Sensory-perceptual qualities of images

Journal of Experimental Psychology: Learning. Memory, and Cognition 1989, Vol. 15, No. 2, 188-199

Copyright 1989 by the American Psvchologscal Association, Inc. O278-7393/89/SOO.75

Sensory-Perceptual Qualities of Images Margaret Jean Intons-Peterson and Beverly B. Roskos-Ewoldsen Indiana University Four experiments demonstrated that such sensory-perceptual features of objects as weight, color, and numerosity affect imaginal performance involving images of those objects. For example, imaginary transport times of objects increased with both the hypothetical weight of the imagined object and the distance traversed. The transport functions were steeper when a map of the terrain was imagined than when it was perceived, suggesting that imaginal performance of heft did not parallel more perceptually guided performance. Corresponding to the view that images activate noncanonical information from long-term memory, mental transport times were longer for maps of familiar terrains than for those of presumably unelaborated unfamiliar terrains. Further, the effects of imaginary color discriminations depended on the familiarity of the object being imagined. Images of customarily colored familiar objects were generated faster when projected onto a surface of the same color than when projected onto a surface of another color, whereas images constructed from unfamiliar targets were recognized more accurately when the target's color differed substantially from that of the ground than when it differed by a smaller amount. The results were predicted by a model that assumed that images may incorporate ancillary characteristics in addition to canonical information.

This article focuses on the effects of sensory-perceptual characteristics of an object upon imaginal manipulations of that object. Phrased differently, the article explores sensory aspects of images. As an example, imagine carrying a cannonball for 100 ft. Then imagine carrying a balloon of the same diameter over the same distance. Would it take longer to mentally traverse the distance with the hypothetical cannonball or balloon? Intuition (and the research cited below) argues for the cannonball. Introspectively speaking, images seem to have sensory components. They may be colored; they may be composed of few or many parts; they may seem big, little, loud, soft, pungent, or perfumed. In brief, they seem to evoke various sensory-perceptual characteristics in addition to their canonical' representations and processes. Pursuing this view, we propose that the imagining of an object activates or primes at least some of the network of associates of the object along with its canonical form. Such a view predicts that imaginal representations of a cannonball might include information about its size, weight, color, uses, kinesthetic properties of hefting it, and so forth as well as information about its (canonical) shape and density, and the view predicts that these representations of a cannonball will differ from those associated with an imaginary balloon. The associates activated as an object is imagined engage the subject's world knowledge, with the result that the time to mentally transport a cannonball, an imagined "heavy" object, would be longer than the time to mentally transport a balloon, an imagined "light" object.

This view implies that ratings of the ease of imagining objects should predict ratings of perceived sensory qualities of objects. It also implies the converse: Ratings of sensate qualities should predict rated ease of imagining. This turns out to be accurate, according to some pilot work. Such a correlational approach leaves much to be desired, however, so we turned to experimentation. To test this view, we manipulated sensory-perceptual aspects of imagery in four experiments. In the first two experiments, we tested the times to mentally traverse different distances while carrying hypothetical objects of different weights. The participants used an imaginary or a perceived map of a known terrain to guide their paths. To investigate the effects of familiarity with the mapped area, in Experiment 2 we contrasted similar hypothetical hefting performance when the imagined or perceived map defined a known (the same campus map) or an unknown (afictitiousisland) terrain. We contended that the associations activated by information about familiar places in long-term memory would provide a potential for lengthening mental transport that would be less likely to occur with a map of a novel place. Another sensory attribute, color, was probed in Experiments 3 and 4, but in different ways. In Experiment 3, participants imagined a colored image of a common object against backgrounds of the same, different, or neutral colors. Because the objects and their colors were familiar, the naming of these objects would be expected to activate a network of associations in long-term memory, which would, in turn, contribute to the image. These attributes were expected to affect image-generation times to a greater extent when the background was the same color than when it was a different

We wish to thank Carla Beck, Stephanie Gaston, Sandra Houshmand, Michael McNaught, and Cara Wellman for their very able assistance as experimenters, and Judith Kroll, Ruth Maki, Marilyn Chapnik Smith, and an anonymous reviewer for their trenchant comments. Correspondence concerning this article should be addressed to Margaret Jean Intons-Peterson, Department of Psychology, Indiana University, Bloomington, Indiana 47405.

' The term canonical is used to refer to the central concept associated with an image of an object. For example, when asked to imagine a dog, people report imagining a particular dog rather than a composite of different dogs. The particular dog imagined is the canonical form of dog for that person. 188

SENSORY QUALITIES OF IMAGES

or neutral color. But what would be the direction of these differences? At least two possibilities exist. One is that the perceptual cues from the surface background's color will "absorb" the "color" of the constructed image if the two are the same color, rendering the image difficult to "see." Another is that the perceptual cues will induce the image's color when the two are the same. The first possibility predicts interference and slower image-generation times with image-background color matches; the second predicts facilitation and faster image generation with color matches. Both possibilities draw on the notion that visual imagery and visual perception share similar underlying mechanisms (e.g., Farah, 1985; Finke, 1980; Kosslyn, 1980; Podgorny & Shepard, 1978), although in the second possibility, the influence would be less direct. In this case, perceptual cues may prime, and thereby facilitate, the elicitation of "extraimaginal" color associations. Contradicting this prediction is the literature on imaginal interference (e.g., Brooks, 1968; Segal & Fusella, 1970), which suggests interference from an identically colored background surface. In Experiments 1 and 3 and part of Experiment 2, the tobe-imagined objects were familiar and nameable, two factors that increase the odds of the objects having associational networks. Suppose this were not true. Association-impoverished stimuli would contribute few, if any, cues to the images that might differentiate between the patterns of imaginal and perceptual performance. That is, although imaginal performance might differ from that of perceptual performance, the overall configurations of the patterns would be parallel. To test this prediction, the stimuli used in Experiment 4 were random dot patterns, rendered even more nonsensical by dividing them into two parts. The subjects were instructed to imaginally integrate two successively presented parts of a dot pattern. To manipulate the ease of discrimination, the color (or shade) of the to-be-imaged target was either close to (but discriminable from) or markedly different from the color (or shade) of the background, and it contained from four to six dots. Accuracy was tested by using a four-alternative recognition test. Marked color contrasts should elicit more distinctive associates than near-color matches; hence, we expected faster and more accurate recognition when the target differed substantially from the background than when it resembled the background. Experiment 1 In Experiment 1, all participants learned the locations of five actual campus buildings on a schematic map of the campus. They then estimated the time it would take them to mentally convey three 3-in. diameter objects (a balloon, a ball, and a cannonball) from each building to every other building. The imaginal-map subjects performed this task in the absence of the map, after being told to consult their imaginary map to make the trip. The visual-map subjects performed the task with the map before them. Both groups also participated in a control condition in which they estimated the time required by a specklikeflyto make the various trips. This condition differed from the others because subjects had to imagine an actor (the specklike fly) other than themselves traversing the distances. It was included as a control,

189

however, in the sense that it replicated other imagery-scanning experiments (e.g., Intons-Peterson, 1983; Kosslyn, Ball, & Reiser, 1978). If we assume that the relative distances between the campus buildings remain constant in the subjects' representations of them, then imaginal traversal times would increase with interbuilding distance but not with the weight of the to-betransported object. In other words, the following relation should hold: t = a(f(x, — >j)), in which x and y are locations. The time, t, to mentally move from location x to location y is some function of the distance between the two. The parameter, a, would remain constant for the objects of different weights. A contrasting position is that the constant, a, would be modified by the weight that must be transported over the distance. Thus, for a constant distance, heavier loads would require longer transport times than lighter loads. These comparisons also provide insight into the similarity of performance between imaginal-map and visual-map performance of the task. If the two are virtually identical, the case would be strengthened for ignoring a distinction between imagery and a visual-map counterpart. If the functions differ, the case for separate consideration of the two kinds of processes would be bolstered.

Method Subjects, design, and experimenters. Forty introductory psychology students (20 female, 20 male) were assigned randomly to 10 replications of the imaginal-map and visual-map groups so that equal numbers of each sex were in each group. The experimenters for all three experiments were not informed about the predictions being tested until after the experiments were completed. Before explaining the predictions, the senior author asked each experimenter about the purposes of the experiment. None of their guesses approached the predictions. Procedure. All subjects first learned the locations of five buildings on the campus by examining a 16 x 18-cm schematic map that contained the names of the five campus buildings and some landmark streets. The map was removed after a brief study period. Then, using a map that was blank except for streets and street names, the subjects were to locate the five buildings by placing a labeled dot in the position shown on the previous map. If their placement of a dot deviated from the correct location by more than 2 mm in any direction the map-learning procedure was repeated. This procedure typically required a maximum of two map reproductions. Four practice trials and the 40 regular trials followed, with the schematic map in view for the visual-map subjects only. On all trials, two buildings were named, followed by the word go. When the first building was named, subjects were to mentally look at the named location on their imaginary map (imaginal-map group) or to look at the location on the actual map (visual-map group). When they heard "go", they were to move to the location of the second-named building. When they arrived at this destination they pressed a button. The subjects were tested on four blocks of the 10 combinations of building pairs. The building pairs were randomly ordered, and across the blocks, each building of a pair was used equally often as the first and second location for each subject. The four blocks, also ordered randomly across subjects, tested each of the four weight conditions. At the beginning of each block of 10 pairs of buildings, subjects were told about the conditions of transport. Imaginal-map subjects were told to imagine carrying a 3-oz balloon, a 3-lb ball, or a 304b cannonball, each with a 3-in. diameter. For the

190

M. J. INTONS-PETERSON AND B. B. ROSKOS-EWOLDSEN

fourth (control) condition, they were asked to imagine a specklike fly zipping from one location to the next. Visual-map subjects were told to pretend they were carrying the objects named above or to pretend that a specklike fly zipped from place to place. The subjects were asked to redraw the locations on a blank map to check their memory for the locations, and they were asked about their interpretations of the experiment. These maps were accurately drawn and will not be mentioned further. Finally, after stating their perceptions of the purpose of the experiment, the students were asked to estimate how long it took them to transport the ball over the shortest distance and the longest distance. The same procedure was followed with the cannonball and the balloon.

Results All of the statistics reported herein used the customary alpha level of .05. In no case did the sex of the subject produce a reliable main effect or interaction, so this variable will not be discussed further. Mental transport times increased with distance, as shown in Figure 1, F(9, 324) = 29.1, MS, = 45.28, and this pattern emerged for both the imaginal-map and the visual-map groups. Clearly, standard imaginal-scanning effects occurred

,-A Cannonball

22000 A-'

21000

O

CO

18 000

0)

A

Imaginal Map

20000 19000

A Ball A'

16 000 CD

E en c

Visual Map

A

17000 CO

with the paradigm. Moreover, mental-transport times increased markedly as the hypothetical weight of the object increased, F(3, 108) = 41.32, MSC = 20.67. This result also appeared for each instruction group. These outcomes imply that images may recruit sensory attributes of the object represented by the image. Another prediction, that of similar patterns of transport times for the two instruction groups, was challenged by a significant interaction among group, distance, and weight, F(27, 972) = 32.86, MSC = 11.21. As suggested by Figure 1, mental-transport times increased more rapidly with both distance and hypothetical weight for the imaginalmap group than for the visual-map group. These trends were tested for theirfitsto linear, quadratic, and cubic polynomials, using the same analysis of variance, which treated group and sex as between-subjects variables, and distance (the 10 building pairs) and weight (of the objects) as within-subjects variables. Tests for linear trends showed that the main effect for distances (the slopes) differed reliably from zero, F(\, 36) = 15.87, as did each of the individual functions plotted in Figure 1. In addition, comparisons of the linear trends for the separate functions of Figure 1 showed that mentally transporting the balloon under the imaginal-map condition differed relia-

f"

15000

A

-*~

^ O Balloon

A

Cannonball

14 000 ./

13 000

o

d

12000 A-'

11000

c ioooo c co 9000 o 8000 CO

/

O/

/

A''

/o— o-'

7000

c (0

6000 5 000 Speck

4000 3000

Speck

2000

iooo 0

II

i

II

1

2

t

3

4

5

6

I

!

I

t

Distances

7

I

8

9

1 0 11

t t I

t

1

i

2

3

4

5

6

7

8

9

(mm) between Mapped Buildings

Figure I. Mean scanning times for distances between mapped buildings when an imaginal or a visual map was used (Experiment 1).

10 11

191

SENSORY QUALITIES OF IMAGES

bly from the balloon-visual-map condition, and the same was true for the ball and the cannonball comparisons. In short, these linear trends were not parallel. Finally, neither quadratic nor cubic trends were reliable for any of the weight conditions under the imaginal map or for the speck condition with the visual map, but both trends were significant for the balloon, ball, and cannonball conditions under the visual map. These latter functions are clearly complex, although a linear fit accounts for an average of 89% of the variance. The slopes and intercepts also were compared by calculating the slope and intercept separately for each subject for each weight condition. Separate analyses were conducted on the slopes and intercepts, using map group and sex as betweensubjects variables and weight as a within-subjects variable. As suggested by the means listed in Table 1, both the slopes and the intercepts increased with the weight of the hypothetical objects, F(3, 108) = 30.51, MS, = .28, and F(3, 108) = 16.70, MS, = 7.18, respectively, but these results were qualified by reliable interactions between the map used and the hypothetical weights, Fslopcs(3, 108) = 6.18, MSe = .28, and FinIert.ep,s(3, 108) = 3.98, MS, = 7.18. Tukey tests were used to compare the slopes and intercepts for the Map-Group x Weight interaction. The results of these tests verified the appearance of Figure 1: For the imaginal-map group, the slopes of the functions for balloon, ball, and cannonball did not differ from one another, although they were reliably steeper than the slopes for the speck-fly condition or for any of the hypothetical weight conditions for the visual-map group. For the visualmap group, the slope for cannonball was significantly steeper than the slopes for the other functions, which did not differ. The major difference was in the intercepts. For the imaginalmap group, the intercept for cannonball exceeded those for ball and balloon, which did not differ, and all three exceeded the intercept for the speck-fly. For the visual-map group, cannonball had a reliably higher intercept than speck-fly; no other differences were significant. Again, the qualification reflected the absence of differences for the speck-fly condition and the generally steeper slopes and higher intercepts for the other hypothetical weights when an imaginal, rather than a visual, map was consulted. The differences between the two map groups might have been induced by demand characteristics. As noted in the Method section, the experimenters did not ascertain the purposes. They, like the participants, expected travel times to Table 1 Experiment 1: Slopes and Intercepts for Mental Transport Times for the Imaginal-Map and Visual-Map Groups Group Imagery map Speck-fly Balloon Ball Cannonball Visual map Speck-fly Balloon Ball Cannonball

Slope

Intercept

.31 1.11 1.22 1.47

1.00 3.67 4.49 6.70

.21 .77 .80 1.25

1.09 2.64 2.87 3.78

increase as both distance and the hypothetical weight of the objects increased. Neither the experimenters nor the participants, however, predicted differences between the two types of maps. The imaginal-map group predicted that it would take 3.01, 3.75, and 4.10 min to transport the balloon, ball, and cannonball, respectively, over the shortest distance, and 7.03, 7.39, and 8.14 min to cover the longest distance. The visual-map group's predictions were, in order of balloon, ball, and cannonball, 2.77, 3.81, and 4.15 min for the shortest distance, and 7.31, 7.49, and 8.09 min for the longest one. The two groups did not differ reliably (F < 1). Thus, the groups did not predict the difference between the two maps, nor did they predict the magnitude of the effects of the different hypothesized weights. It does not appear, therefore, that the results can be explained on the basis of demand characteristics.

Discussion These results indicate that standard imaginal-scanning functions are found for mental transport. This was true for both groups, although the slopes were steeper and had higher intercepts for imaginal-map than for visual-map groups. This "expansion" effect suggests that at least one sensory attribute, the hypothetical weight of an object to be mentally transported from one location to another, affects imaginal performance to a greater extent than performance aided by a physically present map. The data thus favor the view that images are intimately related to knowledge about sensory-perceptual aspects of representations retrieved from long-term memory and disfavor the view that images, as cognitively impenetrable entities, are immune to sensory influence. The data also seem to contradict two other notions. The first is that the differences between the two groups were due solely to the effort of holding an image expended by the imaginal-map group, because such an effort should have affected the speck condition as well. Moreover, such differential effort would be expected to affect the intercepts, but not the slopes, of the functions. The other contradicted notion is that images and perceptions are represented and processed similarly. If they were, the hypothetical weights of the mentally transported objects would have the same effects for both groups. The data do not eliminate the possibility that the generally improverished two-dimensional visual map inhibited the use of richer associations with campus buildings that might have been present in the imaginal-map conditions. Nor did the data distinguish between mental traversal with and without imaginally carrying an object. Consequently, Experiment 2 was conducted to determine whether the use of a novel map, one presumably devoid of embellishing associations, would yield the same results as the use of a map of familiar territory. Another goal of the experiment was to obtain mental walk times when no objects were to be mentally transported. Experiment 2 Two potential explanations for the results of Experiment 1 were investigated in Experiment 2. The first is that Experi-

192

M. J. INTONS-PETERSON AND B. B. ROSKOS-EWOLDSEN

ment l's results occurred because compared with the use of the visual map, the use of the imaginal map permitted mental embellishments of the images that increased processing time. This possibility was tested by trying to eliminate the associations likely to be generated along with the images. The participants learned and were tested on location pairs in novel maps. In this approach, the mental representations for the locations should be highly similar for the imaginal- and visualmap conditions. If the images of Experiment 1 distracted the imaginal-map group more than the visual-map group, performance should be the same for the two map groups on novel locations. A second explanation is that the visual map inhibited reference to the actual buildings. The notion is that the two-dimensional portrayal of buildings might have speeded the visual-map group's responses because buildings typically are perceived as three-dimensional. To try to approximate the imaginal-map group's flexible images in the visual-map group's representations, pictures of familiar campus buildings either were or were not provided along with the map. This explanation predicts that mental-transport times will be identical for the two map groups when pictures appear on the study maps but will differ, with the imaginal-map times longer, when no pictures are on the study maps. Additionally, all participants were asked to rate on a 100point scale the effort of carrying a balloon, a ball, and a cannonball for 100 yards. These ratings were obtained both before and after the map-learning and mental-traversal parts of the experiment. The rationale was that if imaginal transport invokes sensory aspects, the effort might appear greater after all the "labor" for the imaginal-map group than for the visualmap group, which would be less likely to associate weight with the mental activity.

no object was carried. Other counterbalancing details replicated those of Experiment 1. Imaginal maps were used by half of the subjects, and perceived maps were used by the other subjects. This testing ended the first part of the experiment. In the second part, the participants learned the campus map used in Experiment 1, except that half of the imaginal-map group and half of the visual-map group had pictures of campus buildings affixed to their maps. The other halves of the two groups saw maps with no pictures, thus duplicating the map conditions of Experiment 1. Again, counterbalancing and testing procedures were the same as with the island map. After testing on the location pairs for the balloon, ball, cannonball, and no-object conditions, the subjects were asked again to rate the effort required to transport each of the objects 100 yards. They used the same 100point scale. They then were asked to state the purposes of the experiment and to estimate how long it would take them to carry each of the objects over the shortest distance and over the longest distance. The novel-island-map conditions always preceded tests of the campus map to permit comparisons of the novel-map procedure with campus-map performance (Experiment 1) when neither had been preceded by tests of other maps. Moreover, campus-map performance could be examined as a function of prior exposure to no previous map (Experiment 1) or to another map (Experiment 2).

Results

As expected, mental-transport times increased with the distance supposedly traversed, F(3, 108) = 20.7, MSC = .253, and with the hypothetical weight of the object transported, F(3, 108) = 19.67, MSC = .156, (Figure 2). Moreover, these variables interacted, F(9, 324) = 7.04, MS, = .228, with transport times increasing more rapidly with distance for the supposedly heavier objects. More important for our predictions, when the presumed ancillary associations were minimized by using a novel map, the imaginal- and visual-map groups did not differ reliably. The respective means were 11.63 s and 12.07 s (F < 1). Method However, when the maps depicted familiar territory, with its numerous associations, the mean mental-transport time of Subjects and design. Forty new students from the same source the imaginal-map group (16.73 s) reliably exceeded that of were randomly assigned to four groups, which represented the map the visual-map group (15.46 s) when no pictures appeared on used (imaginal or perceived) and the presence or absence of pictures during study of the familiar maps. the study map but not when pictures did appear (imaginalMaterials. Two new maps were prepared. One modified Experimap group = 13.98 s; visual map group = 13.31 s). The ment l's map by mounting small colored pictures of the buildings influence of picture presence on mental-transport times for immediately above the dot pinpointing the building and the building's the imaginal- and visual-map groups for familiar maps sugname. The other new map, which represented a novel terrain, congests that the study-map pictures may have constrained the tained five locations—B, F, H, J, and M—placed on the map to imaginal-map group's images to be more like the visual duplicate the relative locations of the buildings used in Experiment group's representations, facilitating response times for both 1. The participants were told that this was the map of a newly groups. discovered island. It was so new that its unusual locations had not been properly named. In the interim, they were identified as B, F, Exposure to a previous map systematically increased reand so on. sponse times for the longest distances on the familiar map by Procedure. The participants first learned the novel (island) map about 1 s. The effect was somewhat greater for the visual-map and then received three practice trials, one using a balloon, another groups than for the imaginal-map groups, as can be seen by a ball, and the third a cannonball. They then wrote down how much comparisons of the plots for familiar maps in Figures 1 and effort it would take to carry a 3-oz balloon 100 yards, using a scale 2. that ranged from extremely easy, almost no effort at all (1) to Unsurprisingly, it took about the same time to mentally extremely difficult, may not be able to do it (100). The same procedure walk the distance without holding anything as it did to walk was repeated for a 3-lb ball and for a 30-lb cannonball. Subjects were the distance holding the 3-oz balloon. Also unsurprisingly, then reminded about the instructions and tested using two orders of the times for mentally transporting the cannonball always four location pairs for four blocked-weight conditions: balloon, ball, cannonball, and no object. In comparison with the speck-fly condition exceeded the other times. This was true for the 3-lb ball, as of Experiment 1, the absence of an object was used in Experiment 2 well, except for the case in which an imaginal map was used as a more effective way to assess the estimated transport time when to traverse a familiar area and no pictures were used during

193

SENSORY QUALITIES OF IMAGES 34000 • 31000 &

30000

Cannoncall

Imaginal Map - Familiar 1*000

Visual Map-Familiar

(campus)

(campus) . A Cannonttai

MOOO 24000

A""""

•

21000

A

Ball

o

Balloon

aoooo

>

1(000

tZ~**

1(000

A'

Wilk

» / x/p

Ball

« W.Ik " Balloon

14000 11000

o

CD CO

/ //£''

10000 (000 6000 4000

CO

1000

E

0

>

1

'

i

a

3

4

•

a

7

•. t 10

1 2

3

4

•

6

7 8

B 10 .11

14000 -

O5 C

'c c co o

31000 • 30000 -

Imaginal Map - Unfamiliar (Novel Island)

28000 -

Visual Map-Unfamiliar (Novel Island)

21000 -

Cannonoau

CannonbaH

14000 -

c

12000 -

CO

20000 -

0)

taooo -

.Bail

wooo / s

14000 •

a BaMoon .,7* Walk

12000 10 000 sooo. 8000 4000 2000 ol— 1

.2

3

'

'

•

I

—

4

9

8

?

8.

Distances

i

9

10

11

1

2

3

4

3

0

7

8

i

9

i

10

11

(mm) between Mapped Locations

Figure 2. Mean scanning times for distances between locations when an imaginal or a visual map was used (Experiment 2).

study. In this case, the mental transport times for the 3-lb ball were higher than the times for the balloon and lower than the times for the 30-lb cannonball. The estimates of the effort required to actually carry each object 100 yards showed increases with weight, F{2, 72) = 167.26, MSC = 522.76, but they did not show increases with

any other variable or interaction. That the differences in mental transport times did not mirror estimates explicitly derived from general world knowledge also argues against a demand-characteristic interpretation. We also asked the subjects to indicate how long they thought it would take them to actually walk the shortest or

194

M. J. INTONS-PETERSON AND B. B. ROSKOS-EWOLDSEN

the longest distance carrying no object or each of the three weighted objects. These time estimates provide a measure of the extent to which their general world knowledge correlated with their mental-transport times. Although the estimated times increased reliably over distance, F( 1, 36)= 175.51, MSC = 56.12, and weight, F(3, 108) = 36.50, MS, = 38.39, and with their interaction, F(3, 108) = 48.96, MSC = 4.46, neither the map group nor the picture manipulation yielded reliable main effects or interactions (Fs < 1). More important, the estimated walk times were considerably flatter than the mental-transport times, replicating Experiment l's results. In addition, the experimenters were unable to identify the predictions of the experiment. Hence, we do not think that the results can be explained as the product of demand characteristics of one kind or another.

Discussion We must ask whether the results of the familiar map in Experiment 2 are contaminated by order effects. To check this possibility, performance on the novel map was compared with performance on the familiar map in Experiment 1, thus mimicking analyses of Experiment 2. The outcomes did not change. Experiment 2 was designed to investigate two explanations for the results of Experiment 1: The slower mental-transport times of the imaginal- as compared with the visual-map groups were due to (a) time-delaying embellishments of the imaginal group's images and (b) time-speeding inhibitions of building references by the visual-map group. Both processes were implicated by the results of Experiment 2. When representation embellishments were discouraged by the use of a novel map, the two map groups revealed similar performance. When the likely building associations of the two map groups were rendered similar by showing pictures of the buildings on the study map, the two groups performanced at about the same level. But when the associations were unconstrained because no pictures were shown on the study map for the familiar locations, the imaginal-map group showed longer transport times, just as they did in Experiment 1. Again, imaginal associations seem important. We now consider the influence of another sensory characteristic, color, on image-generation times when images of named, colored objects are to be constructed on a background of the same, different, or neutral color. In this situation, as in Experiments 1 and 2, naming a familiar object should activate a network of associations, which, in turn, may affect imaginal performance. Experiment 3 The objects to be imaged, all highly familiar, appear in nature in at least two colors, such as red or green apples. These objects were to be imagined after the subjects fixated on a background of (a) the same color as the object (imagining a red apple against a red background), (b) a color different from the color of the object (imagining a red apple against a green background), or (c) a neutral shade (imaging a red apple against a black background).

Our model holds that naming the color of the to-be-imagined object will activate color associations and increase the likelihood that these associations are represented in working memory. These color associations of the image interact with the color cues of the background, leading to the expectation that the times to generate the images will differ when the background is the same color, a different color, or a color neutral to that associated with the image. But what will the direction be? One possibility is that the background color will "absorb" or mask an image of the same color so that the image's outlines will be difficult to ascertain. This view was favored by all the members of a separate 36-person prediction group. They predicted that generation times would be longer when the color of the background and that of the image were the same rather than different. Another possibility is that when the background duplicates the object's color, it activates or primes color associations with the to-be-imaged object. When the background differs from the imaginary object's color, no such activation would occur. This view suggests that when the background matches the color of the to-be-imaged object, image-generation times will be faster than when the background color is different. A third possibility is that the color of the background should have no effect. Experiment 3 was designed to test these predictions.

Method Subjects and design. Twelve female and 12 male students from introductory psychology classes were assigned to one of two lists of materials. These lists had been constructed to counterbalance the items, as explained below. Students participated in only one of these experiments. Materials. The objects to be imagined were eight items that appear in the environment in at least two colors: red (green) apple, white (pink) dogwood tree, blue (yellow) shirt, tree with green (brown) leaves, gray (blue) sky, pink (red) rose, yellow (white) corn, brown (gray) jacket. For some of the objects, one color may be more frequent than the other, but because one color of each item was imagined against all three types of backgrounds, any differences in frequency would not bias the results. All subjects imagined all eight of the objects, but they were randomly assigned to one of the colors. For example, the first list corresponded to the color named first above (red apple, white dogwood tree, etc.), and the second list contained the other color (green apple, pink dogwood tree, etc.). Three items were used as practice stimuli for all subjects: purple mums (the experiment was conducted in the fall when chrysanthemums were blooming profusely all over campus), aqua fish, and tan armchair. Procedure. The basic plan was to ask the subject to imagine an object after fixating on a television screen that was the same color as the object (e.g., a red apple against a red screen), the alternative color of the object (e.g., a red apple against a green screen), or a neutral (black) screen (e.g., a red apple against a black screen). This plan required subjects to imagine each of their assigned objects three times (once against each background). The order of these 24 trials was randomized, with the constraint that no object be imagined more than twice in succession. An experimental session began with general instructions. The procedure was illustrated by showing three examples of the relations between the colors on the screen and the imaginary object. Thus, subjects imagined purple mums against an orange background, an aqua fish against an aqua background, and a tan armchair against a

195

SENSORY QUALITIES OF IMAGES black background. They were instructed to look at the screen at all times and to begin to imagine the object as soon as the words describing the object appeared on the television screen (printed in white at the bottom of the screen). The words remained in view until the subject responded. Moreover, subjects were cautioned that the screen would be different colors at different times but that they should concentrate on their image and ignore the screen. As soon as their image was "as clear as it can be," they were to press a button marked "image." They then rated the brightness of their image by pressing the appropriate button within an array labeled very dim, dim, neutral, bright, and very bright. Finally, the subjects were given a sheet that listed all 16 critical pairs. The subjects' task was to indicate the objects that they had been able to imagine by writing an / next to any for which they had formed an image. The subjects were questioned about their interpretations of the experiment, debriefed, thanked, and dismissed.

Results Image-generation times. To avoid primacy effects and idiosyncratic differences that might appear during an image's initial generation, we focused on the second and third generations of each image. Image-generation times strongly implicated sensory-perceptual factors. The fastest generation of the second and third attempts occurred when the background was the same color as the intended image (2.11 s), next fastest when the background was black (2.37 s), and slowest with an alternative coloring of the background (2.65 s), F(2, 40) = 3.39, MS, = .597. Similar results were obtained from an analysis of the initial trials with each object. Brightness ratings were highly correlated with generation times. These results directly contradicted the expectations of the prediction group. These intriguing results led us to question whether objects imagined against their complementary color would require particularly long generation times, as might be expected if the imaginal process evokes sensorylike processing. We asked, for example, whether a red apple imagined against a green background or a blue shirt imagined against a yellow one would take longer to generate than yellow corn against a white background or a white dogwood tree against a pink one. The answer was negative, and the means were slightly, but nonsignificantly, in the opposite direction: The mean image-generation times for complementary and noncomplementary backgrounds were 2.12 s and 2.28 s, respectively, F(l, 20) = 1.54, ns. Imageability. Again, most of the subjects reported that they had been able to imagine most of the objects, and no object seemed to be easier or harder to imagine than others.

Discussion The results of Experiment 3 demonstrated once again that sensory-perceptual factors play an important role in imagery. Images were easier to generate against a background of the same color as the object imagined, a result not expected by an independent prediction group. Similar color backgrounds may facilitate image construction by inducing the color. Apparently, retrieval of information about an object from mem-

ory also activates associated cues that may affect the image's generation. Images may convey knowledge, tacit or otherwise. They do not, however, always parallel perceptual processes, as shown in Experiments 1 and 2, and therefore are not reducible to only perceptual representations and processes. Nevertheless, under certain conditions, imagery and perception yield highly similar patterns (e.g., Finke & Shepard, 1986; Intons-Peterson, 1980, 1981; Podgorny & Shepard, 1978). The challenge, obviously, is to identify and predict when the two will be similar and when they will vary. In Experiments 1 and 3, the to-be-imagined objects were highly familiar and were expected to carry surplus information. If the to-be-imaged objects were unfamiliar, as in Experiments 2 and 4, minimal, if any, extra information would be activated. Hence, image generation would depend on roughly the same visual inputs as perception, and the two situations should yield parallel results. Experiment 4 In Experiment 4, we consider the influence of three features on the imaginal integration of two parts of a dot pattern: the ease of discriminating between two colors or shades of dots that comprise the figures, whether thefigureswere constructed from differentiable colors or shades of gray, and the number of dots in the figure. After seeing and then imagining two successively presented parts of a dot pattern, participants were to mentally superimpose the parts and to identify the resultant figure from among four alternatives. To manipulate the sensory components, we varied the contrast between the figure and ground, using Munsell color or gray chips. For one level, the figure-ground discrimination was easy, and for the other level, it was difficult but still discriminable, according to pilot work. We also varied the number of dots comprising the figure from four to six. If sensory-perceptual properties of the components contributing to an image are likely to be associated with the image, both the ease of the figure-ground color discrimination and the number of dots comprising the figure should influence the accuracy of identification of the imaginally integrated pattern from among reasonable alternatives. To make more precise predictions, we incorporated an earlier model (Intons-Peterson, 1981, 1984). This model posits that when a figure is novel or unusual, as was the case in Experiment 4, comparisons of the image with actually presented alternatives will proceed by serially checking each component of a test alternative with the comparable component of the mental image. As a result, this procedure predicts that processing times will increase with the number of novel components. To apply to Experiment 4, the model must be extended to cover accuracy. This is easy to do, for the likelihood of a mismatch increases with the number of potentially deviating dots. Hence, accuracy should decrease as the number of novel components increases. Additionally, accuracy should decrease as visual differences diminish. In other words, if we assume that the representation somewhat incorporates or elicits visuoperceptual information, it should be harder to "see" the dots comprising an imaginally integrated random

196

M. J. INTONS-PETERSON AND B. B. ROSKOS-EWOLDSEN

pattern against a background of dots of a similar color (shade) than against a background of a distinctly different color (shade). Note that in Experiment 3, the imaginally integrated pattern is not assumed to be nameable or to elicit knowledge about uses and other rich associations of objects. Perceptualmemory performance, based on seeing the pattern parts and then trying to integrate them at the time of test, should yield a parallel configuration. Contrariwise, if imagery is not subject to knowledge about sensory-perceptual characteristics, the results should not depend on the difficulty of the color (shade) discrimination or the number of dots.

Method Subjects. Ten female and 10 male subjects were randomly assigned to each of the four between-conditions groups formed by the 2 x 2 design of instruction (imagery, perception) and whether the dots were colored or shades of gray. All participants were introductory psychology students who served as one way to fulfill partial requirements of their classes. Materials. Two decks of 12 three-card stimulus sets were prepared. The first card of each stimulus set showed the left part of the target, the second card showed the right part, and the third card presented the four alternatives, as shown in Figure 3. Four of the stimulus sets had targets composed of four dots, four had five-dot targets, and four had six-dot targets. Two sets at each level of dots were randomly assigned to represent easy discriminations of the pattern dots from the background dots, and two sets were assigned to represent difficult discriminations. All dots were punched from matte Munsell papers. One deck contained patterns constructed from colored dots; the other contained patterns constructed from dots that were shades of gray. The target-background Munsell color combina-

tions for easy discriminations were mauves (5RP 5/4, 5RP 5/6), violets (5P 5/4, 5P 5/6), and light browns (5 YR 5/4, 5 YR 5/6) against greyed aquas (5BG 5/4, 5BG 5/6) and greyed greens (5G 5/4, 5G 5/ 6). The target-background Munsell combinations for hard discriminations were mauves (5RP 5/4, 5RP 5/6) and violets (5P 5/4, 5P 5/ 6) against mauves (5RP 5/4, 5RP 5/6), violets (5P 5/4, 5P 5/6), and light browns (5YR 5/4, 5YR 5/6). In no case was the target and the background from the same color group (e.g., 5RP 5/4 with 5RP 5/ 6), so all targets were discriminable from their surrounds (discriminability was also verified by pilot testing). The target dots were placed randomly, with the restriction that they not define a common, readily named object, as determined by piloting. The third card showed the four alternatives. One alternative was the correct one, of course, and the others were constructed by moving two randomly determined dots one space in a randomly determined direction from the original position. The alternatives appeared simultaneously with numbers below to facilitate identifying the chosen alternative (see Figure 3). The correct alternative appeared equally often in each of the four positions. The same procedure was used to construct the deck of gray dots. The relative ease and difficulty of distinguishing the figures from the grounds of these cards were matched to the counterparts using colored dots. Procedure. For each of three practice sets and the regular sets, subjects saw the first and second cards of each set for 5 s. They then saw the third (test) card of the set, which presented four alternative patterns. Imagery subjects were instructed to look at the pattern on the first card and to signal when their image was "as clear as can be." They then imagined the pattern on the second card, mentally placing it to the right of the first image. The perceptual-memory subjects were simply told to try to remember the left and right pattern parts on the first two cards so as to be able to compare them with the test alternatives. On the test trial, they were to compare the test alternative to a pattern composed of the right part adjacent to the left part. The subjects had as much time as they wished to select their alternative. All subjects were asked to rate the overall vividness of their final images on a 7-point vividness scale. This rating was introduced to the

B

Test Figure 3. Sample materials used in Experiment 4.

197

SENSORY QUALITIES OF IMAGES perception subjects by saying that people occasionally imagine seeing the patterns.

Results As predicted, when the number of dots increased from four to six, the mean correct choices decreased from 1.42 (71%) to 1.05 (52%), F(2, 152) = 7.21, MS, = .227. Similarly, more correct choices were made on easy-discrimination items (M = 1.34, or 67%) than on hard items (M= 1.23, or 61%), F(l, 76) = 9.00, MS, = .169. Thus, the results once again indexed the impact of sensory-perceptual aspects of the stimuli. Moreover, the imagery and perceptual-memory groups showed parallel patterns as functions of the number of dots and the relative ease of the discrimination, although the imagery group recognized significantly more targets (80%) than the perceptual-memory group (49%), F(l, 76) = 29.60, MS, = 1.521. The final variable, the use of colors or shades of gray to represent the figures and grounds, was included to determine whether images were sensitive to differences in brightness as well as hue. They were, in the sense that the patterns described above appeared for both types of stimulus materials. They also showed some differences with respect to instruction, however. Thus, although the main effect for color-versus-gray stimuli was nonsignificant (F < 1), the interaction of this variable with the instruction group was significant, F( 1, 76) = 3.85, MS, = 1.521. Specifically, when the stimuli were colored dots, the advantage of imaginal instructions (76%) over perceptual-memory instructions (56%) was not as marked as it was for gray dots, when the imaginal instruction group correctly identified 84% of the patterns, whereas the perceptual-memory group correctly identified only 42%. These interesting results suggest that imagery may be especially helpful in the more demanding realm of brightness discrimination than of color (hue) discrimination. Postexperimental questioning again indicated that neither the experimenters nor the subjects identified the purposes of the experiment. Additionally, all of the imagery and perceptual-memory subjects described their procedure for comparing the parts with the multiple-choice alternatives in a manner that demonstrated their understanding of the procedure.

Discussion When there was no reason to expect differential preexperimental knowledge to produce a different patterning of imaginal and perceptual-memory performances, none was found, although imagery facilitated performance in general. Both images and memory for percepts were affected by sensory factors, however. Recognition by both imagery and perceptual-memory subjects declined with the difficulty of discriminating the target from its background and with increases in the number of dots comprising the targets. Moreover, the imaginal-over-perceptual-memory advantage was greater for the brightness judgments than for color judgments.

General Discussion The results of the four experiments imply that images are not devoid of associations that customarily attend retrieval of information from long-term memory. To the contrary, images evoke (or are associated with) such sensory-perceptual information as weight, color, brightness, and numerosity. In other words, images are cognitively penetrable. They may be associated with both canonical and noncanonical information, as proposed by the knowledge-weighted imagery model. This acknowledgment poses a dilemma, however, for once we admit the cognitive susceptibility—even cognitive porosity— of images, we become vulnerable to the arguments that the concept of imaginal representation and processing are superfluous and imparsimonious. Instead, a more general representational structure-process, perhaps propositional, should be favored as a powerful, flexible medium that can encompass imagery as another case of humans' ability to reduce the information from their environment into rule-governed, formal arguments between various givens (Anderson, 1978; Chambers & Reisberg, 1985; Kolers, 1983; Pylyshyn, 1973, 1981). Our data contraindicate such a constriction and amalgamation. When subjects estimated the times to mentally transport objects of different weights while relying on their imaginal maps, their times were considerably longer than those of subjects who had a map visible at all times. Thus, imaginal maps and virtually identical visual maps did not deliver duplicate performances. Also contradictory were the results of Experiment 3: Imagining an image on a surface of its complementary color did not produce longer times than imagining the image on a surface with noncomplementary color, as visual characteristics would predict. In Experiment 4, imaginal performance uniformly exceeded perceptual performance, although the overall patterns of responding were similar. In addition, imaginal-brightness discriminations were comparatively easier than imaginal-color discriminations when both were contrasted with perceptual counterparts. Finke (1980) suggested a kinship between imagery and higher (more central) perceptual processes. This view does not help to explain our results, unfortunately, because all of the tasks presumably recruited central processes. Johnson and Raye (1981) offered other hints about possible situations that may or may not induce similar mechanisms. In their model of reality monitoring, they distinguished between internal and external memories, in part by assuming that perceptual cues are more vivid for externally based memories than for internally generated ones. These assumptions appear to contradict our model. This contradiction is illusory, for we are dealing exclusively with what Johnson and Raye called internal memories. Hence, our distinctions could be viewed as a refinement of the reality-monitoring model and as a combination of principles from reality monitoring and elements of commonly espoused imagery models. What are some implications of the results? The first is simply that although imagery and perception may rely on the same underlying mechanisms in some circumstances, they appear to differ in others. Important to this distinction may

198

M. J. INTONS-PETERSON AND B. B. ROSKOS-EWOLDSEN

be the role of knowledge in imagery—the network of associations linked to the to-be-imagined object in long-term memory. Obviously, impoverished or unfamiliar objects will not bring to the imaginal process the same opportunities for enrichment that highly meaningful and familiar objects will. It follows, therefore, that the less meaningful the imagined or perceived material, the more imaginal performance will depend on perceptual input. This seems to be the case because many of the experimenters reporting essentially parallel functions have used such artificial tasks as imagining the relation of two dots to each other as they rotate around the head (Finke & Kosslyn, 1980), detecting the orientation of standard and mirror-reversed capital letters placed in various positions around the diameter of a circle (Shepard & Cooper, 1982), determining whether or not points fell on blocks being filled in in an imaginary matrix (Podgorny & Shepard, 1978), and the dot-integration test of Experiment 4. Less parallelism occurs with more common stimuli and tasks, such as imagining a reversible duck-rabbit figure (Chambers & Reisberg, 1985), faces (Intons-Peterson, 1981, 1984), or the familiar building locations and objects used in Experiments 1 through 3. Some exceptions spring quickly to mind, such as Cooper and Shepard's (1975) identification of left and right hands, although their task may have been sufficiently unusual to account for the discrepancy. Another potential implication or explanation for our results is that they were the product of demand characteristics. Demand characteristics threaten much imagery research. Could they account for our results? It seems unlikely. The experimenters were all naive with respect to the hypotheses being tested, and neither they nor the subjects were able to identify the hypotheses when the experiments were over. These unsurprising results were expected because the manipulations of imagery and other conditions were handled as between-subjects effects. Tacit knowledge about the relation between distance and time may have contributed to the generally upward slope of the functions in Experiments 1 and 2, of course. We think it particularly implausible, however, that the respondents anticipated the specific color-imagery relations of Experiments 3 and 4; indeed, an independent prediction group uniformly mispredicted the outcome of Experiment 3. Also addressed by our results is the issue of the visual and spatial nature of images. Although the prevailing sentiment is that images are amodally spatial, and perhaps not visual in nature (e.g., Baddeley & Liberman, 1980; Carpenter & Eisenberg, 1978; Kerr, 1983; Marmor & Zaback, 1976; Zimler & Keenan, 1983), our results illustrate their visual nature as well. It seems plausible that images retain both nonspatial (e.g., visual) and spatial aspects of their sensory origins. One final comment is in order. We are puzzled by some aspects of Pylyshyn's( 1973, 1981) arguments. When Pylyshyn said that if images are represented separately from propositions, they should not be "alterable in nomologically arbitrary ways by tacit knowledge" (1981, p. 17), does this mean that every time we imagine a particular dog, we must duplicate an initial image of that dog, if indeed an image has been formed? Although the desirability of minimizing the types of represen-

tations and their processes is indisputable, similarly indisputable, to our way of thinking, is the importance of accommodating the variability and diversity demanded by human performance. References Anderson, J. R. (1978). Arguments concerning representations for mental imagery. Psychological Review, 85, 249-277. Baddeley, A. D., & Lieberman, K. (1980). Spatial working memory. In R. Nickerson (Ed.), Attention and performance (Vol. 8, pp. 521— 540). Hillsdale, NJ: Erlbaum. Brooks, L. R. (1968). Spatial and verbal components of the act of recall. Canadian Journal of Psychology, 22, 349-368. Carpenter, P. A., & Eisenberg, P. (1978). Mental rotation and the frame of reference in blind and sighted individuals. Perception & Performance, 23, 117-124. Chambers, D., & Reisberg, D. (1985). Can mental images be ambiguous? Journal ofExperimental Psychology: Human Perception and Performance, 11, 317-328. Cooper, L. A., & Shepard, R. N. (1975). Mental transformations in the identification of left and right hands. Journal of Experimental Psychology: Human Perception and Performance, 1, 48-56. Farah, M. J. (1985). Psychophysical evidence for a shared representation medium for mental images and percepts. Journal of Experimental Psychology: General, 114, 91-103. Finke, R. A. (1980). Levels of equivalence in imagery and perception. Psychological Review, 87, 113-132. Finke, R. A., & Kosslyn, S. M. (1980). Mental imagery acuity in the peripheral visual field. Journal of Experimental Psychology: Human Perception and Performance, 6, 126-139. Finke, R. A., & Shepard, R. N. (1986). Visual functions of mental imagery. In K. R. Boff, L. Kaufman, & J. P. Thomas (Eds.), Handbook of perception and human performance (Vol. 2, pp. 155). New York: Wiley. Intons-Peterson, M. J. (1980). The role of loudness in auditory imagery. Memory & Cognition, 8, 385-393. Intons-Peterson, M. J. (1981). Constructing and using unusual and common images. Journal of Experimental Psychology: Human Learning and Memory, 7, 133-144. Intons-Peterson, M. J. (1983). Imagery paradigms: How vulnerable are they to experimenters' expectations? Journal of Experimental Psychology: Human Perception and Performance, 9, 394-412. Intons-Peterson, M. J. (1984). Faces, rabbits, skunks, and ducks: Imaginal comparisons of similar and dissimilar items. Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 699-715. Johnson, M. K., & Raye, C. L. (1981). Reality monitoring. Psychological Review, 88, 67-85. Kerr, N. H. (1983). The role of vision in "visual imagery" experiments: Evidence from the congenitally blind. Journal of Experimental Psychology: General, 112, 265-277. Kolers, P. A. (1983). Perception and representation. In M. R. Rosenzweig & L. W. Porter (Eds.), Annual review ofpsychology (pp. 129166). Palo Alto, CA: Annual Reviews. Kosslyn, S. M. (1980). Image and mind. Cambridge, MA: Harvard University Press. Kosslyn, S. M., Ball, T. M., & Reiser, B. J. (1978). Visual images preserve metric spatial information: Evidence from studies of image scanning. Journal of Experimental Psychology: Human Perception and Performance, 4, 47-60. Marmor, G. S., & Zaback, L. A. (1976). Mental rotation by the blind: Does mental rotation depend on visual imagery? Journal ofExper-

SENSORY QUALITIES OF IMAGES

199

imental Psychology: Human Perception and Performance, 2, 515— Experimental Psychology, 83, 458-464. Shepard, R. N., & Cooper, L. A. (1982). Mental images and their 521. transformations. Cambridge, MA: MIT Press. Podgorny, P., & Shepard, R. N. (1978). Functional representations Zimler, J., & Keenan, J. M. (1983). Imagery in the congenitally blind: common to visual perception and imagination. Journal of ExperiHow visual are visual images? Journal ofExperimental Psychology: mental Psychology: Human Perception and Performance, ¥, 21—35, Learning, Memory, and Cognition, 9, 269-282. Pylyshyn, Z. W. (1973). What the mind's eye tells the mind's brain: A critique of mental imagery. Psychological Bulletin, 80, 1 -24. Pylyshyn, Z. W. (1981). The imagery debate: Analogue media versus Received November 19, 1987 tacit knowledge. Psychological Review, 88, 16-45. Revision received May 6, 1988 Segal, S. J., & Fusella, V. (1970). Influence of imaged pictures and Accepted May 16, 1988 • sounds on detection of auditory and visual signals. Journal of

Instructions to Authors The editorial policy of the Journal favors publication of integrative articles containing multiple experiments and extensive theoretical development. However, articles reporting single experiments are also appropriate, so long as they report compelling data with a clear theoretical message. Other appropriate forms of submission include brief comments on articles published in the Journal and replies to such comments. In addition, occasional literature reviews advancing our knowledge in the field will also be considered. For further information on content, authors should refer to the editorial in the January 1985 issue of the Journal (Vol. 11, No. 1, pp. 1-2). For information on the other three JEP journals, authors should refer to editorials in those journals. Authors should prepare manuscripts according to the Publication Manual ofthe American Psychological Association (3rd ed.). All manuscripts must include an abstract of 1,000 characters and spaces (which is approximately 100-150 words) typed on a separate sheet of paper. Typing instructions (all copy must be double-spaced) and instructions on the preparation of tables, figures, references, metrics, and abstracts appear in the Manual. Also, all manuscripts are subject to editing for sexist language. APA policy prohibits an author from submitting the same manuscript for concurrent consideration by two or more journals. APA policy also prohibits duplicate publication, that is, publication of a manuscript that has already been published in whole or in substantial part in another journal. Prior and duplicate publication constitutes unethical behavior, and authors have an obligation to consult journal editors if there is any chance or question that the paper might not be suitable for publication in an APA journal. Also, authors of manuscripts submitted to APA journals are expected to have available their raw data throughout the editorial review process and for at least 5 years after the date of publication. Authors will be required to state in writing that they have complied with APA ethical standards in the treatment of their sample, human or animal, or to describe the details of treatment. A copy of the APA Ethical Principles may be obtained from the APA Ethics Office, 1200 17th Street, N.W., Washington, DC 20036. Anonymous reviews are optional, and authors who wish anonymous reviews must specifically request them when submitting their manuscripts. Each copy of a manuscript to be reviewed anonymously should include a separate title page with authors' names and affiliations, and these should not appear anywhere else on the manuscript. Footnotes that identify the authors should be typed on a separate page. Authors should make every effort to see that the manuscript itself contains no clues to their identities. Manuscripts should be submitted in quadruplicate. All copies should be clear, readable, and on paper of good quality. Authors should be keep a copy of the manuscript to guard against loss. Effective in January 1989, the incoming Editor-Elect is receiving all new submissions to the Journal. Submissions that are accepted will be published beginning in the 1990 volume. Mail manuscripts and address all correspondence at all stages of handling to the Editor-Elect, Keith Rayner, Department of Psychology, Tobin Hall, University of Massachusetts, Amherst, Massachusetts 01003. For the other JEP journals, authors should submit manuscripts (also in quadruplicate) to one of the editors at the following addresses: Journal of Experimental Psychology: General, Earl B. Hunt, Department of Psychology NI-25, University of Washington, Seattle, Washington 98195; Journal of Experimental Psychology: Human Perception and Performance, James E. Cutting, Department of Psychology, Uris Hall, Cornell University, Ithaca, New York 14853-7601; Journal of Experimental Psychology: Animal Behavior Processes, Michael Domjan, Department of Psychology, University of Texas, Austin, Texas 78712. When one of the editors believes a manuscript is clearly more appropriate for an alternative journal of the American Psychological Association, the editor may redirect the manuscript with the approval of the author.