HAPTER 8 VISUAL IMAGERY AND

SPATIAL COGNITION

CHAPTEROUTLINE

Codes in Long-Term Memory

The Dual-Coding Hypothesis
The Relational-Organizational Hypothesis
Empirical Investigations of Imagery
Mental Rotation of Images
Scanning Images
The Nature of Mental Imagery
Principles of Visual Imagery
Implicit Encoding
Perceptual Equivalence
Spatial Equivalence
Transformational Equivalence
Structural Equivalence
Critiques of Mental Imagery Research and Theory
Tacit Knowledge and Demand Characteristics
The Picture Metaphor
Propositional Theory
Neuropsychological Findings
Spatial Cognition

T hink of the house or apartment you consider your permanent residence. In part
lar, think about its kitchen. How many cabinet doors does it have? Obviously, ths
question draws on your memory. Most people can answer it after some mental
What sort of work is required? In the process I used, Ifirst recognized that I didnt
the information needed already stored; that is, Ididn't know the answer "off the top
head." So Ihad to determine the answer in another way. Imentally pictured my kitcne
drawing on memory. Then, starting at one end of the room, I scanned my mental pl
 Chapter 8: Visual Imagery and Spatial Cognition 187
countingcabinet doors. My procedure is neither difficult nor original (Shepard, 1966)
be the one commonly used.
seemsto
but
t Se
nature of these
"mental pictures," or visual images, is one focus of this chapter. We
The role ofimages in memory. We will also consider experimentsinvestigating
lookatthe people construct and use visual images and what these findings sug-
will in which will
ways
the about cognition. Finally, we turn to the nature of visual images, considering the
gest representations used to create and store them.
kinds ofr
mental
will Confine ourselves to a discussion of visual images.
Throughout the chapter, we
Recognize, however, that other kinds of mental images exist. Examples include auditory
(such as theimagined sound of your dog barking), olfactory images (such as
images fresh-baked bread), and cutaneousimages (such as the imagined
imaginedsmell of
the your toe being stubbed into the wall). Visual images, like visual perception, have
feeling of most attention within cognitive psychology. Thus, just as when we examined
received the focused on visual perception, in this chapter we will focus on
perception(Chapter 3) we
visual imagery.
of visual imagery has had a controversial history within psychology (Paivio,
The study occasional references to imagery were made at the turn of the 20th
1971). Although of behaviorism essentially dictated that even the concept of an image
century, the rise
Visual images are problematic as objects of scientific inquiry. After all, the
be rejected. of a visual image is
experience just about as private an experience as one can have. If
reallv
Iagsert that Iam forming a visual image of my kitchen, no one but me can tell if I
seen,
have the image or am just pretending. VIsual images, unlike behaviors, cannot be
counted. or controlled by other people. Because visual images can be reported only by
the person who asserts she is experiencing them, that person can distort or bias them,
either consciously or inadvertently. Behaviorists argued that imagery is not the sort of
topic that can be investigated with sufficient scientific rigor or control.
Nonetheless, interest in visual imagery never completely vanished (Paivio, 1971) and in
fact became stronger after the popularity of behaviorism waned in the 1960s. It is difficult
to explain how people perform certain cognitive tasks, such as the one described earlier,
without talkingabout visual images. Moreover, research on memorypeople
suggests that people
who do not.
who report using imagery are better able to recall information than
Sports psychologists, too, have a strong interest in the use of visual imagery. An athlete
who betore competingspends time mentally imagining a smoothly executed,well-timed,
elegant performance has been shown to perform better a bit later when engaging in the
sport (Martin, Moritz, &Hal, 1999). Some research suggests further that imagery can
be used to help people cope with negative emotional events, such as remenbering a
real incident of being rejected,atbandoned, or excluded. Research participants asked to
ViSualize so-called coolaspects of the experience--for example, where they were stand
ng or sitting in relation to other people during the incident--were better able to reduce
her hostile feelings than were participants asked to form images of their visceral reac
tions during the incident or participants not asked to form any images (Ayduk, Mischel, &
Downey, 2002).
and
Psychologists now recognize that to eliminate imagery as a subject of discussion
investigation is to overlook a potentially fundamental aspect of cognition. Hence, visual
imagery hasstregained credibility assaworthwhile topic among most cognitive psychologists.
38 Cognitive Psychology In and Out of the Laboratory

CODES IN LONG-TERM MEMORY

If you recall our discussion of different mnemonic techniques from

rememberChapter
ber that mnemonics are techniques used to help people
For our current purposes, I hope you remember that several
6,
certain you
not al) remem
(though
techniques involve visual imagery: method of loci, interacting
method. You might have wondered why so many mnemonic images, and the
techniques
imagery or howimagery-based mnemonics function differently from infoUsermmpeeagrtmWiooOrnvnii,sC.ud,A
mnemonics. We willconsider two opposing views on this matter no
THE DUAL-CODING HYPOTHESIS

Allan Paivio (1969, 1971, 1983) originated the dual-coding

non-
hypothesis
im
of
agery-basen
explain the workings of various mnemonics. According to
contains two distinct coding systerns (or codes) for representingPaivio, long-term
information to
One is verbal, containinginformation about an item's abstract. linguistic be
memory to
mestomroyed,
other involves imagery: mental pictures of some sort that represent what meaning.
like. Items to remember can be coded by either verbal labels or visual the item looksThe
some cases, both. Paivio's idea is that pictures and concrete words give
verbal labels and visual images; that is, they have twO possible internal images
rise toand, in
codesoh both
representations. Abstract words, in contrast, typically have only one kind of
resentation: a verbal label. code or rep-
One study by Paivio (1965) provided evidence to support this hypothesis.
Particinar.
were asked to learn one of four lists of noun pairs. The first list (CC)
incluuded
which both nouns referred to concrete objects (for example, book-tatble). The second pairsiein
(CA) included pairs in which the first noun was concrete and the second abstract (suchae
chair-justice). The third list (AC) was the converse of the second (such as freedom-dree
The fourth (AA) contained pairs of abstract nouns (for example, beauty-truth). Of a pos.
sible 16 correct responses, participants averaged 11.41, 10.01, 7.36, and 6.05 corect
responses for the CC, CA, AC, and AA lists, respectively.
Paivio (1965) explained the results as follows. Whenever possible, participants spontane
ously formed visual images of the noun pairs. The formation was easiest with concrete
nouns. Paivio (1969) assumed that visual imagery, unlike verbal labeling, increases as a
function of concreteness: the more concrete the noun, the richer the image and the more
elaborated the internal code. This helps explain why pictures (very concrete) are often
remembered better than words (see, for example, Kirkpatrick, 1894; Shepard, 1967,
When items are coded by both images and verbal labels (as concrete nouns can be,.
the chances of the learner's retrievingthem are obviously better. If the learner forgets the
verbal label, he or she might still access the visual image, or vice versa. Items coded ony
by verbal labels are disadvantaged; if the verbal label is forgotten or "misplaced]" the
learner has less to go on.
Further, Paivio (1969) believed that the first noun in a pair (called the "stimulus"nou
serves as a Conceptual peg on which the second (response") noun is hooked. In tis
sense, the stimulus noun serves as a "mental anchor." a place to which the represel
tion of the response noun can be attached. Thus the imaginability of the first nounP
particularly important in improving memorability, explaining why recall in the CA COndiu
was significantly higher than in the AC condition.
 (hapter &: Visual Imagery and Spatial
Cognition 189
otRELATIONAL-ORGANIZATIONAL HVYPOTHESIS
Bower(1970)proposed an alternative to the dual-coding hypothesis, which he called the
elational-organizationalhypothesis. He believed that
imagery improved
hecauseimages are necessarily richer than vertbal labels but because imagerymemory not
produces
assoCiations between the items to be recalled. Forming an image (say,
morewordsin a pair or between a
word and a as in the method of loci) between
location,
two
typically
personto create a number of links or hooks between the information to be
requiresthe andotherinformation. Recall from Chapter 6that the more retrieval cuesa
remembered
information in memory has, the greater are the chances of recalling it. Bower's
argument,
piece then. is that imagery works by facilitating the creatior of a greater number of
of
thatlinkthe
two to-be-remembered pieces of information.
hooks
performed an experiment to distinguish between the dual-coding and the
Bower(1970)
relational-organizational hypotheses. Participants were ivided into three groups, each
differentinstructions for a paired-associates learning task. One group was told to
given repetition" (that is, to rehearse aloud); the second, to construct twoimages
Use overt
rote
interact and were "separated in imaginal space"; the third, to construct an
thattddidnotscene of the two words in a pair (p. 530). Results showed that all participants
interactive
recognized about 85% of the previously seen words. However, recall of those words
Those who used rote memorization recalled about 309% of the paired
differedgreatly.
those who used noninteractive imagery, 27%; and those who formed interact-
associates;
about53%.
ingimages,
to elaborated coding of the paired associates, as the dual-cod-
Ifimagerysimplyled omore participants in thetwo conditions that involved instructions
inghypothesis
predicts, then
similarly. In fact, only those who formed
to form two images ought to have performed the rote memorizers. Apparently, it is not
interacting images showed an improvement over
in which imagery is used. Interacting
imagery per se that helps memory but rather the way
images presumably create or suggest moreeasier links between the target information and other
to retrieve.
information, making the target information
Athough the dual-codinghypothesis continues to attract proponents (see Yuille, 1983),
workings of imagery mnemonics and what
stilunresolved are how well it explains the mnemonics. However imagery mnemon
kind of explanations it provides for nonimagerydo aid memory. To understand how these
many
ics work, there is at least little doubt that what imagery is and how it works,
mmemonics work, it will be necessary toexplore further
topics we turn to next.

EMPIRICAL INVESTIGATIONS OF IMAGERY

Series of studies by Lee Brooks (1968) is widely regarded as yielding some of the best
A different pro
verbal materials or at least invoke conditions
ellce tnat images are distinct from 8.1 depicts different
cesses from those invoked by verbal materials. Figure asked toimagine a
of Brooks's primary experiment. In one condition, participants were clockwise mentally
then to move
letter, such as the outlined capital Fin Figure 8.1(A), andasterisk)
from a particular corner (markedin Figure 8.1 with an and to indicate, for each
example,
extreme bottom of the letter. In this
Corner, whether it was at the extreme top or
the correct responses are "yes, yes, yes, no, no, no, no, no, no, yes."
 Inand of
Out theLaboratory
190 CognitivePsychology

ABIRDIN THE HANDIS NOTTIN THE BUSH.

For each wordin the sentence above.
with an
corner marked indicate whether or not each wordiis a
Start at the noun.
indicate whether or
asterisk and extreme
is at the
not each cormer
top or bottom. (B)
(A)
Brooks's 1968study.
Figure 8.1: Stimuli from
Participants indicated their responses in different ways. One mode of respoNse Weg
verbal: Participants said "yes" or "no," as noted. Another response mode was spatial.
Participants were given a response sheet on whichthe letters Yand N were printed
irregular pattern and were told to point to either a Yor an Nin each row to indicate the
participants took almost 2.5 times longer
responses. Brooks (1968) found that
responded by pointing than they did
when responding verbally.
On a secondtask, Figure 8.1(B), participants were asked to remember a sentence, such
wwhenthey
as "A bird inthe hand is not inthe bush," and, for each Word, to indicate whether it was
no, yes, no, no, yes. no
a concrete noun. In this example, the correCt responses are
responded verbaly s
no, no, yes." As with the previous task, sometimes participants
this task, however na
other times they pointed to Yor N ona response sheet. With
were faster to respond by pointing than they were to respond verbaly (although the cifio
ence in response times was not as great).
One explanation for these results. is as follows. The first task requires the formation
a visual image of an F The visual image probably has at least some picturelike
qualites
(spatial or visual), so a spatial or visually guided response (pointing) would be intertered
imaoeis
with to a greater extent than would a verbal response. lnother words, the visual
more disruptive of, and disrupted by, another spatial or visual type of task (pointing) than
bya vertbal kind of task (talking). The converse is also true: Holding a sentence in memory
a verbal task) is easier to do with a concurrent visual/spatial task (such as pointing than
with another verbal task. Notice that pointing or talking do not differ in difficulty overallbut
vary in difficulty as a function of the task with which they are being performedBrookss
(1968) worksupports the idea that images and words use different kinds of internal codes
(as the dual-coding hypothesis suggests).
Brooks's (1968) task is not the only one that apparently requires people to fom vsua
images. Here is another. Answer the following question: Which is larger, a pineapple
a coconut (Finke, 1989)? To answer the question, you mostlikely constructed a Sd
image of acoconut next to a pineapple and "read" the answer from your image.
Moyer
the two(1973) asked
objects (in similar questions
his study, animals)anddiffered
found that people
greatly. This were faster
effect, to respond
called w
the symbolic
distance effect, works as follows. Other things being equal, you'd be faster to answerthe
question "Which is bigger, a whale or a cockroach?" than the question "Which is bigger,
a hog ora cat?" Interestingly, the same pattern of response times is also obtained when
people look at actual objects (Paivio, 1975). In other words, you'd be faster to answer
 Chapter 8: Visual Imagery and Spatial Cognition 191
thefirst question even if, instead of consulting mental images, you looked at the actual
animalsorat photographs of the animals. This result suggests that images seem to func-
tinsome ways, like pictures. If people merely retrieved verbal
tion,atleast information (for
example,from asemantic network Such as those networks described in Chapter 7), it
wouldbe
difficult to explain this pattern of results.
ROTATION OFIMAGES
MENTAL
precedingstuudiessuggest that people create and use visual images to answer cer-
Thequestions and perform certain tasks. They also suggest that the images created are
tain picturelike(altthough this conclusion has been energetically debated, as
some ways
in
Atthe same time that these findings were reported, other studies showed that
we'llsee). more than simply create images; they could also, apparently, mentally
people could do
transformthem.
famous
One of the most
type was
studies of this and
perfomed by Shepard
Metzler (1971). They showed
participantsperspective line
drawings of three-dimen
8.2
sional objects (Figure
Figure .2: Stimuli from the Shepard and Metzler (1971) study.
presents examples). On each SOURCE: Shepard, R. N., &Metzler, J. (1971). Mental rotation of three
trial, participants would see dimensional objects. Science, 171, p. 701. Copyright G 1971, American
two drawings. In some cases, Association for the Advancement of Science. Reprinted with permission.
the two drawings depicted
the same object but with one rotated by some degree. In the other cases, the drawings
depicted mirror-image reversals; in other words, the objects were similar but not identi
rotations used were
cal. The mirror images were also sometimes rotated. The kinds of
either in the picture plane (that is, as if
the drawing were rotated on the page) (sec)
5
Or in depth (that is, as if the object were
going toward or away from the viewer). time
Shepard and Metzler found that the reaction
amourt of ime jt took participants to
decide if the two drawings depicted the Mean
same object or a miror-image reversal
MAs directly proportional to the angle of
[otation between the drawings.
Hgure 8.3 shows their results. This close
100 120 140 160 180
COmespondence between the angle of 20 40 60 80

rotation of the two drawings and the Angle of rotation (degrees)

participants' reaction times strongly Picture-plane pairs o Depth pairs
SUggests that they performed the task
Dy mental rotation_of one drawing. Metzler (1971) study.
Figure 8.3: Results from the Shepard and
Moreover, the time it took participants SOURCE: Shepard, R. N., &Metzler, J.
(1971). Mental rotation
Come tO a decision was the same Science, 171, p. 701. Copyright
for of three-dimensionalobjects. Advancement of Science.
rotations in the picture plane and in ©1971, American Association for the
depth, This Suggests they were mentally Reprinted with permission.
 Laboratol)
PsychologyInand0utofthe
Cognitive
rotating three-dimensional images, notjust thetWO-dimensional drawings. Had partic
pants beenrotating onlythe latter,their performance would have differed as atunGion
picture plane or in depth.
was in the
whether the rotation
studies by Cooper and ADVANCE INFORMATION
Later showed TEST
Shepard (1973, 1975)mentally.
also
that participants
rotated more recognizable. Identity
Orientation

alphabetlet
stimuli, such as
drawings_ofhands. In R
ters or Shepard,
&
one study (Cooperwere some 100 msec
1973), participants of the
2000 msec
400
times given a drawing trial, fol 700
letter to be used on a 1000
lowed by a cue showing the ori
stim
entation to which the test Fiqure 8.4: Cooper and Shepard's 1973experimentaldesion
would be rotated, before
ulus
the test stimuluS appeared. It SOURCE: Cooper, L. A., &Shepard, R. N. (1973). The time required to pre
presented pare for a rotated stimulus. Memory and Cognition, 1, p. 247.
these twO CUes were
(for example, 1973, Psychonomic Society, Inc. Reprinted with permission. ¬ Copyright
early enough
1,000 milliseconds before the
test stimulus appeared), then 1100
the participants' performances
were the same for all angles of 1000 -
Group data (N =8)
rotation. Figure 8.4 depicts the B-100
experimental conditions, and reaction
(msec)
time

Figure 8.5 shows the results.

900

Note the shape of the curves in (B-400

800 4
Figure 8.5, which suggests that
participants were able to men 700
Mean
tally rotate. their images either.
B-700
clockwise.or.cOunterclock 600 A
Wise, depending on. which.
direction Iled to alesser angle. 500
These results differ from those B-1000
of Shepard and Metzler (1971), 400
as a comparison of Figures 8.3
and 8.5 shows, presumably 60 120 180 240 300 360
because alphanumericch¡r Orientation of test stimulus (degrees, clockwise from uprighi)
acters have a known"upright" No information -- 100 msec
- - f - 400 msec
position. whereas Shepard and Combined
Metzler's line drawings do not - A 700 msec - 1000msec information
By the way, one reason for the
"peaks" in reaction times at
180 degrees might be that par Figure 8.5: Results from Cooper and Shepard's 1973 study.
required topre-
ticipants were uncertain about SOURCE: Cooper, L. A., &Shepard, R.N. (1973). The time Copyright
pare for arotated stimulus. Memory and Cognition, 1, p. 248.
which direçtion to rotate the
figure and thus hesitated. 1973,Psychonomic Society, Inc. Reprinted with permissIOn.
 Chapter 8: Visual Imagery and Spatial
Cognition
participantsin these experi- Standard
Are
mentally rotating the whole Reflected Standard
only at Reflected
or are thhey looking
ments
stimulus, this ques-
parts? To answer
certain Cooper (1975) performed
Lynn
tion, presented participants Form A
that 6 points Fom E
iregular polygons,) such as
stualies_
12 points
with shown in Figure 8.6. The
those wereformed by connect-
polygons
randomlyScattered number of
Form
8 points
a with more complex polygons 16 points
points, greater number of
resultingfrom a
points.
Participants were first trained
discriminate between original Form C Form G
Todmirror-image reflections of the 8 points 16 points

polvaons. Next, they were shown

and

ether the original

polygons or the
refectionsat different angles of
rota- Form D Form H
to determine 12 points
were asked 24 points
ton and was the
whetherthe object
depicted JFigure 8.6: Stimuli from Cooper's 1975 study.
original or a reflection of the original.
Cooper(1975) found that the reaction times once again increased linearly with the angle of
rotation andthat therate of rotation was the same for allthe polygons, regardless of their
complexity,.If participants were attending only to parts of the polygons, then performance
Ought to have differed as a function of the polygon complexity.Instead, it appears that
participants mentally rotatedIentire polygons, treating the very simple polygonsin exactly.
the same manner as they did the very compleX ones.
In another study, Cooper (1976) showed that mental rotations, like_physical rotations, are
ontinuous in nature. Her demonstration worked as follows. She determined, for each
nerson. his or her rate of mental rotation. To do this, she showed participants a polygon
at a particular orientation. The polygon was removed, and participants were asked to
start mentally rotating it in a clockwise direction. As they were doing this, a te_t shape
(the polygonor its mirror image reflection) was presented in some orientationithe test
shape was presented at the orientation Corresponding to the orientation at which the
participants' visual images would be expected to be, their reaction times were always
fast As the disparity between the actual orientation of the test shape and the expected
oientation of the visual image grew, the reaction times to respond grew longer.
Ihese resuts in particular suggest that mental rotation worka like physical rotation. If you
the drawing
uaw a snape on a piece of paper and slowly rotate the paper 180 degrees, Similarly,
Wil pass through intermediate orientations: 10degrees, 20 degrees, and so0 on.
appears from Cooper's (1976) work that rotating images pass through intermediate
angles of orientation.
whether
Since Cooper's landmark studies, other cognitive psychologists haveatstudied unusual angles.
and how people use mental rotation in recognizing objects presented
Consi der, for (B). How do you recog
example, the object(s) depicted in Figure 8.7 (A) and that
nize (A) as ofdepicting (B)? One orpossibility is you mentally rotate
an image (A) until the
it reaches some ascanonical,
same object standard, orientation of depiction,
 Cognitive Psychology In and 0ut of the Laboratory

such as that shown in (B). Tarr and Pinker

(1989) and Gauthier and Tarr (1997a, 19975)
provide evidence of mentalrotation in rec
ognizing two-dimensional shapes drawn to
resemble asymmetric characters. Biederman
and Gerhardstein (1993), in contrast, argue
that when people view three-dimensional.
objects (or line drawings of them), as long as
the distinctive geons fthe basie-geometrie
components shown in Figure 3.10) of the
object remain visible, people can recognize (A)
the object without performing mental rota (B)
tion. This debate )is very much ongoing Figure 8.7: Two views of achair.
However, notice that both sides of the debate
employ conceptsand models used to explain perceptual phenomena.
SCANNING IMAGES

The research reviewed so far suggests that people can construct and transfor a.
visual images. This evidence also seems to suggest that images are in many Wavs E
pictures:They contain visual information, and the kinds of transformations
themseemto correspond to simlar transformations on pictures. Another seriesperformed
of studcies,
caried out by Stephen Kosslyn, investigated the spatial properties of images. The se
typicaly required participants first to form avisual image and then to scan it, moving
one location to another in their image, a process known as imaginal scanning. The ita
is that the time people take to scan reveals something about the ways images represent
spatial properties such as location and distance (Finke, 1989).
In one study, Kosslyn (1973) had participants
study drawings of objects such as those shown
in Figure 8.8. Notice that these drawings are
elongated either vertically or horizontally and
that each has three easily describable parts:
two ends and the middle. After the initial
learning phase, participants were told to form
an image of one of the drawings and then to
"look for" a particular part (for example, the
petals of the flower). Some participants were Figure 8.8: Stimulifrom Kosslyn's 1973 study.
told to focus first on one part of the image (for
example, the top or the left) and then to scan, looking for the designated part. Kosiy
results showed that the longer the distance from the designated end to the locaton
the part, the longer it took people to say whether the part they were looking tor we
the drawing. So, for example, participants told to form an image of the flower and to
scanning at the bottom took longer to "find" the petals (at the top of the drawng)
they did to "find" the leaves (in the middle of the drawing). Presumably, this is because u
visual image formed preserves many of the spatial characteristics of the drawings
that parts of the drawings that are separated in space are also separatedintheimage.
argued
The results of the study were not entirely clear, however. Lea (1975), for instance, i7tho
that perhaps the reaction times increased, not because of increased distance
 Chapter 8: Visual Imagery and Spatial Cognition 195

because of the number of items in the image that had to be scanned. Notice
image,but that, if one started
from the bottom, one would scan over the roots
floWer example
the
in the, leaves on the way to the petals but only over theroots to get to the leaves. Lea
and
reportedresults
supportingthis interpretation.
Ball, and Reiser (1978) performed another series of studies of image
reply,Kosslyn,
n theyfirst created a map of afictional island and had participants memo-
Scanning.Indone, seven
thelocations
of objects shown on the map, depicted in Figure 8.9. Notice that
rize
seven objects allow for the construction of 21 distinct paths-for example, from the
the
thelake and from
the tree to the hut. The paths vary in length, from 2cm to 19 cm.
treeto
to focus
Participantswereinstructed
on one object. Afew seconds
mentally named lanother
the experimenter
later,
object on the island,
and participants
asked|toimagine scanning
werethen
by imagining
n this second objectmoving across
small black speck
line. They were
the map in astraightbutton when they
instructed to push a and
"arrved" at the second object,
recorded,
their reaction times were between
The reaction times to scan the dis
objects were correlated with
al.
tance between objects (Kosslyn et more
1978); that is, participants tookdistant
time to Scan between tWO
objects than they did to scan between
two nearby ones. This reinforced the I Figure 8.9: Stimuli from the Kosslyn et al. (1978) study.
idea that images preserve spatial rela when the stimulus was a
tions. Related studies by Pinker (1980) showed similar results box).
three-dimensional array of objects (toys suspended inside an open
is in some ways
Kosslyn's work suggests that people's scanning of their visual images between two parts,
similar to their scanning of actual pictures: The greater the distance least some spatial
the longer it takes to scan between them. Images apparently depict at
These conclusions
information, and people can retrieve this information from their images. 1980).
pictures" (Kosslyn,
have strengthened the metaphor of images as kinds of "mental
however, is work by Barbara
Adding some interesting wrinkles to Kosslyn's conclusions, maps. Before reading further,
lVersky (1981)on people's systematicerrors in memory for
following cities: Seattle;
Close this book,drawa map of the United States,and in it theBoston; Portland, Maine;
put
Portland, Oregon; Reno; Los Angeles; San Diego; Chicago;
Philadelphia; New York: and Washington, D.C. Presumably, to carry out this task you
of a map of the United States, formed
de drawing on a previously stored mental image from staring at a vinylplace
Pemaps in your fourth-grade geography class or maybe even
mat showing the 50 states.
questions: (a) Which city is farther
NOW, refering to your image, answer the following
farther west, New York City or Philadelphia?
On, Boston or Seattle? (b) Which city is Diego? Now look at Figure 8.10, which shows
9 Which city is farther east. Reno or San
 Portland, ME
Toronto
Paris
Seattle
Portland, OR Boston Monaco,
Madrid
Reno, Chicagd New York City
Washington, DC Philadelphia
40° Rome
Los Angeles Algiers
.San Diego

Equator

IFigure 8.10: Map of Europe and the United States with selected cities (cylindrical projection).

the actual locations of these cities. If you are like B. Tversky's Stanford
ticipants, you made errors on questions (a) and (c). Tversky (1981) argued Unithatversity
maps are systematically distorted because people use different heuristics, or peopl
rules
e'sot
thumb, in orienting and anchoring oddly shaped units such as continents or
principles of perceptual organization,such as those discussed in Chapter 3, states.
peoplo tUsing
"line up" things to make them more orderly. Thus, South America is "remembered" inon
image as being directly south of North America instead of southeast of North America a
itactually is.
A similar principle applies to your siting of the various cities on your
map. You prob
ably know that the state of California is west of the state of Nevada, a fact largely true.
However, parts of Nevada are west of parts of California. In fact, San Diego is east of
Reno, not west. And Seattle is significantly north of Boston. But your knowledge of the
states' relative locations, combined with your propensity to make your mental image of
the map more aligned, contributes to systematic distortions. These distortions are one
way in which mental images are not like mental pictures.
Another way is found in the work of
Chambers and Reisberg (1992). They
first asked their research participants
to form an image of the creature shown
in Figure 8.11(A). You might recognize (A) (C)
the creature as the ambiguous "duck/ (B)
rabbit" shown in many introductory
psychology textbooks. Figure 8.11: Test stimuli for Chambers and Reisberg's
Sometimes experiments:
experimenters told participants that the (C) modification(A)onunmodified fiqure, (B) modification on the ducksDM,
the rabbit's nose.
creature was a duck; other times, they
said it was arabbit. They presented the actual drawing for only about 5 seconds (enough
time to form an image of the figure but not enough time to
"reverse" the figure).
 participantshad formed an
lnagery and Spatial (ognition
image, 197
once and (B) or (A) and (C),
(A) and they were
were asked presented
to choose with a pair of
pesented,You'll notice that the
elther
Chambers and Reisberg distinctions
(1992) between any pair
duck/rabbits,
which had actually been
were well above found that when
detect. are subtle and
hard to
a uck
imagingalteration
they
chance
in (B) is to the duck's bill]
at
but detecting
part
the ic ip ants
differencethought they were
Bthe[the.alteration in (C) is to the
rabbit' s could not clearly between (A) and
and (C) who had formed an initial
image
nose).
of Exactly the oppositedistinguish between (A)
reasonfor this effect is that people paid rabbit. Chambers and pattern emerged
tor
those a
the creature's "face" and less to
the more attention to the Reisberg believe
the back of the region they took to
be showsthat people who
result form images of the same creature' s head. In any case,
the
different construals or meanings to the stimulus, physical stimulus, but who give
Chambers and Reisberg report from their previous actually form different images. In
participants spontaneously reversed their work that even with hints and fact,
few elooking at the picture of the duck/rabbit
everyone image did. of the duck/rabbit, althoughprompts,almost
So far as if it were
This review Sounds
mental images. But a recent study byalways beneficial to be able to construct and
use
cOunterexample. They studied people Knauff and Johnson-Laird (2002)
reasoning with what are called provides a
problems, Such as the following: three-term series
Tandy is furrier than Bussey.
Bussey is less furry than Eskie.
Which dog is furriest?
Ta athos varied the kinds of terms used in the
hoth visually and spatially, such as above-below orproblems. Some were easy to envisage
front-back. In other words, it's easy to
mentally image one person being in tront of or behind another, and it is easy to
denict the three terms. Let's say, for example, that the premises state: mentally
Tandy is in back of Bussey.
Bussey is in back of Eskie.
Then it is easy to spatially depict the relative positions of the three dogs without forming
avisual image of them, as in this "map":

(ront) Eskie Bussey Tandy (back)

Notice that this representation doesn't show visual details of any of the dogs. Knauff and
Johnson-Laird (2002) created other problems with different content, which were easy
to torm mental images of but not quite as easy to form spatial representations of (e.g.
Ceaner-dirtier, fatter-thinner). If you form a representation of "Tandy is dirtier than Bussey."
oinsiance, you probably do it by constructing a visual image of one dog more covered
nmud than another one. There were also control problems, which were not easy to Torm
aly Knd of image or spatial representation of (e.g., better-worse, smarter-dumber).
Results showed that visual relations (for example, cleaner-dirtier) slowed down perfor-
mance relative to either control problems (for example, better-worse) or visuospatial
problems (in back of-in front of). Plausibly, the mental effort devoted to constructing the
viasual images used up mental capacity that could have been focused solely on drawing
logical conclusion. Thus, imagery is not always a boon to cognitive performance-alot
epends on the nature of the task at hand.
 theLab
PsychologyInandOutof
Cognitive
198 IMAGERY
MENTAL
NATUREOF share some properties
THE thatimages looking at with
kinds ofresults reviewed
mental transfomations
Allthe typicallyreporttheir
suggest
sofar done on
experience ofimages
as mental pictures,
images seem very similar to transiormationg pictures
heand
People questions: Just what are images?
leads directlyto the these like and unlike the properties What kinde
done on
pictures. This have, and how are
properties do
images that rea
of
pictureshave? implications for the way
Presumably, answers to such
and used.
questions
Research
have
on visual
imagery, then, can potentialy
represented and|organized. Our
informationusa
tel is
stored, retrieved,
knowledge
great deal
representation
about how and
information is mentally
concepts (Chapter 7) focused primarily
another kind of
on coverage
verbal inforna- of

tion. Research on visual

and used.
imagery suggests there may be
informationthat
has been intense cognitive psychology.
is stored in
nature of visualimages at
taking a close look theimage-as-mental-
Wewil
(1989) pr-ipnicictpulrees
Debate overthe debate here,
the
review highlights of this discussion, we will first
review Ronald Finke's
metaphor. To organize critiques ofthis research and of the image
Then we will examine -mental
of visualimagery.
picture metaphor.
IMAGERY
PRINCIPLES OF VISUAL
imagery, taken together, are meant to deserihs
of visual
Finke's (1989) principlesproperties There are five principles, and e
of visual images.
fundamental nature and characteristicof imagery.
covers a different aspect or
Implicit Encoding
principle of visual imagery states that "mental imagery is instrumental in
Finke's first of objects, or about physical relation
retrieving information about the physical properties
objects, that was not explicitly encoded at any previous time" (1989, .
ships among places from which some information can b¹
7). This principle implies that images are
intentionally stored. Imagery can thus be
obtained, even if that information was never
don't have a directly stored answer.
used to answer questions for which you probably were asked about the number of cabinet
The task at the beginning of this chapter-you case in point. My guess is that
doors in the kitchen of your permanent residence-is areason to count kitchen cabinet
much
if you are like most people, you've never had directly in long-term memor,
doors. So this information was probably not represented
However, the information was implicitly encoded, meaning it wasimage stored unintentionaly
of your kitchen.
along with other information that allows you to construct a visual
To answer the question, then, all youneed to do is form the visual
image, scan it, ano
cOunt cabinets.
Brooks's (1968) task, in which people had to answer questions about an outlined capltal
Eprovides another illustration. Most people have never bothered to check whether eac
corner of an outlined capital F is at the top or bottom of the letter. Yet peopleimplicity
are a
to perform this task, presumably because the required information has been
encoded together with the information that allows them to form a visual image of an
 apler 8: Visual lmagery and
Spatial Coqnition 199
eeptualEquivalence
second principle of visual imagery has to do with the
Finkes of visual images and the similarities
constructon
functionally equivalent to perception of real
objects and events. Itbetween the
"imagery iss
systemare activated when objects or
perception to the extent that similar states that
vsual events are
oreventsare actually perceived" (1989, p. 41). In otherimagined words, as
manywhen mechanisms
of the
the same
in theof
kinds
same objects
prOcesses used in mental
internal visualization are used in visual
perception as well.
study by Perky (1910) bears on this
Anearty werelooking at an object (such as a principle.a Perky had participants
they
a blank screen.
After they tomato, banana, an orange, aleaf) imagine
that
at
staring by one experimenter reported having
while another two formed the image, they were briefly
while
distracted
experimenters
thatprojectedfaint pictures of the objectsthe participants wereimagining. operatedPerkyan apparatus
found that
manyofthe participants were unable to distinguish between their own images and the faint
pictures. Presumably,
this is because images share
many similarities with faint pictures.
Arelatedgroup of studies, including many more experimental controls, was reported by
MarthaFarah (1985). Participants were asked to form an image of a certain
example, an H or a T. Very sOon after, they were sometimes presented with oneletter-for of these
lettersbut at a
lowlevel of contrast, making the letters very
difficult to see. Those who
imagined aletter first were more accurate at detecting the actual
they were at detecting another letter. These results suggest that imagery presented letter than
can "prime" the
visual pathway used in detecting an actual stimulus (Finke, 1989). Some authors even
regard visual imagery as perceptual "anticipation":
the visual system "getting ready" to
Actually see something (Neisser, 1976).
SpotialEquivalence
Enke's third principle of visual imagery has to do with the way that spatial information,
such as location, distance, and size, is represented in visual imagery. The principle states
that "the spatialarrangement of the elements of a mental image corresponds to the way
obiects or their parts are arranged on actual physical surfaces or in an actual physical
space" (1989,p. 61).
Much of the evidence for this principle comes from the scanning studies by Kosslyn
and associates, reviewed above. The general finding is that the amount of time it takes
peopleto scarn from one element of a isual image to another corresponds to the distance
between the elements in a physical representation. Thus, the spatial relationships among
elements of adrawing or object (for example, relative locations, distances, sizes) allseem
to be preserved in the visual image of the drawing or object.
Deparating the visual characteristics from the spatial characteristics of an image (or object
U rawing)is quite difficult. But an ingenious series of studies by Nancy Kerr (1983) has
apparently Succeeded at this task. Hers was a map-scanning study, very similar to that of
Kosslyn et al. (1978) described earlier. Howeve, in this case some of the participants were
Congenitally blind and learned the "map" by feeling objects (each of which had a distinct
Pe) placed on a flat surtace, Once participants had learned the locations, they heard
the experimenter name a pair of objects and were asked tofocus mentally on one and to
imdistance
agine movi ng ja raised dot from that object to the second. Ker found that the greater the
between objects, the longer it took both blind and sighted participants to scan.
 PsychologyIn and Out oftheLaboratory
200 Cognitive
(1978), suggesting
echoed those of
Kosslyn et al.
are similarto
visual that visual
Results of this study The spatial properties
properties.
has spatial because congenitally blind
people-without representations
Vision-apparently
not be visual, images.
make use of spatial
TransformationalEquivalence
way
visual imagery has to do with the that images
tallytransformed. It states
of that "imagined transformations and physical transtormat
Finke's fourth principle
exhibit corresponding dynamic characteristics and are governed by the same \aws

The best(1989,
motion" p. 93).for this principle comes from the studies of mental rotation. Rera
evidence
that the findings from those studies suggest that mental rotation apparently works in t
same way physical rotation does: It is continuous, with rotating objects moving thrOu
final orientation, The time it takes
orientations on their waytotheir
mental rotation depends on how much rotation is to be done, as with physical rotatio,
intermediate sto pefom,
object, and not
with physical rotation of an object, the whole extends just parts of
And, as equivalence beyond
work with mental
transformation
is rotated. The principle of
kinds
however, in assertingthat other objects.
real
of transformations will
images in rotatia,
the same way they work with
Structural Equivalence
the ways that images are ormert.
Finke's fifth principle of visual imagery has to do with
mental images Corresponds to that
and assembled. It states that "the structure of coherent, well organized o
actual perceived objects, in the sense that the structure is
p. 120).
can be reorganized and reinterpreted" (1989,
your artistic skils and incina.
Imagine that you need to draw a picture of an object or (if object. How would vou
tions are as poor as mine) that you need to look carefully at an
do this, and what properties
of the object would influence
the difficulty of your task?
Generally speaking, the larger
the object, the more time it
Would take to look over or to
draw. Also, the more compli
cated the object-that is, the Aannnh

more different parts the object

had-the harder it would be
(and the longer it would take)
to look at carefully or to draw.
Apparently, the construction of (B)
visual images works the same
way. Visual images are formed,
not all at once, but in pieces (A)
that are assembled intoa final
rendition (Finke, 1989). Figure 8.12: Stimuli from the Kosslyn et al. (1983) study.
 (hapter 8: Visual Imagery and
Spatial Cognition
Kos0yn, Reiser, Farah, and Fliegel (1983) studied image
complexityofthe object to be imagined. Participants were generation as it relates to the
that,differed in amount of detail, such as those in asked to form images of pic-
tures
1.3times as long to form an image of the Figure 8.12(A). It took
about
images of outline
drawings. In a
detailed pictures participants
as it did other
pantsto form
such as those shown in Figure 8.12(B) as
related study, the authors used partici-
forms
descriptions. For instance, Figure 8.12(B) could be
stimuli, all of which allowed forgeometric
described either as "five squaresdifferent
shapeof a
cross" or as "two
overlapping rectangles." Participants first read a inthe
corresponding figure, then covered it up and
then
sawthe
et al. found that people given the first
description,
formed a visual image of the
Kosslyn
figure.
than did people given the second
description took longer to form the
image description,
evenbe fasterthe
wasthe same. Notice, by the way, that it would probably though
to draw or look over physical stimulus
conceived of it as two
Figure8.12(B)ifyou rectangles than as five squares. With images,
the
apparently, greater the complexity of the conceived structure of the object, the longer
takesto assemble an image of it.
it
CRITIQUES OF MENTALIMAGERY RESEARCH AND THEORY
In the introductionttothis chapter, I noted that the study of imagery has been controversial
inpsychology, anditis time now to examine the controversy. Although almost everyimag-
erystudy has been subject to some debate (Finke, 1989, provides several examples), we
focus on three general and interrelated themes. The first concerns criticism of imag
eryresearch. In particular, the criticismis that the experiments themselves give enough
"hints." either explicitly or implicitly, for people to perform by relying on their beliefs and
knowledge rather than relying strictly on visual imagery. Asecond critique questions the
metaphor of images as pictures. Athird kind of criticism is more theoretical, questioning
the need to talk about imagery as a distinct kind of internal code. We will consider each
critique in turn.

Tacit Knowledge and Demand Characteristics

Pylyshyn (1981) argued that the results from many imagery studies reflect participants'
underlying and implicit tacit knowledge and beliefs about the task ratther than their con
struction and manipulation of visual images. He paid special attention to image-scanning
experiments. Participants' scanning time is proportional to distance scanned, Pyyshyn
asserted, because they know that the amount of time it takes to physically scan between
tWOpoints in a visual display depends on distance and because they expect the experi
ment to demand this kind of performance.
Finke (1989) explained how this knowledge and expectation could distort results. Imagine
30u want tO move an object (say, your coffee cup) from one location (the right side of your
scenes in western bars)
yO another (the left side of your desk). You could (à la moviesafer to pick up your cup
y0 Side the cup across the desk. but it would probably be
move
placeIt in the new location, Imagine, for the sake of argument, that you cOuld one set of
it from
up instantaneously, regardless of the distance (maybe "teleporting" expected that the
COordinates to the other, àla Star Trek). Suppose, however, you believed or
depend on ne
untof time it will take tomove the coffee cup to the new location should and
total distance from You could adjust your time by pausing
the old to the newlocation.amount down on
holding of time before you placed it
the cup Over the new location for some
202 (ognitive Psychology In and Out of the Laboratory
in moving the cup would be proportional to
the desk. Then your reaction time would depend artbitrarily on the
the cup moved, although the time time
you
may be mentally pausing in
chosethe disiaA
Pylyshyn's(1981) argument was that people
of their beliefs andexpectations about what the image-
experimenterssCyanrig
experiments because
them to do. Tasks that are affected by people's beliefs and expectations ar
Some tasks make it obvious to participants
Pylyshyn to be cognitively penetrable.
instructions, the tasks themselves, or
tasks ought to be performed. The how to behave. Such a task is said to
about the situation cues the person
task "demands"
characteristics(One, 1962). In other words, the psychology
participants in
Som e
have
somehow the th indegm
thata ne
g ls
son behavein a certain way. Typically, experiments pe
and may behave artificialyjust to perform in ways the try to pleaA
they believe willI satisfy

Moreover, sometimes experimenters unconsciously

give subtle Cues to
ex per im erte g
Intons-Peterson (1983) has argued that these experimenter
influenced at least some of the imagery investigations.
menters conduct a number of imagery studies.
Some of the
She had parti
expectancy effects
undergraduat c
eipa
experimenters expeD-
nt,
to believe that the results Would turn out one way; the other
as
to believe the opposite. In all the studies, participants performed the
experimenters werweree lleedd
expected them to.
In one study, Intons-Peterson (1983) used four
undergraduate
expe
experimenters,
all
rimentery
"for their intelligence, dependablity, good judgment, and maturity" (p. 396). None known
familiar with the imagery literature. Each experimenter was assigned to supervise a
18 participants in three different conditions in a mental rotation study. Some total oj
participants
were "primed" by either seeing orimagining a stimulus before each trial; participants in a
control condition received no primes. Two of the four experimenters were told to
that imaginal primes (primes that participants were asked to imagine) Would be expecimore
effective than perceptual primes (primes actually presented to participants). The other t
experimenters were toldthe opposite: Perceptual primes would be more effective tha
imaginal primes. Although all stimuli were presented by microcomputer, and
experimenters were not in the same room with the participants, except initially when the although
read instructions, the results mirrored the experimenters' beliefs. Participants superviser
byexperimenters who believed imaginal primes would be more effective than perceot al
primes produced data to support that belief; participants supervised by the other exper
menters produced data that resulted in the opposite findings.
Intons-Peterson (1983) found similar results in imaginal scanning experiments. She con
cluded that participants in imagery experiments were sensitive to subtle, unintentionl
Cues given by experimenters, including slight differences in intonation or pauses when
reading instructions. Intons-Peterson further argued that imagery research, by virtue of
the subjective nature of the phenomenon under study, may be especially vulnerable to
demand characteristics and experimenter expectations. Although she did not assert that
results from all visual imagery experiments are the result of experimenter effects and
demand characteristics, she did warn that visual imagery researchers must take speciad
care to minimize these effects.

The Picture Metaphor

Much of the discussion so far has suggested an analogy between pictures and image
Some psychologists speak casually of visual images as "mental pictures." The questu
 Chapter &: Visual Imagery and Spatial Cognition 203
doesthe analogy go? As Pylyshyn (1973) pointed out, pictures and
Howseveral ways. Perhaps
far the most important images
difference is that you can
without first knowing what it's a picture of (say, if someone physically
in
ditfer
at
a
picture
photographandVou ask, "What's this?",buyou cannot"Mook" at an image
wordlessly
bokyoua
is. After all, images are internal
hand's youfirst know what it constructionsformed with
somevntention
wnless
in mind. You don't just spontaneously create random images; rather. you
particularthings.
imagesof
om disrupted, and disruptable, in different ways. You can
pictures andimages are
the result that arbitrary parts of the objects depicted disap-
photographin half, with
Second,
more meaningfully, and when they fade, only the meaningful
utaImagesare organized
1989).
Deat disappear r (Finke,
s
pants
seem to be more easily distorted by the viewer's interpretations than are
images
Last, photographs. Remember Bartlett's (1932) work on story recall? (See Chapter
Sor
needto review.) We saw that people's memory for stories changes Over time
pictures

6itoften depends on their initial or subsequent interpretations. So also with images.

you

Carmichael Hogan, and

and Walter(1932) presented participants with patterns such asthose
8.13 with one of
two labels (diferent participants were given different labels).
inFiqure reproductions of the patterns (presumably based on imager) were
Particjpants' later
accordance with the label initially provided, as shown in the figure.
distorteddin
Nickerson and Adams (1979) have shown that people make many errors when
Similarly, their images of familiar objects. Try drawing a picture of a US penny
tyingtoreproduce it to areal penny. Is Lincoln facing the right
looking at one and then compare
motto in the right place? How about the date? Notice in this case that
without
way?ls the right a penny, as
image probably has far less information in it than would aphotograph of with several
your
can see in Photo 8.1. Even when given a multiple-choice recognition test, at selecting
very good
pennies, Nickerson and Adams's participants were not
you
drawings of
the correct representation.
Processes class,
Astudent in my recent Cognitive creative
Rebecca Plotnick (2012), performed a very
(1979) study
study based on the Nickerson and Adams She asked
but using familiar corporate logos as stimuli.
from a set of
undergraduates to choose the real logo
foils, either in context (e.g., on a coffee cup or com
encountered,
puter screen, as they would normally be
as in the examples shown in Figure 8.14 (A) and (B)) or
presented in isolation(see Figure 8.14 (C) or (D). Overall,
the participants were slightly better at identifying the
CoTect choice when the logos were presented in con
eXT (b8% Correct) than when the logos were presented
in isolation (48% correct). IPhoto 8.1: Compare this photograph of a
penny with the sketch you just made.
What's the Significance of finding differences between
images and pictures? Visual images are thought to be information. Although many cognitive
ns of internal coding and representation of a distinc mental code, and although
psychologists believe that visual imagery exists as qualities,the evidence to date sug-
they believe the code has many visual and/or spatial
gests that the visual image-as-picture analogy works only roughly.
204

Word Stimulus Word

list I|

ReprfoidguurCeed
Reproduced figures
list I
figure
Curtains in Diamond in
a window a rectangle

Bottle Stirrup

Crescent Letter C
moon

Hat
Beehive

Eyeglasses Dumbbells

Four
7 Seven
4
Ship's Sun
wheel

Hourglass X Table

BeAn
Kidney Canoe
bean

Pine tree Trowel

Gun Broom

2 Two Eight 8
Figure 8.13: Materials from the Carmichael et al. (1932) study.
 205

b)
Windows Windows'p.
a) b)

FE Windows" WindowsP
d)
d)

a6Ur,

OFFEE Windows Windows

(A) (B)

b)

SUEWAY SUBWAY a) b)

d)
NBC NBC
UEWIAY SUBWA? c) d

e f) NBC NBC
USWAY SUBWAY e

NBC (D) NBC

(C)
ngure 8.14: Corporate loqos from Plotnick (2012) study. Go to www.sagepub.in/galotti_CP5e for colored image.

Propositional Theory
Abroader criticism of work on imagery is theoretical and is aimed at the very premise
images serve
behind the
a$ adistinctfield.mental
Proponents propositional theory reject the idea that
of
code for representing information. Instead, propositional theorists
 PsychologyIn and Out of theLaboratory
206 Cognitive
but propositional in
believe a single code,
neither visual nor verbal
represent all information nature(J.
R. A
&Bower, 1973), is
As we saw in
used
Chapter
to
7,
store and mentally
propositions are a means of specifying relationships
idea that New York
York is a city located to
Anderson
different concepts.
might
(New York,
Boston
For example,
be represented
Boston).
the
by the
Propositions
following propositions:CITY closely
can belinked in networks, with two(New York);theWwest
related betwe n
joined by sharing a number of propositions. WESTOF
(1973) asserted that propositional theory could explain
explain the results of
Pylyshyn information
experiments. His idea was that all
propositions. Participants in visual imagery
is mentally represented and
experiments might look as if
representations, but they they sto im ag
red
Would actually bewere
be using
ery
by
sulting or manipulating internal visual the same kind of representations that cOn-
internal propositional representations, that underie
processingof verbal material, such as sentences
or stories. the
assertion. Kosslyn first
Two studies by Kosslyn (1976) attempted to test this assertion.
association strength between animals andtheir physical attributes. For instance,tested
for
people "claws" are more strongly associated with "cat" than is "head," although most
the
course, have both. Kosslyn found that when people did not use use imagery, they werecats, of
to verify that cats had claws(high association value, small visual part of a cat) than fastetor
verify that cats had heads (low association value, large visual part of aa cat).
theory would predict that the higher the association value. the more propositions Propositiona
the two items and thus the faster the verification time (Finke, 1989). relating
However, when participants reported having used imagery to do the task, their reaction
times went in the opposite direction. Here, they were faster to verify visually larger parte
with low association values than visually smaller parts with higher association valus
Apparently,using imagery results in performance that propositional theory does not predict.
What does it matter in the real world whether people use imagery as a means of coding
information? Understanding how and under what circumstances people mentally reore
sent information is crucial to explaining how they carry out avariety of cognitive tasks. If
they use different codes for different tasks, and if we can make good predictions about
when they use which code, we can perhaps predict when they are likely to be able to do
things easily and when they will have more difficulty with a task.
 PsychologyInandOutoftheLaboratory
Cognitive
the findings from purely cognitive tasks
and that
visually and/or spatialy theories of how
imaginal processing ought to
produced by people's

SPATIAL COGNITION
tacit
tuncion.
The study of visualimagery can be construed as part of a broader picture: spatial
navigate in and through space
tion or how people represent and
store, and use mental representations of (Montello,
spatial
20
is, how do we acquire, to point B? One example of a spatial entity might
themto getfrom point A our environment, presumably
enbeat ies,
and ugp
showingcogite
of sSome part of
map"-a mental depiction among them.
landmarks and spatial
relationships ra
or lying
example, right where you are sitting (or standing,class is down), point in the
For
ofthe buildingin which your cognitive psychology
knowledge of the
stored Opinions
relationship
held. To do
this, presudimarecteba
between your current
andcall
you upon some
thespecified location. vary as to how maplike a cognitive map really i
might note a similarity here to the debate over how picturelike a visual image
maps are mental constructs people use to
case, most agree that cognitiveespecially one that is too large to be
tially through an environment,
(Kitchin,1994).
immediatnavielygpeatercevspeda
|
Barbara Tversky (2005) notes that in the realm of spatial cognition, there are really a
of different kinds of spaces to be distinguished. The way people think
numberdepends
space abod
on which kind of space is under consideration. Each kind of space seems
tohave different attributes and organization.
includes knowing whe
The first kind of space is the space of the body. This space
the different parts of one's body are located at any given moment (such as knowino tht
foot is wrapped aroundt
my right foot currently rests flat against the floor but my left
bottomof mydesk chain); knowing what other objects diferent body parts are interactin
with (my fingers with the keyboard;my derriere with the seat of the chair); as well as feelin
underheated office). Iuse
internal sensations (my stuffy sinuses, the slight chill from my spatialy
my knowledge of the space of the body to direct different parts of my body
something.
Ireachfor something, duck to avoid something, or walk or run toward
ara
The second kind of space is the space around the body. This space refers to the
immediately aroundyou: the room youare in, say, or the region in which you can easily
objects in this
perceive and act on objects. Tversky's work suggests that people localizefront-back axs
space along three axes that are extensions of the bodv. One axis is the
another is the up-down axis, and the third is the left-right axis. Studies by Tversky and
colleagues had people imagine being in a particular space and then locating an imaginay
object in that space. People heard narratives describing them standing, say, In a u
lobby or a museum, with objects on all six sides of their body (front, back, at ne te
at the feet, tothe right, to the left). Next they were asked to imagine direction.,
themseveTimesto
a different direction and then to locate objects while facing this new to
retree
fastest, and times
"retrieve" objects at the head and feet were consistently
objects along the left-right axis were consistently slowest (Tversky, 2005).
and
The space of navigation refers to larger spaces--ones we walk through, explore,
travel to and through. In Tversky's (2005) words:
 Chapter &: Visual lmagery and Spatial
Cognition 209
(onstiuents
ofthespace of navigation include places, which may be buildings or parks or piazzas or
moOuntains.as well as
countries or planetsS or stars, on yet
larger scales. Places are interrelated
directionsin a reference frame. The space of
des or paths or
of s o must be integrated from different pieces of
navigation is too large to perceive
terms
conparable
n place it information that are from
Likethespacearoundthe body, it can be acquired from descriptions andnot immediately
diagrams,
one
fom
a swellas
notablymaps,
fromdirect experience. One remarkable feature of the human mind is the abil
conceiveof spaces
that aretoolarge to be perceived from one place as integral wholes.In order
iyto
eofspacesof navigationas wholes, we need to paste, link, join, superimpose, or otherwise
offinformation. (p. 9)
separatepieces
ntegrate
give directions
to someone, we are dealing with the space of navigation.
When
we
adopta "route" perspective and give those directions in terms of landmarks
we till you come to the gas station, then take aright and go until
Whether two blocks
straight,
(0 barn, then turn left") or a"survey" perspective and give directions in
thered
see
view ("Watson dorm is two blocks east of the chapel and one
you
e
bird's-eye
of Goodsell Observatory"), we communicate some spatial informa-
of
lemssoutheast
bock representations we form within the space of navigation aren't always
Howeverthe
tion.
ACCUrateor
Thus, Tversky (2005) prefers the term cognitive collage to the
complete.
cognitive,map. Cognitive collages are subject to systematic errors and distortions,
term earlier inaccurate beliefs about whether Seattle or Boston is
with people's
We sawW
as north.
farther
were instantiated| in asenior honors thesis conducted by one of my students,
Theseideas
Dara-Abrams(2005). Drew createdla task in which participants (students at Carleton
Drew were shown cutout versions of different campus buildings and asked to place
College) possible. Figure 8.15 shows an actual map of the campus
on a map as accurately as
Pach cutouts correctly placed; Fiqure 8.16 shows the cutout task as it was presented
the
towith
researchparticipants.
maps in the cutout task
showed that participants were likely to make "neater
Results
Dara-Abrams, 2005). That is, they systematically arranged the cutouts along orthogonal
running north-south and east-west
ines. making buildings line up more neatly on lines to rotate buildings so that
than the buildings are actually arranged. They were also likely
orientations-that is, to avoid placing the
they all lined up along vertical or horizontal campus buildings do have a "diagonal"
buildings diagonally, even though several of our (1992) on people's memory
orientation. Such work replicated earlier findings by A. Tversky
for maps. For example, people frequently think of South America as directly south of
America, even though it is really more southeast. The twist in Dara-Abrams's thesis
North participants was learned through their actual
is that the space of navigation used by his
navigation on the campus, not from a map-learning session.
updating (Sargent, Dopkins,
Some recent work in spatial cognition concerns spatialthat as organisms move through
Phlilbeck,& Chichka,2010; Wang et al.. 2006). The idea is
mental representation of where things in the
ace, they must continually revise their location. Zhang, Mou, and McNamara (2011)
imOnment are with respect to their current
give the following illustration:
who is to
you are entering your department's main office and see a colleague offce is
VeE Suppose that the main
chat. After turning, you need to know
L TOu stop and turn left to have a
to the right of you. (p.419)
 the Laboratory
210 Cognitive Psychology In and Out of

Carleton College
6

ARSOF ICE
S*GaOUNIOA
ARB GARAGE
PARR WOUSE
FARM HOUsE

LOWER
ARBORETUM
SOFTMAL
LILAC HILL

WEST FIELD COUNCL

RING
RECREATIOM
CENTER
STEWSIE
ISLAND

CENTER FOR MAI-FETE

THREF
OAKS
MATHEMATICS ISLAND
AND' P
cOMPUTING
BOLIOU HALT
WEST R
LAURENCE0tRRARY
GYMNASIUM GOUL UPPER
FACIUTIES
P ARBORETUM
LAIRD HALL GOODSELL
LEKHTON
HALL
OBSERVATORY S9ODHUE
oUN HALL
UARD SAYLES-HILL OF SCIENCE
CARMPOS MUSK
STADIUM CENTERAP HALL
ANGUAGEAND DIÑNG
THE
SEVERANCE BALD
HALL HULINGS
SPOT HALL
MUDD AALL MYERS S THEC
ECAVE
RWILLs
HALL
OF SCIENCE HALL

BURTON NOURSE
HALL
P BELL
NLSON MUSSER
P FIELD
ALEN BVGSIERHALL
E MALVILE CONCE
E
P
G PRENTICE DAVS WALL SKINNER M DRIAL
CASSAT AnON
MEMORIAL AL HALL CENTER
CHAPEL TENNIS COUNTS
NASON RS
BRODKS 1OHNSON
STIMSO HOUSE BIRD WATSON
OMENS HALL E FACULTY
cOLIER MAT ALUMNI SPERRY CUB ANNEX
CoTL
HUNT GUEST
DOW DACIE HOUSE MEADIE
GEFFERT MENRICKSON MOSE P SECCOMBE
BERC HANEY
CoLWELL FACULTY
MUNTINGTON BENTON CLUB
HUNT STRONG
MIUTON

KONES
RYBERC WHITTIER

IEWETT PAUSH
NUTTING DoUGLAS

CENTRAL PARK

WEITZ CENTER
FOR CREATIVITY
ROUTHSTRE
P ACCESSIBLE AROG
ACCESSIBLE RESTROOM
ACCESSIBLE BUILDING ACCESSIBLE ENTRANCE
ACADEMIC RESIDENTIAL oTHER

Figure 8.15: Actual map of the Carleton College campus.

 Dpduldi kognition 211
Cutout-Arrangement Task
On chefollowingscneenyou will be provided d wich cucOUC pleces hac represe bulldln on the
arletoncampus. When you place the mouse aursor on a utout plece, its name will appear in
e boxloaovd h che lower rh comer of dhe sCreen. Pease arran de cutoUr pleces wMoln
chebhckborder to best represent the respectve locaions of the bulldlngs. lo move a cutout
Ik wich che mouse.To rocace a CUtOUC
anddrop plece, placa thelhboard,
cursor oer Once
dohedrg
plece, cutout piece and press either he left or riht arrow buton on he he you
concer
of dhe.
he complecedarrangngall cutoUt pieces wihin the black
border, press dhe Condnue
bucton,
please ask che eperimener now.
have any quesions,

Continue

Im Finished

Figure 8.16: Cutout task<presentedto participants in Dara-Abrams's2005 study.

 PsychologyIn and Outofthe Laboratory
Cognitive
An important issue that arises is whether representations people Construct as they ar
egocentric (that is, made with reference
to the
OWn position
navigatingthrough space)are
in the space or allocentric(that is, made independently).Important factor
that seem to affect this issue include the size of the space (e.g., a smal indoor roOr
versus a larger outdoor area), the number of distinct and recognizable objects in th
Obsever
updating task.
space, and
Montello the specific
(2005) has argued that navigation consists of two major components: \ocomc
tion(movingthe body over terrain) and wayfinding(planning and making decisions abou
whereto go and howto getthere). Theinvestigation of how people (and animals) navigat
shows the integration of a number of cognitive processes we have previously encour
memory, and knowledge representation, as
including perception, attention, yet to come. wel a
tered, reasoning, decision making)
some topics (planning,

Summary
representations of perceptualexperiences. be There
Visual images are mental
1.
olfactory, cutaneous, and other images,each thought to a
also auditory, experience.
tal representation of a perceptual
states that when information can be coded
2. The dual-coding hypothesis of memory the memorability of that informatin
both by a verbal label and bya visual image, coded only by a verbal label
be
is enhanced relative to information that can
existence of these two distinct codes
3. Not all psychologists believe in the propositional code is
However, despite the theoretical possibility that only one cognitive psycholo
used to perform the visual imagery tasks described, many sort of a distinc
existence of some
gists are persuaded by the evidence of the
visual-spatial code.
4. Research on visual imagery has suggested that
images function in some ways
operations and trans
like internal pictures, undergoing certain kinds of mental appear to operate in
formations. These mental operations and transformations
and transformations.
ways similar to corresponding physical operations
have pointed out limitations in the
5. However, other researchers and theoreticians a number
image-as-picture metaphor. Images work differently than pictures in concluded
have therefore
of ways. Some investigators, Such as Farah (1988), representing information
that "imagery is not visual in the sense of necessarily using
the sense of
acquired through visual sensory channels. Rather, it is visual in
some of the same neural representational machinery as vision" (p.
315).
encoding,
6. Finke (1989) has proposed five principles of visual imagery: (a) implicit equiva
(b) perceptual equivalence, (c) spatial equivalence, (d) transformational
lence,and (e) structural equivalence.
7. Neuropsychological findings, taken in conjunction with older studies, can help
distinguish among diferent proposals. The studies that show activation of the
visual cortex when people form imagery provide convincing evidence that tne
processing of visual images and the processing of visual perceptual information
share a neural substrate.