Content uploaded by Anezka Kuzmicova
Author content
All content in this area was uploaded by Anezka Kuzmicova on Sep 22, 2015
Content may be subject to copyright.
Mimesis: Metaphysics, Cognition, Pragmatics
ed. G. Currie, P. Koťátko, M. Pokorný
273–315. London: College Publications, 2012.
FIDELITY WITHOUT MIMESIS:
MENTAL IMAGERY FROM VISUAL DESCRIPTION
Anežka Kuzmičová
ABSTRACT: In this paper, I oppose the common assumption that visual
descriptions in prose fiction are imageable by virtue of perceptual
mimesis. Based on introspection as well as convergent support from
cognitive science and other disciplines, I argue that visual description
(and the mental imagery it elicits), unlike narrative (and the mental
imagery it elicits), often stands in no positive relation to perceptual
mimesis because it lacks a structural counterpart in perceptual
experience. I present an alternative way of defining the kind of mental
imagery elicited by visual descriptions, and propose a number of text
variables underlying the imageability or non-imageability of any such
description.
What does one really gain from the critically acclaimed visual
descriptions in the novels of Dickens, Franzen, McEwan and
hundreds of others? What do they add to mimesis?
When we read prose fiction, many kinds of mimesis are at
stake. There is mimesis of conceptual thought generation,
mimesis of emotions portrayed and aroused, mimesis of
historical or scientific fact, mimesis of speech and so forth. One
kind of mimesis particularly relevant to the non-scholarly
reader, and particularly neglected by literary scholars in spite of
the current cognitive boom, consists in emulated experience
(e.g., Halliwell 2002: 22) of the world as apprehended pre-
verbally, by the sensorimotor apparatus alone. For instance, if a
narrative rendition of bright midday sunshine in high summer
elicits the near-experience of needing to squint, then the
Anežka Kuzmičová
274
passage in question is likely mimetic in the sense relevant for
this paper. For lack of a better word, I will further refer to this
kind of mimesis as perceptual mimesis (see also Scarry 1999: 6ff).
It should be pointed out, however, that I assume perceptual
mimesis to involve the entire sensorimotor array, including the
proprioceptive and kinaesthetic modalities (e.g. the senses of
limb and organ position, velocity, effort, acceleration and so
forth) that are less frequently associated with perception
proper.
It is generally assumed that the reader’s mental imagery is
a prime vehicle of perceptual mimesis. Insofar as a piece of
fiction succeeds in eliciting sensorimotor (especially visual)
images of its content, it is regarded as perceptually mimetic. In
the common parlance of book reviewers, essayists and literary
scholars, a particularly strong mimetic effect of the perceptual
kind is usually attributed to visual description (Wolf 2004: 339;
Nünning 2007: 113). To my knowledge, nobody has put this
near-automatic association between perceptual mimesis, mental
imagery, and visual description to closer scrutiny. Several
authors (Esrock 1994: 38; Scarry 1999: 55; Grünbaum 2007: 311)
have briefly countered the widespread assumption that there is
a straightforwardly direct relationship between the amount of
visual detail provided through description on the one hand and
the imageability of a text on the other. A few attempts have
been made to account for the mechanics of visual (Scarry 1999;
Burke 2011: 56–85) or multimodal (Kuzmičová, forthcoming)
imagery elicited by fiction at large and by narrative in
particular, but there is no systematic account of the imagery
elicited by visual description (when elicited at all). This paper
aims at filling the gap.
In section 1, I will briefly specify what I mean by visual
description and present a further rationale for describing the
Fidelity Without Mimesis
275
visual imagery it elicits. In section 2, I will proceed to the main
body of my argument and make the following point: Unlike
instances of narrative proper, visual descriptions and the mental
imagery they elicit are not perceptually mimetic, because they lack
an experiential correlate in the world as apprehended pre-
verbally. In this stage of the argument, special emphasis will be
put on the pre-requisite of experientiality and, to a somewhat
lesser degree, on the closely related pre-requisite of suspending
the verbal. In section 3, I will argue that even though these specific
points of contrast preclude perceptual mimesis, they allow for a
fruitful analogy between images from visual description and
another kind of visual mental imagery, namely images from
voluntary visualization (e.g., one’s purposeful image of what a
particular bike model looks like). This analogy will be based on
the following: images from visual descriptions, just like images
from voluntary visualization (and in contrast to those images that
are perceptually mimetic), are always expected, feeble, and
essentially finite. Finally, in section 4, I will further elaborate on
the analogy in order to infer a tentative set of rules of imageability
generally applicable to visual description. The proposed rules will
be supported by introspective analysis, aided by extant analyses
of voluntary visual imagery and by research on reading and
language processing at large. References to other cognitive-
scientific research, as well as to literary scholarship on the general
topic of description, will be made throughout the paper when
appropriate.
Reduced to the most basic questions and answers, the
main argument of the paper can be summarized as follows: Is
imagery from visual description perceptually mimetic? (No.) If
it has no correlate in perceptual experience, what other sort of
experience does it resemble, if any? (The experience of
voluntary visual imagery.) What makes visual descriptions
Anežka Kuzmičová
276
difficult to image in the first place, and what makes the
imageable ones imageable? (Visual descriptions in general tend
to run athwart the experiential makeup of visual mental
imagery, exceeding the limits of what can be accommodated in
a visual mental image. Visual descriptions only become
imageable when they operate within these limits.)
1/ Why visual description?
One could go on forever trying to formulate a comprehensive
definition of visual description. For the sake of brevity, I will
instead refer to a prototypical example:
An oval splayed out with whale-bone, [the cap] started off with three
pompons; these were followed by lozenges of velvet and rabbit’s fur
alternately, separated by a red band, and after that came a kind of bag
ending in a polygon of cardboard with intricate braiding on it; and
from this there hung down like a tassel, at the end of a long, too
slender cord, a little sheaf of gold threads. It was a new cap, with a
shiny peak. (Flaubert 1995: 16)
This description of Charles Bovary’s cap is by far the most
frequently quoted one among literary theorists of description
(see Bal 1982). Here are a few suggestions as to why this may be
so: Firstly, the passage ascribes “properties to entities within a
mental model of the world” (Herman 2009: 90). Secondly, the
entities and their properties are represented “in stasis, in
simultaneous relation, and (they) are organized by spatial
markers like adverbs of place.1 Verbs in the present, past, or past-
progressive tenses depict states” (Mosher 1991: 442). Thirdly,
references to the properties are post-posited with regard to the
central entity, which thus constitutes the “global introductory
1 Or, in this particular example, by temporal adverbs acting as spatial
markers.
Fidelity Without Mimesis
277
theme” of the description (Hamon 1982: 159; emphasis mine).
Having recognized these generic features of literary description,
let us now turn to the features that in my view are prototypical
for visual descriptions in particular.
Firstly, the passage isolates an inanimate object rather
than a person, animal, landscape or other kind of complex
spatial configuration. Inanimate objects tend to be comparably
insignificant in fiction insofar as they are the least likely to have
direct impact on the story (see also Barthes 1989). An inanimate
object cannot be readily imaged (like landscapes or other spatial
configurations) or identified with (like people or animals)
inwardly, by projecting one’s body inside it, and with it one’s
mind. An inanimate object is as close as one gets to objectivity,
and therefore also to a description that is purely sensory, visual.
Secondly, the inanimate object described in the passage is a
manufactured rather than a natural one, and like most
manufactured objects described in modern prose, it is an object
of daily use. Unlike natural objects, manufactured objects are
fully dependent for their identity on how they happen to be
instantaneously used (Atran 1990: 63). Thus lacking an objective
essence, manufactured objects are little more than what they
appear to be in a given situation. This makes them the perfect
content for visual descriptions, the descriptions of appearance.
All examples in this paper will consist in visual descriptions
bearing significant family resemblance to the above prototype.
Yet I believe that part of what I have to say about mental
imagery from visual description may apply to descriptions
quite remote in kind, visual or non-visual.
However, description is not only a type of text, but also,
by virtue of its intuitive noticeability, an autonomous mode of
text processing. That is to say, there is more to descriptions than
their typical features encoded in text. There is also something it
Anežka Kuzmičová
278
is typically like to be reading a description. This experience in
turn, albeit subject to many variables in its final quality, may be
correlated with specific cognitive processes prior to
consciousness. Description can only exist against the background
of other types of text and processing. In most cases, its other is
narrative, the dominant text-type of prose fiction. As far as prose
fiction is concerned, it is by contrast to narrative that description
is usually defined, and rightly so; it is the contrast to narrative
that makes it pre-reflectively noticeable in the first place.
Description processing entails, first and foremost, a marked pause
in a chain of events. More specifically, once narrative processing
has given way to visual description processing, the reader
temporarily loses track of, and any connection whatsoever with,
the preceding story. An eclipse of awareness takes place as it
were; the reader focuses on one type of content only, the basic
content of visual description: “that something is there and like
that” (Wolf 2007: 34). In real time, this could last a fraction of a
second or several minutes.
What visual description processing entails apart from a
clean-cut contrast to narrative temporality, and what is meant
by narrative, will be explored throughout the rest of this paper.
Importantly, it should be noted at this point that I do not
assume my notion of visual description processing to cover all
possible visual description experiences. Depending on context
and the instantaneous focus of the reader, visual descriptions as
a text-type can be experienced in a number of different ways.
What I rather assume is a continuum of possible description
experiences where my notion of visual description processing
constitutes one of two extremes. The opposite extreme consists
in description experiences indistinguishable, in terms of mental
imagery and perceptual mimesis, from experiences of narrative.
As for my examples of visual description as text-type, certain
Fidelity Without Mimesis
279
types of visual descriptions, such as descriptions of human
characters, landscapes or other spatial settings, are disregarded
exactly inasmuch they seem more likely to prompt experiences
of a less distinctly descriptive, i.e. more narrative, kind.
For any account of perceptual mimesis of the visual kind,
descriptions in general and visual descriptions in particular
would seem the natural place to begin. While narrative proper
may be equally, or in fact more, efficient in prompting visual
mental images, visual descriptions alone (when imageable at
all) secure the highest possible fidelity of mental image with
respect to the text. For instance, upon reading about a “broom”
with no further description, my mind may image whatever it
pleases. Most often, it will image by default the kind of broom I
am most familiar with from my firsthand experience of the
world. Most often, this will be unproblematic. But what if the
broom, or the story as a whole, turns out to be set in a foreign or
otherwise distant context? In such a case my mental image may
be proved incorrect anytime by a subsequent passage
suggesting that the broom is to be ridden on by a medieval-
style witch, or that it sweeps aboard a spaceship in zero gravity.
Although my initial image is then by no means disqualified as
mental experience, it fails to pertain to the story-world in
question. On the other hand, should my “broom” be described
when first mentioned, its visual description may perhaps
delimit my freedom of imaging but it prevents me from
conjuring incorrect images.
Although some of the mechanics of readers’ visual mental
imagery have previously been accounted for by literary
scholars (Scarry 1999; Burke 2011: 56–85), none of the accounts
has exposed or even acknowledged visual description’s unique
potential to stipulate bottom-up rather than top-down
processing. Drawing implicitly on a romantic notion of
Anežka Kuzmičová
280
imagination as essentially a free activity, these scholars have
treated imagery without properly considering its prosaic debt
to the specific wording of a text. Michael Burke, for instance,
suggests that readers tend to furnish fictional interiors with
visual images based on their childhood homes and that they
often do so regardless of textual counter-evidence. Such
“whimsical” top-down imaging may possibly be considered
truly experiential in ways largely outreaching the domain of
vision (e.g., in terms of its affective impact on the reader), but
from the viewpoint of content fidelity, it could just as well be
regarded as mere mind-wandering. To restate one of my
opening formulations with a little more precision, a piece of
fiction is perceptually mimetic insofar as it triggers mental
images of the world as we pre-verbally apprehend it. But the
images must also be images of certain fidelity with respect to
the text. Otherwise there would be no way of determining that
they really arose as an effect of a specific passage in a specific
piece of fiction, rather than as an effect of fiction reading in
general, or language use in general, or for no particular reason
at all. This is why visual description would appear to be highly
relevant to the study of perceptual mimesis.
An intuitive grasp of this unique ability to specify the
visual is probably what makes the common association of
perceptual mimesis with visual description so appealing. When
checked against random intuitions about actual practices of
reading, however, much of its appeal vanishes. Indeed, visual
description might make us image far less frequently and far less
vividly than suggested by the rhetoric of book reviews and book
promotion materials. For instance, there must be a reason why
non-scholarly readers, so notoriously keen on vicarious
experiencing, show a tendency to skim, or even skip (Allington
2011), particularly lengthy descriptions. Furthermore, when I ask
Fidelity Without Mimesis
281
fellow literary scholars for book recommendations featuring
vivid visual descriptions of, say, manufactured objects, they
invariably cite passages in which manufactured objects are
simply mentioned without being described. This may well arise
from the fact that visual descriptions of manufactured objects
occur less frequently than simple mentions do, but only to a
certain point. I have established elsewhere (Kuzmičová 2012)
that simple mentions often conjure imagery more reliably and
more vividly than visual descriptions. My objective then was to
account for the fundamental processes underlying the most
multimodally saturated kind of perceptual mimesis, the one
resulting in the reader’s instantaneous sense of presence in the
three-dimensional world of a story. Now is the time to explain
wherein the main difference between descriptions and simple
mentions lies, and to finally analyze the mental images
prompted by visual description in their own right.
I now turn to the main difference between imagery from
visual descriptions and imagery from simple mentions, which, I
would like to argue, is not one of degree, but one of kind. That
is to say, whenever my mind conjures up a visual image of an
object based on my processing of its visual description, the
resulting experience does not amount to some weak variety of
my presence in the story-world, or a weak variety of perceptual
mimesis for that matter. In fact, the resulting experience is not
at all perceptually mimetic. Images prompted by visual
descriptions are essentially different from other readerly visual
imagery because they are generated differently. While simple
mentions and other narrative instances of fiction generate images by
virtue of their experientiality, visual descriptions can only generate
images by virtue of their imageability.2 The former part of this
2 When aesthetician Elaine Scarry (1999) subsumes all fiction-induced
imagery under the (further undefined) notion of perceptual mimesis, she
Anežka Kuzmičová
282
assertion will be elaborated in the next section. The latter part
will be elaborated subsequently.
2/ Why not perceptual mimesis?
Why is experientiality proper not at work in the processing of
visual descriptions? A definition of perceptual experience is
needed here. Underlying my assertion is an understanding of
perception as preconditioned by bodily interaction. Continuous
interaction with our immediate environment, be it overt (action,
i.e. bodily movement) or covert (psychophysiological processes
related to pre-conscious or conscious action simulation), has
lately been identified at various levels of inquiry to be the basis
of our sensorimotor apprehension of the world. A textbook
example of the inextricable link between interaction and
perceptual experience are the clinical cases of so called
experiential blindness. Congenitally blind patients whose
vision has been restored by surgery tend to take physical
objects for blurs in their visual field as long as they remain
unable to couple their visual sensations with relevant
sensorimotor patterns of interaction. (Noë 2006: 5ff) At a pre-
experiential level, there are indications that visual attention for
objects involves neurophysiological processes inherent to action
preparation (Rizzolatti and Gallese 1988). There are countless
convergent sources like these, spanning vast areas and
methodologies of inquiry from isolated brain imaging studies
of visual and linguistic processing (see e.g. Martin 2007; Fischer
and Zwaan 2008) to comprehensive enactivist phenomenologies
of the self (Gallagher and Zahavi 2007). The growing body of
interdisciplinary research pointing toward a centrality of
fails to isolate description as a distinctive text-type and mode of processing.
Also, her examples of imageable prose are predominantly narrative rather
than (visual) descriptive.
Fidelity Without Mimesis
283
interaction in experience is commonly subsumed under the
aggregate label of embodied (or grounded or situated)
cognition. Given the strong definition of experience proposed
by the framework of embodied cognition, visual description
construed as stasis and as subsequent temporary detachment
from the object described (see also Grünbaum 2007) has no
purely experiential correlate in the actual world. For in visual
description, interaction has come to a temporary halt.
That is not the case with simple mentions. Simple
mentions of object names, unless we are dealing with a
catalogue rather than with narrative, tend to be part of
interaction insofar as they stand for grammatical subjects or
objects attached to non-copular verbs, e.g.:
In the kitchen closet I found a practically new broom[.] (Baker 1998:
20)
This is a matter of syntactic fact rather than necessity. In not-
very-elegant prose, the broom in this sentence could be
minutely visually qualified, e.g.:
In the kitchen closet I found a practically new bright red ridged plastic
broom.
The point is that simple mentions happen to adhere more
closely to their referents as pre-verbally experienced. Upon the
reading of the above sentences, the reader’s embodied mind has
no problem identifying an interaction to emulate (the firsthand
act of opening a door and finding a broom in the penumbra of a
closet), forming thus a multimodal sensorimotor image, an
instance of presence, a mediated experience proper of a world
out there.
Anežka Kuzmičová
284
One could argue that all descriptions, along with the
objects described, are likewise embedded in interactive situations
because they belong to larger narrative wholes that always
feature such situations. However, as suggested by the
introductory definition of visual description as a mode of
processing (section 1), the surrounding story is relegated to
outside the reader’s consciousness as soon as a visual description
is encountered and pre-reflectively identified as such. Visual
description means per definition an instantaneously experienced
lack of continuity with any narrative (and interactive)
embedding, and consequently with any emulated firsthand
experience.
One could further object that by token of the theories of
embodied cognition, even the seemingly most passive
observation of the world entails covert interaction, and that
visual descriptions are in this respect no different from the
firsthand experience of such observation. One could say, in
other words, that visual description emulates in the reader an
act of firsthand yet passive visual experience and that it is
experiential in the same way as narrative renditions of overt
interaction are, only less perspicuously. It may be particularly
tempting to say so with regard to descriptive passages that are
framed by explicit or very strongly implied references to acts of
perceiving. For instance, the context of the visual description of
Charles Bovary’s cap strongly suggests that the cap is in fact
being scrutinized by the boy’s contemptuous classmates. Yet
again, as long as the reader’s mind remains aware of such
framing, which seems particularly unlikely in a description of
such flamboyance, we are not dealing with visual description
processing proper. Should the same objection be raised so as to
encompass all visual descriptions regardless of framing, it must
be countered by the following clarification:
Fidelity Without Mimesis
285
In the instant of switching, for a second or a fraction of a
second, from narrative to the mode of visual description
processing, the linguistic nature of reading robustly emerges
toward the threshold of the reader’s consciousness. The reader
thus assumes, if barely reflectively, the cognitive stance of
someone who is being informed that a certain object has certain
visual properties. In line with this argument, several literary
theorists (Hamon 1981: 21; Cobley 1986: 397) have noted that
description in general entails an increased presence of the
narrator. Importantly, the reader has no way of simultaneously
maintaining the cognitive stance of someone who is
approaching an object of certain properties pre-verbally, the
stance of a perceiving experiencer. Instead, the properties are
taken in as foregrounded information in a framework of
communication.3 Whatever the actual syntax of the description
in writing, the mental propositional “syntax” (if there was such
a thing) of visual description processing would follow roughly
this pattern:
There is a broom. It is practically new and made of bright red ridged
plastic.
rather than:
There is a practically new bright red ridged plastic broom.
Unlike in the processing of (certain instances of) narrative
proper, it becomes impossible under such circumstances to
3 This part of my argument seems to dovetail with Michael
Riffaterre’s. Speaking of description in general, literary theorist Riffaterre
maintains that description’s “primary purpose is not to offer a
representation, but to dictate an interpretation” (Riffaterre 1981: 125).
Anežka Kuzmičová
286
bracket off this quasi-communicative dimension of visual
description and achieve as it were a full, pre-verbal sense of
presence. This results in the immediate noticeability of visual
description in the course of reading, its autonomy as a receptive
mode, as well as in the skimming or skipping of descriptions by
impatient readers, which goes hand in hand with another
typical attribute of visual description as text-type: its low
memorability. Unless the wording or subject of a visual
description is perceived as particularly striking, the reader is
often left with a sense of amnesia as soon as the description is
over, not to speak of one’s minuscule chances of retaining the
rough contents of a description beyond an immediately
subsequent stretch of text.
In cognitive psychology, a pronounced tradeoff has been
found to operate in the (English) lexicon and its processing
between imageability on the one hand and phonological and
orthographic uniqueness on the other (Westbury and
Moroschan 2009). Words with the lowest number of direct
phonological neighbors, i.e. words that differ the most in their
structure from the rest of the lexicon, happen to denote
referents with low or no imageability (e.g., “thought”).
Conversely, words with highly imageable referents are the least
conspicuous as to their structure (e.g., “broom”) and thereby
also the most easily confused with other words (“room”,
“boom”, “brook” etc.). Suggesting that marked “verbality”
somehow interferes with imagery, these findings can be taken
in support of the above assumption that a sense of being
verbally informed of perceptual facts is necessarily
discontinuous with a sense of direct perception. Overall, they
further disclose how treacherous visual description is by
nature. It is intuitively known by the reader to denote
something quite familiar and easy to image, yet in the end it is
Fidelity Without Mimesis
287
always found to consist in a unique, unfamiliar4 concatenation
of qualifiers.
Besides, unless the qualifiers in question situate the object
in uncommon context, e.g. by suggesting the kind of twig
broom used by witches and others in the Middle Ages, they
also happen to be more or less tangential to any sort of
interaction. In most visual descriptions they will be tangential
by necessity, because those properties that really pertain to
interaction are already encapsulated in the central object name
itself. Empirical studies have shown that in the cognitive
processing (and thereby also in the linguistic labeling) of
manufactured objects, each object category—broom, cup or
flower pot—is delimited exactly by the particular subset of its
properties that are immediately relevant to interaction (Rosch et
al. 1976). For instance, any container that can be used for
planting flowers, qualifies as a flower pot. Thus it would make
little sense to describe a flower pot by recounting the properties
of being hollow and closed at the bottom. Indeed, few
straightforwardly prosaic descriptions take such course.5 In
general, visual descriptions tend rather to exploit the countless
accidental properties, those having no direct relation to how
objects are essentially interacted with.
The above mentioned impossibility to bracket off the
linguistic medium lies also in the very nature of isolated objects
and their visual properties. In firsthand visual experience, as
long as objects are principally apprehended as objects of
interaction, visual properties are those that are self-evidently
4 An interesting exception being visual descriptions by way of
epitheton constans, which are always standardized within a corpus and thus
largely familiar.
5 The opposite cases generally signal an anti-representationalist
authorial agenda, resulting in estrangement.
Anežka Kuzmičová
288
given. This may be why visual description seems to occur
rarely in mundane oral narrative and conversation in general,6
and when they do, they may mostly be meant to foreclose
baseline misconception, rather than help the interlocutor walk
in another perceiver’s shoes. Unless I am describing an object
for its purely aesthetic qualities, e.g. a work of art or a piece of
clothing, the pragmatics of my infrequent spontaneous
descriptions tends to be other than that of prompting imagery:
I want the interlocutor to help me find my purse in the mess of
my office, or to pick up the right kind of baby food at the
grocery store. I want the interlocutor to know what these things
look like, not necessarily to see them with the mind’s eye.
Lastly, apart from being self-evident, the visual properties
of an object given in firsthand experience are in most cases all
simultaneous and one with the object itself. In comparison to
narrative renditions of interactive situations, this puts visual
description at a disadvantage vis-à-vis the inherent temporality
of language and turns it once more into a gross abstraction from
perceptual experience. Looking at the ancient broom in my
childhood home kitchen, I can certainly conceive of its visual
properties in a linear sequence, one by one. But I cannot do so
without recourse to inner speech, without hearing my mind
briskly articulate at least some of the sounds in “brown”,
“wooden”, “shabby”, “orthogonal”.
To sum up, the aim of this section was to isolate two
fundamental characteristics of visual description. The two
characteristics taken together disqualify visual description from
6 As discourse theorist William Labov (1972: 370) would put it, visual
properties of static objects are seldom reportable, i.e., they tend to lack the
quality of being inherently worth telling. Labov (1972: 389) also expressly
notes the rarity of qualifying syntactic structures in his own material, the
African American Vernacular.
Fidelity Without Mimesis
289
perceptual mimesis. Ultimately, the two characteristics taken
together bear a significant deal of responsibility for visual
description’s lower overall imageability when compared to
narrative. Firstly, I have shown that visual description, unlike
narrative, defies firsthand perceptual experience. Secondly, I
have shown that mental imagery from visual description, when
prompted at all, is significantly less pre-verbal in nature than
mental imagery from narrative. While the latter characteristic is
a matter of degree (at a certain level of awareness, the reader
always knows that he or she is dealing with a piece of verbal
fiction), the former is a question of imagery from visual
description either bluntly being or bluntly not being
experiential and perceptually mimetic. But to say that imagery
prompted by visual description in a very particular sense is not
experiential is not to say that it does not amount to experience.
What sort of experience it amounts to will be suggested in the
following section.
3/ What other sort of experience?
What do I mean by stating, at the end of section 1, that visual
descriptions can only generate imagery by virtue of their
imageability? In part I am referring to their dissociation from
firsthand perceptual experience. There is no experience in them;
there is only the experience of them. But I am also hinting at the
one sort of experience to which imagery from visual description
bears resemblance: the experience of (unseen) objects as
visualized in a voluntary mental act of imaging. Voluntary
mental imaging is the kind of imaging engaged in when one
tries to image what an object (e.g., a particular bike model)
Anežka Kuzmičová
290
looks like based on one’s memory or declarative knowledge, or
when one fantasizes about a perfect something.7
It is not my intention to suggest that imagery from visual
description is mimetic with regard to the above mental acts.
Once again, unless the above mental acts are explicitly rendered
in the text and retained as such in the reader’s focus, i.e., unless
the mode of processing inclines toward narrativity and fails to
be one of visual description proper, the reader’s experience is
clearly dissimilar from voluntary visualizing or indulging in
fantasies. For instance, the acts of voluntary imaging are
temporally open-ended, the imager possessing the ultimate
power to extend their duration endlessly. Imaging in visual
description processing, on the other hand, is always framed by
the reader’s assumption that strict temporal constraints have
been set beforehand. A visual descriptive stretch of text can
easily turn out to feel somewhat lengthy to a reader, but a
voluntary imager never continues imaging beyond what feels
right for the moment. Furthermore, while in the act of
voluntary imagery it is the imager alone who is the originator
of the experience and who thus largely exerts control over its
content, imagery from visual description arises upon external
instruction, with all the rigor and lack of control this entails.
Given these and other discrepancies, an important clarification
must be made at this point: I am going to consider similarities
between voluntary mental imagery and imagery from visual
7 Unlike philosopher Evan Thompson (2007), I assume that it is
possible, although not necessary, to image an object by an act of will without
simultaneously enacting, on the level of consciousness, a firsthand
perceptual experience of that object. Voluntary visual images involving
emulated perceptual experience (e.g., images of what it is like to be looking
at a particular object), perfectly common as they are, will be excluded from
the present discussion for lack of analogy to imagery from visual
description.
Fidelity Without Mimesis
291
description pertaining only to how the product (the image), not
the act of production (the imaging), is experienced. Separating
the two phases of experience conforms with established
phenomenological practice (Casey 2000: 38).
The following questions may arise: Why voluntary visual
images in the first place? Why not consider other sorts of visual
images, such as the far more frequent products of compulsive
imaging, the fleeting yet intense mental images that can take us
by surprise whenever we happen to think of, or talk or hear
about, something highly imageable? One could even wonder
whether these latter images really are not closer to images from
visual descriptions after all, given their uncontrolled character.
But they are not. They differ from voluntary visual images and
from images prompted by visual descriptions in several
respects. Firstly, they differ in that they can, and do, take us by
surprise. Voluntary visual images, on the other hand, are
always expected, and the same is true of images from visual
description. As soon as a visual description has been
encountered in a text and identified as such, the (modern)
reader automatically assumes that visual imagery will
somehow be addressed. This is not to say that one is never
surprised by the specific contents of an image prompted by a
visual description, just like one can sometimes be surprised by
the specific turns one’s voluntary imagery has taken. But one is
never surprised that an image has arisen.
Apart from always being expected, voluntary visual
images as well as images from visual description tend to be
experienced as markedly feeble (see also Scarry 1999: 4; Casey
2000: 3). Their feebleness distinguishes them further from
involuntary visual images. Surprise alone could be the reason
why images of the involuntary, fleeting kind appear as much
more saturated. However, the sheer possibility of surprise lies
Anežka Kuzmičová
292
at the heart of a yet deeper difference, one that comes down to
the question of perceptual mimesis.8 For instance, why am I
surprised by the compelling image of the handlebar of my bike
suddenly emerging in my consciousness as I let my mind
wander freely on a tired afternoon? Why do I experience the
image as strikingly vivid? I am surprised because initially I was
thinking of something else than my bike and the visual details
of its handlebar. Otherwise the image would have been
expected. And I experience the image as strikingly vivid
because this something else that I was initially thinking about
was in fact an instance of interactive perceptual experience:
dropping off my son at daycare this morning, then biking to the
station to catch the bus for the University. Consequently the
involuntary image of my bike has experiential qualities
comparable to the perceptually mimetic imagery prompted by
certain instances of narrative. These qualities are absent in any
visual image resulting from a voluntary attempt to visualize a
static (see also Jajdelska et al. 2010) isolated object. At the same
time, these qualities seem to be constantly in demand, visual
imagery being reflexively assessed by the standards of
perceptual experience. Hence the sense of enfeeblement
inherent to images from visual description.
The third and last feature to be recounted in this section is
finitude. This feature too is best conceptualized upon
comparison with involuntary visual images and images from
narrative. It too derives indirectly from the lack of perceptual
mimesis, in the following respect: bearing traces of perceptual
experience, involuntary visual images and images from
8 By contrast, Elaine Scarry (1999: 104) contends that all fiction-
induced imagery supersedes the feebleness of voluntary visual imaging. By
way of explanation, she emphasizes that imagery in reading is involuntary
inasmuch as it is constrained by external instruction.
Fidelity Without Mimesis
293
narrative are residually dynamic. This is not the case for
voluntary visual images or images from visual description,
which are static through and through. Mental images of all
kinds may be considered temporal insofar as they have a
certain if minimal duration, yet only the dynamic ones contain
a promise of something more than what is immediately
presented. For instance, when I as a reader have experienced a
visual image of a broom while emulating the experience of
retrieving it from a closet, my broom image tends to recur for a
little while before it fades away completely. Sometimes it
changes slightly between the various stages of recurrence and
then it is no longer, strictly speaking, the same image as before.
Nevertheless, an image of a broom does recur without a broom
being mentioned anymore. It echoes throughout the dynamic
extension of my covert enactment of the bodily movements
involved when stretching my arm, grasping the broomstick,
retrieving the broom. When such enactment is particularly
strong, perhaps outright noticeable in the muscles of my arm
and hand (see also Kuzmičová 2012), the image can keep
recurring for a considerable period of reading time.
Alas, images from visual description, similarly to
voluntary visual images of isolated objects, do not have the
same tendency to deliver promises of surplus visual experience,
or to promise anything in the first place. They must be cued
anew if they are to recur. In their static nature, they are destined
to yield to other, dynamic and interactive experiences at the
very next intersection with narrative. Here are a few examples
of how this can happen:
[The camera] was mounted on an altazimuth bracket above the back
door. Its casing was of brushed aluminum. It had a purplish gleam in
its eye. ¶ [1] Gary returned the bottle to the liquor cabinet, moved to
the sink, and ran water in a bucket. (Franzen 2001: 230)
Anežka Kuzmičová
294
By a pile of magazines was a coffee cup—tall, in thin white porcelain,
one of a set of six [2] bought by Patrice at Henri Bendel’s in New York.
[3] Aldous raised it to his lips. (McEwan 2011: 97)
In cases where the immediately subsequent narrative refers to a
direct interaction with the central object, the initial image may
be transformed into a perceptually mimetic one, continuing its
life in a new format. While segment [2] in the latter passage
bears but a vague resemblance to this scenario (suggesting a
hypothetically direct interaction that involves, but is not limited
to, the central object), segment [3] provides a more clear-cut
example of how mental images from visual description can live
a narrative afterlife. In all other cases where visual description
is interrupted by narrative, i.e. in the cases represented by
segment [1] above, images from visual description appear once
and then vanish without extension and the reader’s image
experience is readily informed by this. The reader thus
experiences that, apart from being expected and feeble, images
from visual description are essentially finite, in ways that
images from narrative are not.
Needed or not as my above observations may have been
in themselves, the ultimate aim of this paper is to valorize them
for more practical, predictive purposes. They are meant to help
determine what it might be that makes a visual description
elicit mental imagery, in spite of the lack of perceptual mimesis.
However, defined as expectedness, feebleness and finitude, the
principal experiential features of images from visual
description are still too broad to instruct a text-oriented
analysis. In the next section, while I revisit visual description as
text-type, a number of sub-features and further observations
will be grouped with the two of the above features that are
Fidelity Without Mimesis
295
directly relevant to image content, in the following order:
finitude (subsection 4.1), feebleness (subsection 4.2).
4/ More on the image and when it arises
In assessing the readerly experience of narrative, it is possible
and natural to explain imageability by reference to structural
analogies between text and firsthand perceptual experience
(Kuzmičová 2012), with a vast body of empirical perception
research at one’s disposal. In the absence of perceptual mimesis,
functional analogies (or functional discrepancies) can only be
charted between text on the one hand and visual mental image
on the other. Given the general elusiveness of mental imagery,
there is by contrast little empirical evidence to rely on, and
introspection becomes as indispensable as ever.
Based on introspection, the idea that fiction can be made
imageable by emulating the inherent characteristics of
voluntary imagery has previously been suggested by
aesthetician Elaine Scarry. In Scarry’s (1999) valuable account,
fiction becomes imageable by virtue of analogy when the
predicaments of voluntary imaging are made explicit (e.g.,
when a character is struggling to visualize a cherished face), or
alternatively, when objects of certain qualities (e.g., trans-
lucence, floral supposition) are represented. Even if Scarry did
take notice of the idiosyncratic mode of visual description,
there would still be significant differences between her
approach and mine. Most notably, Scarry singles out image-
able particulars such as the visual properties of being
translucent or flower-like. Meanwhile, my subsections 4.1 and
4.2 aim at distinguishing between markedly imageable and
non-imageable classes of visual properties, or parameters (e.g.,
color, shape), regardless of value (e.g., blue or yellow,
rectangular or circular). The respective presence or absence of
Anežka Kuzmičová
296
each and one of the selected parameters in an instance of visual
description will be set against the default parameters and
limitations of the visual mental image (voluntary or prompted
by visual description), and its effect on description imageability
will be predicted.
4.1/ Default parameters (finitude)
The experienced finitude of imagery from visual description is
closely related to its experienced feebleness. The former will
now serve as a background for positive characterization. The
latter will subsequently frame an account of what properties a
mental image from visual description does not, and in some
cases even cannot have.
The kind of image finitude discussed here may be further
conceptualized by comparison to a picture. Having no intention
to take sides with either the descriptivist or pictorialist camp of
the age-old imagery debate in cognitive neuroscience,9 my
suggestion is that mental images from visual description are
picture-like at least in two relevant respects: Firstly, because of
their static nature, they are experienced as two-dimensional
(see also Casey 2000: 92). It may be by virtue of this
resemblance that one of the most prominent types of so called
ekphrasis (i.e., the ancient rhetorical device of visual rendition)
was the verbal representation of (more or less) two-dimensional
visual artworks. As objects, such artworks are largely defined
by their complex pictorial surface, while they tend to be
uniform in overall contour shape. Interestingly, the mundane
objects visually described by modern fiction that are presently
9 Arguing about the pre-experiential cognitive format of mental
imagery, the descriptivists posit that mental imagery comes down to
propositional structures, while the pictorialists reject the possibility of such a
reduction. (See e.g. Thompson 2007 for a review.)
Fidelity Without Mimesis
297
in focus—brooms, cups, flower pots—depend more heavily on
contour shape for their identity than on surface. This fact is
reflected in the makeup of the correspondent mental image.
When I read about a flower pot or simply fancy to image one
out of the blue, I may or may not be able to tell afterwards how
porous the earthenware was, but I will always roughly know
the slope of its wall. The contour of the flower pot may be
incomplete (there may be no way of telling whether or not there
was a rim at the upper edge), but it will always be there and it
will always be perceived as final because the flower pot will be
given to me in two dimensions. By contrast, the contour shape
of a flower pot imaged in the processing of narrative will be
perceived as open to modification due to the possibility of
interaction, by means of which a virtual third dimension is
brought into existence.
Secondly, mental images from visual description tend to
be oriented in a way resembling of certain canonical types of
pictorial representation. If their contour is to adhere to the
object they represent, they cannot be, and obviously are not,
multiperspectival in the manner cubist paintings are. In their
feebleness, they can hardly be said to comply with the
standards of realist perspectival painting. They are imaged
under a perspective nevertheless, and most often a markedly
pictorial perspective at that, namely the one optimally revealing
the distinctive contour shape of the object in question. When
imaging a broom decoupled from perceptual experience, i.e.,
when processing its visual description, the broom as an entity is
given to me in the most perspicuous way: vertically, perhaps in
a slight angle to the orthogonal axis, its bristles facing the
bottom of my mental visual field (see Figure 1). This is how a
broom is normally depicted when immediate comprehension is
at stake, e.g., in pictograms or illustrated dictionaries.
Anežka Kuzmičová
298
Object orientation in involuntary images or in images
prompted by narrative, by contrast, is always situated. For
instance, should I image the same broom as part of a perceptual
experience, e.g. in emulating the act of sweeping, the
broomstick would become, in a rather compelling manner,
disproportionately short and thick in my mental image and the
overall contour of the broom would alter.10 Moreover, the
image as a whole would no longer symmetrically occupy the
center of my mental visual field, but rather gravitate toward its
lower right hand side (I am right-handed). Albeit compelling in
the mental, such a broom image would seem highly indefinite
and ambiguous as to its content if transposed into an actual
two-dimensional picture (see Figure 2). There are of course
instances of perceptual mimesis in which brooms appear under
the same perspective as depicted in pictograms or as imaged
during visual description processing. The difference is that in
mental imagery from visual description, brooms are rarely
oriented otherwise.
An approximate contour filled with a sketchy surface as
afforded by an initial orientation is all there is to visual mental
images of isolated objects such as brooms, cups or flower pots.
That is to say, they are all there is by default, at the very instant
a reader has pre-reflectively understood that an object
description is about to begin unfolding, but before any post-
posited visual attributes have been taken in. This particular
stage of imagery is what I call the default mental image. It arises,
for instance, with the underlined portions of the following
examples:
10 The perspective thus assumed would coincide with what cognitive
psychologists call ”canonical perspective”, i.e., a perspective by which
typical interaction is facilitated (Palmer, Rosch, and Chase 1981).
Fidelity Without Mimesis
299
[T]hree dirty mattresses, each rolled up in a blanket: which occupied
one corner of the room during the day, and formed a kind of slab, on
which were placed an old cracked basin, ewer, and soap-dish, of
common yellow earthenware, with a blue flower[.] (Dickens 1998: 551)
My Austrian sniper’s rifle with its blued octagon barrel and the lovely
dark walnut, cheek-fitted, schutzen stock, hung over the two beds.
(Hemingway 1962: 11)
The tablecloth was thick, smooth and blue. Heavy Indian cotton, a
thin turquoise line through blue checks. Small frayed holes here and
there. (Roberts 1993: 14)
Whatever parameters the reader brings in on top of contour
shape and orientation pertain less to the image as such (but see
below, section 4.2) than to the reader’s preconceived notion of
the object in question. If the reader lives in a world where most
brooms are brown, or if the reader assumes that most brooms
are brown in the particular world of the particular piece of
fiction, then it is sheer conceptual knowledge that makes the
reader tacitly believe that a fictional broom is brown.
Obviously, the particular contour and orientation imposed on
one’s default mental image of a broom are mediated by
conceptual knowledge as well. They happen to coincide with
what one’s culture knows as prototypical. The main difference
from other visual parameters, e.g. color, is that contour shape
and orientation alone are necessary for manufactured objects
such as brooms to appear as what they are. As far as mental
images of manufactured objects are considered, the other
parameters are accidental. A purple emerald will no longer be
an emerald, but a broom made of purple china will still appear
as a broom unless its practical function is considered, which
anyway never happens in visual description proper, where
appearance is the only thing at stake (see also section 2).
Anežka Kuzmičová
300
A strawberry turned upside down will lose nothing of its
essence, but a flower pot turned upside down will suddenly
appear somewhat less like a flower pot. This, along with a fair
deal of introspection, is what lies behind the above suggestion
that contour shape and orientation is all one really sees in,
rather than reads into, the default mental image.
But what does the particular status of the two closely
interconnected parameters of contour shape and orientation
imply for imageability in cases when contour shape or
orientation are explicitly mentioned in a visual description?
Paradoxically, nothing much. Their centrality to the definition
of each object category (and the relatively low variability of
shape within each object category) seems to make the two
parameters relatively useless, and perhaps even relatively little
used, in visual descriptions.
That a particular contour shape or orientation is
mentioned at all, usually implies that several different contour
shapes or orientations are afforded by the object category in
question. While less typical contour shapes (rectangular flower
pots) need not cause difficulty for imagery, less typical
orientations (chairs lying on their backrests) tend to be more
treacherous but mostly viable thanks to our ability to perform
mental rotation (Shepard and Metzler 1971). Importantly,
unless the object in the specific contour shape and orientation is
highly unexpected (spheric brooms, banana shaped coffee
cups), it is accommodated by the initial mental image without
resistance, but also without the reader taking particular notice.
It has been noted by Michael Burke (2011: 145), and partly
also by the proponents of the classical theory of literary
estrangement (Shklovsky 1990: 1–14), that mental contents
really become noticeable only when a mismatch takes place
between the reader’s top-down preconceptions on the one hand
Fidelity Without Mimesis
301
and bottom-up textual input on the other. Once we adopt the
same idea for visual mental imagery, the following conclusion
avails: if contour shape and orientation are the only two default
parameters in mental images of isolated objects, then explicit
references to these parameters should represent the only kind
of textual input that can induce a match or mismatch proper,
i.e. a match or mismatch concerning the image as such. Yet
paradoxically, match or mismatch seems to make little
difference to contour shape and orientation. Mentioned or
unmentioned, matching or mismatching, it is as if the two
parameters were rarely noticed in their own right exactly
because their presence in the mental image is inevitable
anyway. Compare for instance the following passages:
She carried the Quimper dish on her upturned hands. … A big dish,
roughly oblong in shape, with rounded shoulders. Its thickness and
heaviness were emphasized by the bold strokes of its painted
decoration, dark orange, dark pink, and navy blue. (Roberts 1993: 91)
Gary … took the last of the six signs that a Neverest representative
had sold to him. Considering the cost of a Neverest home-security
system, the signs were unbelievably shoddy. The placards, roughly
oblong in shape, were unevenly painted and attached by fragile
aluminum rivets to posts of rolled sheet metal[.] (Franzen 2001: 225)
Above, a contour shape qualifier (“roughly oblong in shape”)
was removed from the former passage and planted in the latter
passage. Neither one of the two mental images (of the dish, of
the home-security sign) lost or gained any of its initial power, in
spite of the fact that home-security signs are more likely than
dishes to be roughly oblong in shape, and in spite of “roughly
oblong” suggesting slightly different shapes for the two objects:
an oval one for the dish, a rectangular one for the home-security
sign. This is not to say that certain contour shapes or
Anežka Kuzmičová
302
(especially) orientations are not more likely to be imaged than
others. The point is simply that there is nothing about visual
descriptions referring to contour shape or orientation per se that
makes them either imageable or non-imageable. This is not the
case for those parameters for which any possible match or
mismatch pertains not to the level of the image, but to the level
of invested conceptual knowledge. Those are the non-default
parameters. Among them, I will argue, some truly have the
power to make a visual description imageable, while others are
for various reasons detrimental to description imageability.
Tapping firmly into the conceptual, the non-default parameters
overall seem more likely to become noticed, to capture one’s
attention in the course of reading. The ones listed in the
following subsection tend also to make a noticeable difference
for one’s mental imagery.
Figure 1
Broom contour prompted by
visual description
Figure 2
Broom contour prompted by
narrative
© Jim Shaw © Schünnin
Fidelity Without Mimesis
303
4.2/ Other parameters, limitations (feebleness)
The absence of each one of the below visual parameters adds to
the perceived feebleness of the default mental image from
visual description. By recounting these parameters, I will thus
continue recounting the many ways in which mental images
from visual descriptions, especially in the default, are
experienced as feeble. However, there are countless aspects to
what meets the eye in perception and I have no ambition to
offer a comprehensive overview of all conceivable visual
parameters. Rather, my intention is to propose a general way of
classifying visual parameters according to their imageability,
while identifying salient representatives of each category.
Throughout the proposal, new distinctions will need to be
drawn between the various levels of the notion of imageability
that are at play. A diagrammatic summary will finally be
provided in Figure 3, where further examples of each category
will also be proposed.
My ultimate aim is to pinpoint visual parameters that
have a pronounced impact, be it positive or negative, on the
imageability of a visual description of an object as encountered
in a piece of fiction. In this respect, the only thing I have been
able to establish thus far is that the parameters of contour shape
and orientation do not seem to have much impact. On the other
hand, the parameters of contour shape and orientation
obviously are imageable, even more so than any others, given
their privileged status in mental imagery. The ubiquitous
notion of imageability thus begins to bifurcate: First, there is the
basic imageability of a particular parameter in itself. In this sense, a
parameter either is or is not imageable depending on whether it
can be readily represented in a mental image. Then, there is the
impact the same parameter may or may not have on the
imageability of a visual description. Unlike contour shape and
Anežka Kuzmičová
304
orientation, all the below parameters have a pronounced
impact. It is assumed in what follows that a negative score on
basic imageability automatically entails a negative impact on
description imageability. Meanwhile, a positive score on basic
imageability does not, interestingly enough, guarantee a
positive impact. But let us begin with the easy cases.
Some visual parameters simply cannot be represented in
an image from visual description. Their presence in a visual
description is then necessarily a hindrance to mental imagery.
Rather than contributing to a visual presentation, the words
referring to these parameters leave other sorts of imprints on
the reader’s consciousness, thus disturbing the mental image;
they are reflected upon qua higher-order concepts or qua verbal
expressions, or simply skimmed or skipped. Size is a salient
example of this class of parameters. In explaining why size
cannot be imaged, I will once more revisit Elaine Scarry. In her
treatise, Scarry argues that blossoms are amongst the most
easily imageable of all possible contents. For explanation,
Scarry (1999: 47) refers to the typical size of a blossom, which
she says is commensurable with the size of the physical space
occupied by mental images, i.e. with the size of one’s forehead.
Even though I do not share Scarry’s passion for flowers,
and even though I do not posit that the mental visual field is
experienced to span a stretch of physical space, my assumption
about the non-imageability of size is grounded in a similar
premise, namely, that the spatial magnitude of visual mental
images is invariable across contents. That is to say, no matter
how small or big an object in reality, its visual mental image is
readily enlarged or diminished as if to nicely fill the blank of
the mental visual field, leaving a perfectly proportionate
margin (see also Casey 2000: 54). Evidence from empirical
studies on experimenter guided mental imagery concurs with
Fidelity Without Mimesis
305
this intuition. People have been found to consistently image
smaller objects as if they were closer and vice versa (Kosslyn
1978). Because the blank to be filled is not physical in any
respect, I am more inclined to view mental images as sizeless
rather than uniformly sized. Either way, explicit reference to
size, especially when absolute (e.g., “one foot long”) as opposed
to relative (e.g., “long”), seems ostentatiously useless and
distractive in visual description as far as the content of mental
imagery is concerned. In the following examples, for instance,
any hitherto conceived mental image may recede or even
disperse as soon as the reference to size is processed:
He showed me one of [the guns], a smoothly jagged piece of metal
over a foot long. It looked like babbitting metal. (Hemingway 1962:
182)
Beyond stands the lamp, in the right corner of the table: a square base
six inches on each side, a disk tangent with its sides, of the same
diameter, a fluted column supporting a dark, slightly conical
lampshade. (Robbe-Grillet 1965a: 144)
Yet other visual parameters are imaged with great ease and
tend to have a positive impact on the imageability of a visual
description. Color is a salient example of this class of
parameters. That visual mental images prompted by names of
manufactured objects are generally experienced to appear in
shades of white and achromatic grey unless a color is explicitly
mentioned, is an insight based on introspection (see also Scarry
1999: 22). There are, however, empirical indications toward
such a view of the default mental image. For instance, a brain
imaging study of embodied cognition (Simmons et al. 2007) has
identified a neural substrate common to the processing of
object-color word pairs (e.g., “eggplant-purple”) and actual
color processing. Interestingly, the same study has shown that
Anežka Kuzmičová
306
object words decoupled from explicit color attributes do not
activate the cortical areas in question. Pre-consciously, objects
seem to be processed as colorless. Nevertheless, while the
stimuli in the above study consisted in both natural and
manufactured object words, my own introspective hypothesis
regarding conscious and near-conscious imagery does not
extend beyond the latter. Rather, I am inclined to describe
mental images prompted by names of natural objects—
“eggplant”, “strawberry”, “emerald”—as tinted by the color
typically associated with the object (see also section 4.1). To be
more precise, I am inclined to thus describe mental images
prompted by the names of any objects which are very strongly
associated with one particular color. These objects (and the
specific colors they are associated with) vary in part across
cultures and individuals. The group happens to coincide
largely, but far from entirely, with the category of natural
objects. It probably includes bricks, but not bell peppers.
Whether default mental images of manufactured objects
really are entirely achromatic or just extremely feeble in hue,
the parameter of color is essentially different from the
parameter of size in that its experienced absence from the visual
mental image is no necessity. Not only can color be easily
accommodated by a visual mental image. It is often
accommodated with benefit, boosting the image beyond the
threshold of the reader’s attention. The rare potential of
externally induced color to inform visual imagery is further
confirmed, as it were, by empirical research into the so called
Perky effect. In the initial Perky (1910) experiment, participants
were asked to produce mental images of diverse objects (e.g., a
banana, a book) while unknowingly facing a white screen on
which dim pictures of the same kinds of objects were being
projected. The original data suggests that exposure to pictures,
Fidelity Without Mimesis
307
generally speaking, affects the content of concurrent mental
imagery. Further research (Reeves 1981) has shown, however,
that the Perky effect is up to six times stronger when the target
picture is colored compared to when it is achromatic.
Finally, the basic imageability of color being beyond
doubt, its positive impact on description imageability is
perhaps most easily avowed by recourse to practical examples.
In my view, any mental image produced by these randomly
chosen visual descriptions fades drastically when references to
color are thought away:
Right at the back would be a narrow bed covered in ultramarine
velvet and stacked with cushions of all colours. (Perec 1990: 24)
The sofa was upholstered in yellow and blue satin, shiny and tight,
finished with rolled gold cord and tassels. A hard little matching satin
bolster tucked in at either end. Gold claws at the end of twisted
wooden legs. (Roberts 1993: 54)
To encourage him, Baxter at last takes the knife from his pocket. As
far as Perowne can tell, it’s an old-fashioned French kitchen knife,
with an orange wooden handle and curved blade with no sheen.
(McEwan 2006: 215)
The third and final category of parameters is those that may be
perfectly imageable in themselves, but have a negative impact
on the perceived imageability of a visual description. The fact
that such discordance is possible at all suggests that the notion
of imageability, bifurcated as it already was for the purpose of
the preceding analysis, in fact trifurcates. In between the basic
imageability of each individual parameter and its impact on
overall description imageability, there is the mediating variable
of a parameter’s respective possibility or non-possibility to be
accommodated in the general object image. This possibility or non-
Anežka Kuzmičová
308
possibility comes down to the inherent makeup of the mental
image as described above. Namely, it derives from the
prominence of contour shape in object imagery, and from the
feebleness of the surface filling the space delineated by a
mentally imaged contour. More specifically, the category
encompasses whatever can be imaged in its own right but
cannot meet the mind’s eye when an object is imaged as a whole.
The parameters belonging to this third category all reduce, in
one way or another, to surface detail and to non-contour shape,
i.e. to aspects of shape that do not inform the general object
contour (as projected into the two dimensions of a mental
image). One salient representative of this category is what I will
henceforth call, for lack of a better expression, the parameter of
visual complexity. The underlined portions of the following
descriptions all roughly amount to visual complexity:
The table is a metal disc pierced with innumerable holes, the largest of
which form a complicated rosette: a series of S‘s all starting at the
center, like double-curved spokes of a wheel, and each spiraling at the
outer end, at the periphery of the disk. ¶ The base supporting the
table consists of a slender triple stem whose strands separate to
converge again, coiling (in three vertical planes through the axis of the
system) into three similar volutes whose lower whorls rest on the
ground and are bound together by a ring placed a little higher on the
curve. (Robbe-Grillet 1965b: 94–95)
A local craftsman had made the buffet for Thérèse’s grandparents. …
A solid piece in worn pine, darkened with age, satin-smooth. Its top
pair of doors was carved with reliefs of oakleaf garlands. Two fat
swags that hung down, one on each door. (Roberts 1993: 11)
The placards were unevenly painted and attached by fragile
aluminum rivets to posts of rolled sheet metal[.] (Franzen 2001: 225)
Fidelity Without Mimesis
309
It is fair to say that reference to visual complexity, the verbal
rendition of the detailed architecture of things, is over-
represented in literary visual descriptions. Visual complexity
thus adds to why visual descriptions, in aggregate, end up
appearing so surprisingly non-imageable, given the reader’s
intuitive readiness to see with the mind’s eye. It has previously
been suggested by literary scholars and cognitive psychologists
alike that in order to be imageable, visual descriptions of
various kinds (descriptions of faces; descriptions of complex
spatial settings) need to preserve a holistic (Jajdelska et al. 2010)
and unitary (Lopes 1995: 23) view of what is being described.
Visual complexity obviously flouts these principles, breaking
objects into details of structure (such as fragile aluminum
rivets) and details of surface (such as perforations forming
complicated rosettes). Any initial object image is then broken
down accordingly.
Although the reader may not cease to experience visual
imagery while processing references to visual complexity,
the images experienced are no longer experienced as images of
the central object proper. A sense of discontinuity obtains (see
also Casey 2000: 91), with negative consequences for the
imageability of the visual description overall. Mental imagery
from visual description, at least when discrete objects are
considered, thus differs from perceptual experience and from
perceptually mimetic (e.g., narrative) mental imagery in that
nothing can be represented in it without simultaneously being
represented as being in focus. And if whatever is in focus
optimally fills the visual mental field, then each mental image
consists only and exclusively in whatever is in focus. Hence the
necessary lack of continuity between mental images of objects
and mental images of object parts.
Anežka Kuzmičová
310
Figure 3
Never
No
No
Negative
Size
Volume
Separately
Yes
No
Negative
Visual
complexity
Occluded
features
Optionally
Yes
Yes
Positive
Color
Luminance/
Texture
By default
Yes
Yes
Neutral
Contour
shape
Orientation
Parameter
Is imaged
Imageability I:
Is imageable
Imageability II:
Can be accommodated
in object image
Imageability III:
Has impact
on description imageability
Salient example
Further example
Fidelity Without Mimesis
311
5/ Conclusion
“Fidelity without mimesis,” reads the title of this paper. Let us
once more recount its main implications. Firstly, visual
description as well as the mental imagery it sometimes elicits is
non-mimetic with respect to perceptual experience. Secondly, if
visual description is to elicit mental imagery in the first place, it
depends for its imageability on a different, non-perceptual kind
of fidelity than imageable narrative, namely on its fidelity to the
experiential makeup of voluntary visual images. When it
deviates from this makeup, visual description decreases in
imageability. But when mental imagery is elicited after all, it is
further distinguished from other forms of fiction-induced
imagery by its level of fidelity to what is actually encoded in the
text. Hence a second way of reading “fidelity”.
There is of course much more to visual descriptions than
visual parameters. Not only are there parameters relevant to
vision that have a strong potential to engage embodied,
interactive processing (e.g., weight, surface texture). Attached
to the third fundamental feature of mental imagery from visual
description, i.e. to its inherent expectedness, there are also the
countless issues of visual descriptive style. It would be
untenable to posit that description imageability is unaffected by
stylistic variation. But a detailed inquiry into imageable and
non-imageable descriptive lexicon and syntax remains to be
carried out. Moreover, a call for a yet larger enterprise lurks in
the conceptual network of the above argument, namely the call
for a systematic, positive analysis of all the other attentional foci
that can piggyback on visual description processing. For
instance, when is a visual description more likely to address
conceptual reflection or draw one’s attention to its linguistic
structure rather than elicit visual imagery? What are the mutual
relationships between the three? What are their respective
Anežka Kuzmičová
312
relationships to mimesis, and what kind of mimesis are we
talking about? Such a systematic analysis, should it be viable,
would significantly contribute to the charting out of the
regularities of prose fiction reading overall. Literary theorist
Philippe Hamon’s aphoristic remark about description “being
the crucial point at which the readability (of fiction) is
organized” (Hamon 1982: 167) would thus acquire new, clearer
significance.11
Stockholm University
anezka.kuzmicova@littvet.su.se
REFERENCES
Allington, Daniel. 2011. “‘It Actually Painted a Picture of the Village and the
Sea and the Bottom of the Sea’: Reading Groups, Cultural Legitimacy,
and Description in Narrative (with Particular Reference to John
Steinbeck’s The Pearl).” Language and Literature 20 (4): 317–332.
Atran, Scott. 1990. Cognitive Foundations of Natural History: Towards an
Anthropology of Science. Cambridge: Cambridge University Press.
Baker, Nicholson. 1998. The Mezzanine. London: Granta.
Bal, Mieke. 1982. “On Meanings and Descriptions.” Studies in 20th Century
Literature 6 (1–2): 100–148.
Barthes, Roland. 1989. “The Reality Effect.” In The Rustle of Language, trans.
Richard Howard, 141–48. Berkeley: University of California Press.
Burke, Michael. 2011. Literary Reading, Cognition, and Emotion: An Exploration
of the Oceanic Mind. New York: Routledge.
Casey, Edward S. 2000. Imagining: A Phenomenological Study. Bloomington:
Indiana University Press.
Cobley, Evelyn. 1986. “Description in Realist Discourse: The War Novel in
Narrative Poetics.” Style 20 (3): 395–410.
11 Part of this research was carried out at the University of Alberta
during a research stay funded by the Swedish Foundation for International
Cooperation in Research and Higher Education.
Fidelity Without Mimesis
313
Dickens, Charles. 1998. Pickwick Papers. Hertfordshire: Wordsworth Classics.
Esrock, Ellen J. 1994. The Reader’s Eye: Visual Imaging as Reader Response.
Baltimore: The Johns Hopkins University Press.
Fischer, Martin H., and Rolf A. Zwaan. 2008. “Embodied Language: A
Review of the Role of the Motor System in Language
Comprehension.” The Quarterly Journal of Experimental Psychology 61
(6): 825–850.
Flaubert, Gustave. 1995. Madame Bovary. Trans. Allan Russell. London:
Penguin Popular Classics.
Franzen, Jonathan. 2001. The Corrections. New York: Picador.
Gallagher, Shaun, and Dan Zahavi. 2007. The Phenomenological Mind: An
Introduction to Philosophy of Mind and Cognitive Science. New York:
Routledge.
Grünbaum, Thor. 2007. “Action between Plot and Discourse.” Semiotica 165
(1–4): 295–314.
Halliwell, Stephen. 2002. The Aesthetics of Mimesis: Ancient Texts and Modern
Problems. Princeton: Princeton University Press.
Hamon, Philippe. 1981. “Rhetorical Status of the Descriptive.” Trans. Patricia
Baudoin. Yale French Studies 61: 1–26.
———. 1982. “What is a Description?” In French Literary Theory Today: A
Reader, ed. Tzvetan Todorov, trans. R. Carter, 147–78. Cambridge:
Cambridge University Press.
Hemingway, Ernest. 1962. A Farewell to Arms. In Three Novels, 1–332. New
York: Scribner.
Herman, David. 2009. Basic Elements of Narrative. Chichester: Wiley-
Blackwell.
Jajdelska, Elspeth, Christopher Butler, Steve Kelly, Allan McNeill, and Katie
Overy. 2010. “Crying, Moving, and Keeping it Whole: What Makes
Literary Description Vivid?” Poetics Today 31 (3): 433–463.
Kosslyn, Steven M. 1978. “Measuring the Visual Angle of the Mind’s Eye.”
Cognitive Psychology 10 (3): 356–89.
Kuzmičová, Anežka. Forthcoming. “The Words and Worlds of Literary
Narrative: The Tradeoff between Verbal Presence and Direct Presence
in the Activity of Reading.” In Stories and Minds: Cognitive Approaches
to Literary Narrative, ed. Lars Bernaerts—Dirk De Geest—Luc
Herman—Bart Vervaeck. Lincoln: University of Nebraska Press.
———. 2012. “Presence in the Reading of Literary Narrative: A Case for
Motor Enactment.” Semiotica 189 (1–4): 23–48.
Anežka Kuzmičová
314
Labov, William. 1972. Language in the Inner City: Studies in the Black English
Vernacular. Philadelphia: University of Pennsylvania Press.
Lopes, José Manuel. 1995. Foregrounded Description in Prose Fiction: Five Cross-
Literary Studies. Toronto: University of Toronto Press.
Martin, Alex. 2007. “The Representation of Object Concepts in the Brain.”
Annual Review of Psychology 58: 25–45.
McEwan, Ian. 2006. Saturday. New York: Anchor Books.
———. 2011. Solar. New York: Random House.
Mosher, Harold F. 1991. “Toward a Poetics of ‘Descriptized’ Narration.”
Poetics Today 12 (3): 425–45.
Noë, Alva. 2006. Action in Perception. Cambridge, MA: The MIT Press.
Nünning, Ansgar. 2007. “Towards a Typology, Poetics and History of
Description in Fiction.” In Description in Literature and Other Media, ed.
Werner Wolf and Walter Bernhart, 91–128. Amsterdam: Rodopi.
Palmer, Stephen E., Eleanor Rosch, and Paul Chase. 1981. “Canonical
Perspective and the Perception of Objects.” In Attention and
Performance IX: Proceedings of the Ninth International Symposium on
Attention and Performance, Jesus College, Cambridge, England, July 13–18,
1980, 135–51. Hillsdale: Lawrence Erlbaum.
Perec, Georges. 1990. Things: A Story of the Sixties. Trans. David Bellos.
Boston: Godine.
Perky, Cheves West. 1910. “An Experimental Study of Imagination.” The
American Journal of Psychology 21 (3): 422–52.
Reeves, Adam. 1981. “Visual Imagery Lowers Sensitivity to Hue-Varying,
but not to Luminance-Varying, Visual Stimuli.” Attention, Perception,
& Psychophysics 29 (3): 247–50.
Riffaterre, Michael. 1981. “Descriptive Imagery.” Yale French Studies 61: 107–
125.
Rizzolatti, Giacomo, and Vittorio Gallese. 1988. “Mechanisms and Theories
of Spatial Neglect.” Handbook of Neuropsychology 1: 223–46.
Robbe-Grillet, Alain. 1965a. In the Labyrinth. In Two Novels by Robbe-Grillet,
trans. Richard Howard, 139–272. New York: Grove Press.
———. 1965b. Jealousy. In Two Novels by Robbe-Grillet, trans. Richard
Howard, 33–138. New York: Grove Press.
Roberts, Michèle. 1993. Daughters of the House. London: Virago.
Rosch, Eleanor, Caroline B. Mervis, Wayne Gray, David M. Johnson, and
Penny Boyes-Braem. 1976. “Basic Objects in Natural Categories.”
Psychology 7: 573–605.
Fidelity Without Mimesis
315
Scarry, Elaine. 1999. Dreaming by the Book. New York: Farrar, Straus and
Giroux.
Shepard, Roger N., and Jacqueline Metzler. 1971. “Mental Rotation of Three-
Dimensional Objects.” Science 171 (3972): 701–3.
Shklovsky, Viktor. 1990. Theory of Prose. Trans. Benjamin Sher. Normal:
Dalkey Archive Press.
Simmons, W. Kyle, Vimal Ramjee, Michael S. Beauchamp, Ken McRae, Alex
Martin, and Lawrence W. Barsalou. 2007. “A Common Neural
Substrate for Perceiving and Knowing about Color.” Neuropsychologia
45 (12): 2802–10.
Thompson, Evan. 2007. “Look Again: Phenomenology and Mental Imagery.”
Phenomenology and the Cognitive Sciences 6: 137–70.
Westbury, Chris, and Gail Moroschan. 2009. “Imageability x Phonology
Interactions during Lexical Access: Effects of Modality, Phonological
Neighbourhood, and Phonological Processing Efficiency.” The Mental
Lexicon 4 (1): 115–45.
Wolf, Werner. 2004. “Aesthetic Illusion as an Effect of Fiction.” Style 38 (3):
325–51.
———. 2007. “Description as a Transmedial Mode of Representation:
General Features and Possibilities of Realization in Painting, Fiction
and Music.” In Description in Literature and Other Media, ed. Walter
Bernhart, 1–87. Amsterdam: Rodopi.
