis and emotional involvement, plus capturing “the interaction of
meaningfulness and felt experience in literary reading.” (p.1)
In stylistics itself, the general trend seems to have been away from formalism (Leo
Spitzer, Rene Wellek and Austin Warren, Stephen Ullman, later Leech, Short, and
Widdowson in Britain) towards a more contextualized stylistics, and thus “reconnecting
more fully with the older and longstanding rhetorical tradition;” (p.2) other developments
have been in text linguistics and pragmatics, discourse analysis and sociolinguistics,
computational and corpus linguistics; to all these, cognitive poetics also adds a
psychological and sociocultural dimension. The picture is almost complete with renewed
interest in conceptual metaphor, figure and ground, schema- and world-theories:

“Conceptual metaphor theory suggested new ways of examining creative language in

poetry and ways of understanding extended metaphors and thematics in longer fiction. The
work on figure and ground had obvious implications for understanding literary
foregrounding, significance, deviance and value. Schema theory and various theories of
world-building offered ways on which fictional worlds and performed poetic personas
could be better understood. Schema theory, possible worlds theory and text world theory
all suggested various ways to explain the fact that interpretive communities could share
roughly consensual readings at the same time as individual readers could hold varying
interpretations.”(p3)

Drawing as it does on cognitive science, cognitive poetics relies on the same principles as
its source discipline: the concept that meaning is embodied (mind and body are
continuous); the notion that categorization (see Tsur before) is a feature of prototype
effects, while categories are provisional; finally, the idea that language and its
manifestations in reading and interpretation is a natural and universal trait in humans.
Peter Stockwell’s own Cognitive Poetics (2002) describes first the micrological dimensions
of cognitive poetics (figure and ground, prototypes, deixis, cognitive grammar) and then
the macrological dimensions (schema poetics, possible worlds, mental spaces, metaphor
and parable, text world theory and models of global comprehension; all of these are used to
explore such issues as literariness (with fewer and fewer adepts lately), defamiliarization,
intertextuality, deviance, canonization, characterization, perspective, fictionality and so on
and so on.
 Tsur’s problem continues to attract attention, namely that lesser advances have been
recorded in accounting for aesthetic effects, side by side with the role of feelings, and
emotions, while the main focus remained on matters of meaningfulness; text world theory
(Stockwell, Werth and Gavins) seems to be a solution here, in that it explores the ways in
which a certain word or universe is enriched and experienced emotionally in the process of
reading:

“My point here is that world-based models go beyond a simple propositional account
and start to draw in considerations of felt experience, empathy, identification, atmosphere,
and impact. These are all dimensions that are a crucial part of the literary reading
experience, but they have not really been systematically addressed until recent and
forthcoming cognitive poetic work.”(Stockwell, p.6)

Not accidentally, some fruitful work has recently been done here in the analysis of
drama: the complexities of the discourse world of the theatre, audience, stage and actors
and the interaction of these elements with counterparts in the constructed text worlds
generated in the course of a dramatic performance (Dan McIntyre, E. Lahey, T.
Cruickshank…) This may be explained by the fact that dramatic texts may be pointing to a
continuity between literary and non-literary settings of language use, so that a new
principle of cognitive poetics emerges: there is no such thing as an exclusive literary
language, i.e. there is no such thing as literariness; both everyday and natural language
have a prominent creative dimension, so that no clear disjunction is accepted between
poetic and ‘non-poetic’ language; and thus a certain principle in certain types of stylistics
ends here. One author (Derek Attridge, 2002) suggests the use of singularity as the sense a
reader gets that the literary experience is not quite like anything else; it is not a feature, but
an event that takes place in reception.
Like Tsur again, Stockwell seems to be interested in the application of cognitive
frameworks in the understanding of literary effects and, implicitly, of aesthetic value. His
distinction is between professional readers and unprofessional ones, or readers who read
for reading’s sake, not for some ulterior purpose: Literary study in universities bears little
resemblance to the sorts of things non-professional readers (i.e. the huge majority of
readers//we ourselves have started questioning the existence of this ‘huge majority’) do in
literary reading.”(p.10) What he has in mind is such features that are barely discussed in
university lectures and seminars: atmosphere, tone, identification, excitement, involvement,
resistance, disgust…, i.e. the motivating factors for literary reading. Readerly involvement,
the sense of transformation and self-implication received quite a lot of attention in recent
years (Richard Gerrig, D. S. Miall, D. Kuiken…)
Stockwell’s final question is also hesitant: “If cognitive poetics can account for any
reading, then to what extent is it a theory at all?” Well, recourse to prototype again, since
cognitive analyses can identify prototypical readings produced either by individuals or by
communities; some readings may be widely shared and conventional others are
idiosyncratic and eccentric, i.e. they diverge from the norm, in which case we could speak
of a stylistics of reading or reader response. Nothing seems to be very new here, whence the
unsurprising conclusion:
 “A theory of literary reading (cognitive poetics) is merely a specific part of a general
theory of language, and this general theory is grounded in empirical evidence. This is not to
say that cognitive poetics in itself is a scientific theory in a straightforward sense, but it
does assert that there is a scientific basis for the tools which cognitive poetics provides for
explorations in literary reading.” (p.11)

Reuven Tsur shares in Stockwell’s skepticism, but he takes one step further and denies
Stockwell’s contribution itself, especially in what regards the latter’s work in deictic
categories; what Stockwell seems to be doing is classify, label and illustrate these
categories, when the task of cognitive poetics is to shift attention from labeling and
classifying them to accounting for their effect; in other words, Stockwell is too much
preoccupied with meaning (in a discussion of Shelley’s “Ozymandias”) and too little with
feeling; according to Tsur, in Stockwell/s practice “cognitive analysis… sometimes consists
in rechristening well-worn terminology into new, ‘cognitive’ terms.” And more:

“Everything that is language or literature goes through the cognitive system of authors,
readers, and critics. However, a discussion becomes cognitive not when it resorts to a
certain terminology, but when certain problems are addressed which cannot be properly
handled without appealing to some cognitive process or mechanism.” (Tsur above, p.18)

8. 4. Cognitive Science—More or Less

If interpretation is the aim of reading, inferencing is central to the process of reading, so

David S. Miall (very often with Don Kuiken) dedicates much of his work to inferencing.
Since all writers can mean more that they say, inferencing is highly important in discourse
processing, and he quotes Arthur C. Graesser, a psychologist who, in several articles,
considers the categories of knowledge-based inferences that map onto the representation of
the narrative, for instance, in working memory: referential (like the anaphoric “he,” “she,”
”it”) role assignment for each verbal category (time, space, object, agent, patient…), causal
relationships (linking one proposition to what went on before), character motivation,
theme, characters’ emotion, consequence, author intent, reader emotion…
Such a theory of inference is not only usable in discourse processing, but also I n a
number of other poetic domains like understanding the minds of characters or in
metaphoric mappings (Lakoff and Johnson), in deixis theory (keeping track of space and
time, the characters’ perspectives and relationships among them), the role of time itself in
narrative (not only story time and discourse time, but also the time of the reader, the time
of the narrator, the time of the plot, the time of actions at discourse level, the time of events
at story level, the time of characters, plus variation in time of the narrative discourse,
including the well-known scene/summary distinction), foregrounding and defamiliarizing,
character understanding (spaces in which they are embedded, relative position and
importance in the story, literary characters as mental models, psychological traits, their
aims and emotions, etc.)… The problem of inferencing in this author’s view seems to be
typical of the field of cognitive poetics as a whole.
Very much like Tsur, Miall avoids emphasis on interpretation, and points out the role of
feeling in literary response, though, he thinks, this remains a largely uncharted area. For
cognitive poetics, the question is not “What is this poem/drama/novel saying?”, which will
result in multiple interpretations, but rather “what were your feelings and emotions while
reading it?”; feeling situates readers in relation to complex modes of experience, memory,
and social understanding, just as literature in general can change readers’ modes of feeling
an modify them, side by side with the reader’s self-concept. Consequently (and again),
empirical research on reading must be seen as the centre of cognitive poetics.
Miall and Kuiken propose that feelings during literary reading be characterized at four
levels: suspense and amusement as reactions to an already interpreted narrative; feelings
that derive from perceived affinity with an author, narrator, or narrative figure (“I like
Dostoyevski, or I like Hamlet, and that’s it…”); feelings of appreciation, which are, in fact,
aesthetic reactions; and the fourth level, the most complex one, which involves the
modifying power of feelings (see above) that appear to be triggered by the narrative and
formal components of literary texts (phonetic iconicity among them, i.e. the sound patterns
of the text, especially in poetry or poetic prose). So, during reading, these feelings interact,
sometimes in the form of metaphors of personal identification, to modify the reader and his
self-understanding. There are, of course, typologies of feeling responses, but the main point
is that of understanding the role or roles that feeling performs during reading.
Summing up their view on the contributions of feeling to literary reading, Miall and
Kuiken (“A Feeling for Fiction…,” 2001) re-emphasize these four domains:

“(1) evaluative feelings toward the text as a whole, such as the overall enjoyment,
pleasure, or satisfaction of reading a short story; (2)narrative feelings toward specific
aspects of the fictional event sequence, such as empathy with a character or resonance with
the mood of a setting; (3) aesthetic feelings in response to the formal (generic, narrative, or
stylistic) components of a text, such as being struck by an apt metaphor; and (4) self-
modifying feelings that restructure the reader’s understanding of the textual narrative and,
simultaneously, the reader’s sense of self. While there is no sharp demarcation between
these four domains in readers’ experience—a given moment may contain elements of more
that one feeling process—we propose that each feeling domain depends upon
characteristically different structures and processes.” (p.3)

They next investigate some properties of modifying feelings, the generative power of
feelings, and the catharctic relationship, all of these on the basis of empirical evidence from
one of two stories. A special section is dedicated to anticipation and feeling, on the premise
that the prefrontal cortex is responsible for anticipation (see the numerous studies in the
past decades on beginnings): other responses are likely to be mediated by the right
hemisphere (see above), such as prosodic aspects of foregrounding, figurative language,
and narrative structure.
We may end up our not very convincing tour among representatives of cognitive poetics
by briefly referring to Raymond W. Gibbs”s The Poetics of Mind (1994) whose main
assumption (very much like that of Mark Turner about narrative) is that everyday
language is widely and ineradicably metaphoric; so, not only is there nothing like
literariness (or metaphoricity, for that matter), but we understand absolutely all linguistic
constructions in terms of what might be called, rather technically, figural projections of
image schemas. Everyday mind, continues Gibbs, is fundamentally shaped by various
poetic and figurative processes which, incidentally, develop very early in children (a special
chapter on “The Poetic Minds of Children”). As a result,
 “Cognitive science cannot approach adequate explanations of human mind and
behavior until it comes to terms with the fundamental poetic character of everyday
thought.”(p.454)

Figurative imagination is a part of human cognitive processes in general, so cognitive

poetics is almost equivalent to cognitive science, something that is meant to make it either
simpler, or infinitely more difficult to characterize.

8. 5. Overview

The main focus of cognitive poetics—in order to justify its cognitive dimension—is on how
readers process the language of text, so that psychology (processing), linguistics (language),
and text interpretation are very much part of its investigative purposes. We have seen, in
more that a couple of instances, that what it does more than other approaches to literature
is to explore the emotional aspects of information processing in reading.
A summing up of the principles underlying cognitive poetics is difficult at this point,
since there seems to have been no systematic approach to all the elements of poetics (author
and implied author, reader and implied reader, the graphitic, semantic, syntactic, and
figurative levels, narrator and narrate, point of view, perspective and focus, characters,
situations and events, symbolism, allegory and parable…) in order to see how each of them
is part of a cognitive project. A list can however be proposed, and it contains the oldest
principle of the embodied mind (Mark Johnson’s The Body in the Mind, Varela,
Thompson and Rosch’s The Embodied Mind, George Lakoff, E. Sweetser, etc.); form and
iconicity (mainly in poetry); the cognitive consciousness (conceptualization, intuition,
feeling, and emotion); metaphorical thought; creativity (creating emergent structures by
the process called blending—Fauconnier and Turner, The Way We Think); distributed
cognition; the role of audiences (interpretive communities)…

DRAGOS AVADANEI -COGNITIVE SCIENCE AND THE HUMANITIES, Ed.

„Universitas XXI”, Iaşi, 2010