Content uploaded by Helmut Leder
Author content
All content in this area was uploaded by Helmut Leder on May 15, 2014
Content may be subject to copyright.
1 Introduction
The human mind is equipped to deal with different aspects of the environment that
vary in their demands for specific responses. These demands require corresponding
response mechanisms (eg fight or flight), which have been subjected to adaptive refine-
ment during the course of human evolutionary history. One basic mechanism of the
human visual system responds to low-level stimulus features. Such stimulus features
influence reactions, such as those related to preferences and aesthetic responses (Bar
and Neta 2006; Fechner 1876). For example, it has been shown that people prefer
stimuli that are symmetrical rather than asymmetrical (eg Jacobsen and Ho
«fel 2002;
Tinio and Leder 2009); often, but not always, complex rather than simple (eg Imamoglu
2000, but see Phillips et al 2010); and large rather than small (Silvera et al 2002).
In this study, we investigated one such visual feature: contour
ö
whether curved or
sharp contour influences how objects are perceived and judged aesthetically (Bar and
Neta 2006, 2007; Leder and Carbon 2005). Contour is a defining element in how an
object is seen. In the area of product design, there has been a recent trend towards
designing objects that are generally more rounded, compared with the designs of approx-
imately three decades ago (Carbon 2010). Objects such as cars, furniture, and electronic
devices have become more curved in their appearance. This trend has been paralleled
by recent studies supporting the positive bias in preference towards stimuli with curved
contours. For example, it has been shown that people prefer car interiors with curved rather
than sharp elements (Leder and Carbon 2005). Several other studies (Bar and Neta
2006, 2007; Silvia and Barona 2009) that used various types of stimuli have supported
and expanded our understanding of this preference for curvature.
Bar and Neta (2006, 2007, 2008) suggested that the reason why sharp objects are liked
less is that they appear as potential threats. They further posited that the difference in
preference judgments between curved and sharp objects might stem from increased arousal
in response to sharp objects. Bar and Neta explored this hypothesis using fMRI to examine
Emotional valence modulates the preference for curved
objects
Perception, 2011, volume 40, pages 649 ^ 65 5
Helmut Leder
Faculty of Psychology, University of Vienna, Liebiggasse 5, 1010 Vienna, Austria;
e-mail: helmut.leder@univie.ac.at
Pablo P L Tinio
Division of Education, Queens College of the City University of New York, New York, USA
Moshe Bar
Martinos Center at MGH, Harvard Medical School, Charlestown, MA, USA
Received 8 October 2010, in revised form 15 May 2011
Abstract. Previous studies have shown that people prefer objects with curved contours over objects
with sharp contours. However, those studies used stimuli that were mainly neutral in emotional
valence. We tested here the interplay between visual features and general valence as positive
or negative. After replicating curvature preferences for neutral objects, we used positive (cake,
chocolate) and negative (snake, bomb) stimuli to examine if emotional valence
ö
through response
prioritisation
ö
modulates the preference for curved objects. We found that people indeed preferred
the curved versions of objects to the sharp versions of the same objects, but only if the objects
were neutral or positive in emotional valence. There were no difference in liking for objects with
negative emotional valence. This is evidence that the aesthetic response is adaptive, in this case
prioritising valence over contour as demanded by the general semantic classification.
doi:10.1068/p6845
differential activations in the brain as a function of exposure to curved or sharp
objects. The responses to sharp as compared to round objects indeed were associated
with greater activation in the amygdala (Bar and Neta 2007), an area of the brain linked
to the fear response and general arousal (eg Adolphs 2002; Adolphs et al 1999a, 1999b;
Whalen 1998). This is evidence of a fear-related response elicited by sharp contours.
Thus, psychological studies indicate that people prefer curved contours. Although
fashion may change this preference, it nevertheless seems to be a basic visual primitive
with an evolutionary-based function (Carbon 2010). The stimuli used in past research on
contour preference consisted of basic shapes such as circles (Hevner 1935) and polygons
(Silvia and Barona 2009), but also real-world objects such as furniture and other house-
hold goods (Bar and Neta 2006, 2007). Importantly, those stimuli were characteristically
neutral in emotional valence. However, many objects that we encounter in our everyday
lives are associated with previous experiences
ö
some positive, some negative. Such seman-
tic meanings could impact preferences for objects. This could be the case if an object's
visual features and semantic meaning impact our preference in a hierarchical way.
Because of the influx of environmental cues available at any given moment, different
response mechanisms may be simultaneously activated, and in some circumstances, such
mechanisms might compete for priority (eg Cosmides and Tooby 2000; Tooby and
Cosmides 2005). Thus, if a particular situation demands several simultaneous responses,
it is crucial that priority is placed on producing the response for the most pressing
ö
in terms of survival
ö
demands. Thus, in terms of behavioural responses that are
directed towards survival and other more specific adaptive problems, objects that
are associated with a positive affective value should be approached and those associ-
ated with negative past experiences avoided. However, what happens if two sources
that affect preference
ö
semantic meaning and contour
ö
conflict? Would preference
depend on contour or higher-level semantic affective value?
We report here a study in which the influence of emotional valence on the prefer-
ence for curvature was examined directly by using stimuli that varied in valence. The
assumption that the threat posed by an object results in the object being disliked may
be true. If so, the threatening element of an object could override its positive elements,
such as its curved contours. Therefore, we specifically examined the possibility that
people would only show preference for objects with curved contours if the objects are
associated with semantically non-threatening information
ö
either neutral or positive.
In this case preference for curvature should be present for stimuli with neutral or posi-
tive valence. However, when people encounter objects with negative valence, semantically
threatening information becomes prominent and subsequently overcomes the influence of
contour. This would be the case if the objects with negative valence, perceived as threat-
ening, elicit an aversive response. It would show whether this response then overrides the
effects of visual features such as contours.
This study consisted of two parts. The aim of the first part was to establish the
effects found by Bar and Neta (2006, 2007) that people like curved objects more than
sharp objects in the case of objects with neutral valence. This first part was conducted
to ensure that the effects would be found in a different context
ö
in terms of physical
setting, language (instructions in German), and slight modifications of the procedure
(eg experimental control software and presentation screen). Having established these
effects, we tested the same participants in a second part, which was aimed at directly
assessing response prioritisation in situations where there is competition between the
effect of semantic information (eg snake versus lollipop) and the liking bias elicited
by contour (sharp versus curved). In addition to liking ratings, we also determined
valence ratings to better characterise the emotional valence factor. We also deter-
mined arousal ratings, as there is some evidence that negative emotions generate higher
levels of arousal than positive emotions (Ekman et al 1983).
650 H Leder, P P L Tinio, M Bar
2 Part 1. Preference for curvature
ö
stimuli with neutral valence
2.1 Method
2.1.1 Participants. Thirty-seven psychology students (twenty-four female) from the University
of Vienna participated in the study. Their age ranged from 18 to 29 years with a mean
age of 22.00 years (SD 2:51 years). The nature of the procedures was explained to,
and informed consent was obtained from, each participant prior to data collection.
All participants had normal or corrected-to-normal vision and none was aware of the
purpose of the study.
2.1.2 Materials. Three sets of stimuli with neutral emotional valence used previously
by Bar and Neta (2006, 2007) were employed. The first set included 140 pairs of real
objects, with each pair consisting of a curved- and a sharp-contoured version; that is,
there were two versions of each object (eg stapler): a curved version (stapler with
rounded contours) and a sharp version (stapler with sharp-angled contours). The two
versions of the same object were visually similar in dimensions, but opposite in their
curvature characteristics. The second set of stimuli included 140 pairs of abstract
patterns, each pair consisting of a curved- and a sharp-contoured version, as with the
real objects. The third set was a control set of real objects (eg table, hammer, lantern,
and leaf) that were also neutral in valence, and consisted of 80 objects with approx-
imately equal numbers of curved- and sharp-contoured objects. The full set can be
seen on http://barlab.mgh.harvard.edu/Objects.htm.
2.1.3 Procedure. All stimuli (9.03 cm69.03 cm) were presented in greyscale on a grey
(193,193,193) background using Presentation software (Presentation, Albany, CA, USA;
version 10.3, www.neurobs.com). The general structure was as follows (in order of presenta-
tion): instructions, practice trials, and main trials. Each trial consisted of the following
sequence: a fixation cross for 500 ms; the stimulus for 84 ms; a cue for 1916 ms; and
an inter-trial interval for 2000 ms. The short presentation duration was identical to that
used by Bar and Neta (2006) and was chosen to show, as in their study, that contour
processing occurs during the very early stage of visual representation. In order to become
familiar with the trial structure, participants were given 60 practice trials (10 round real
objects; 10 sharp real objects; 10 round abstract objects; 10 sharp abstract objects; and
20 control objects), which were identical in structure to the main trials. The high number
of practice trials was consistent with Bar and Neta's (2006) procedures. The stimuli
used in the practice trials were not included in the main trials. For the main trials,
each participant viewed one version of each object (either round or sharp, counter-
balanced across subjects) and all the control objects.
Participants provided their responses during the time interval in which the cue
``Like? or Dislike?'' was presented on the screen. Two buttons on the keyboard corre-
sponded to the two response choices. The participants were instructed to provide their
ratings spontaneously. In addition, they were tested individually and the presentation
of the objects was fully randomised.
2.2 Results and discussion
The dependent measure was calculated as the proportion of ``like'' responses to the
total number of responses. Means, sampled over participants, were 0.71 for round objects,
0.68 for sharp objects, 0.29 for round abstract patterns, and 0.21 for sharp abstract
patterns. A repeated-measures analysis of variance (
ANOVA
) with contour (round and
sharp) and object type (real objects and abstract patterns) as within-subjects factors
was performed. In general, there was a main effect of contour (F
136
10:46,p0:01,
Z
2
p
0:225); thus participants preferred curved over sharp stimuli; and a main effect
of object type (F
136
76:43,p50:01,Z
2
p
0:68) with real objects preferred to abstract
patterns. There was also a significant interaction between contour and object type
,
,
Emotional valence and contour preference 651
(F
136
4:28,p50:05,Z
2
p
0:106), indicating that the difference in liking between
round and sharp was greater for abstract patterns than real objects. Directed t-tests
confirmed these higher preferences for the round over sharp real objects (t
36
1:82,
p50:05) and abstract patterns (t
36
3:20,p50:01).
Thus, the results confirmed earlier findings (Bar and Neta 2006, 2007; Silvia and
Barona 2009) that curved objects are liked more than sharp objects. This effect was
observed both for real objects and abstract patterns, with both sets of stimuli possess-
ing inherent neutral valence. Next we examined whether the preference for curved
contours would interact with the valence of an object. Using two sets of stimuli that
varied in their emotional valence, we aimed to directly examine semantic valence as
mediator between competing responses.
3 Part 2. Preference for curvature
ö
stimuli with positive and negative valence
The aim of this study was to directly compare the effects of contour on preference
judgments for objects that had positive and negative emotional valence. In addition to
measuring liking ratings, we also measured valence and arousal ratings.
3.1 Method
3.1.1 Participants. The same group of participants as in part 1 was used.
3.1.2 Materials. A new set of objects (see figure 1 for examples) was produced for part 2.
The visual characteristics (eg size, brightness, contrast, and background colour) of these
stimuli were standardised to match the set used in part 1. However, in contrast to the
previous set, the objects in the new set were chosen because they were objects commonly
known to have strongly positive (eg slice of cake, chocolates) or strongly negative (eg
snake, battle-axe, bomb) emotional valence.
The stimuli consisted of 20 pairs of real objects with positive valence (eg birthday
cake, pralines, sailboat, ice cream, and teddy bear), with each pair consisting of a round
and a sharp version; and 20 pairs of real objects with negative valence (eg razor blade,
sword, toilet brush, spider, dentist's chair), with each pair consisting of a round and a
sharp version. Thus, there were four groups of objects: 10 round positive; 10 round
negative; 10 sharp positive; and 10 sharp negative real objects. The full set can be seen
on http://ppcms.univie.ac.at/uploads/media/Images Perception.pdf
The stimuli were produced by first listing objects that fit into these valence categ-
ories. Next, we searched for pictures that portrayed these objects in isolation (plain
backgrounds) and that depicted them at a normal angle. We then produced two differ-
ent versions of each object
ö
a curved and a sharp version
ö
by carefully digitally
manipulating all possible contour elements. Thus, unlike the stimuli in part 1, the two
versions of each object in part 2 differed only in curvature.
3.1.3 Procedure. The study consisted of three blocks. The first block involved liking
ratings of the stimuli in the same manner as in part 1. The second block involved forced-
choice ``pleasant'' or ``unpleasant'' valence ratings. Finally, the third block consisted of
arousal ratings on a seven-point scale with `1' indicating `calm' and `7' indicating `exciting'.
All participants performed the liking block first. Then, the order of the second
and third blocks
ö
the valence and arousal ratings
ö
was balanced across participants.
This order assured that liking (the primary measure) was always measured at the first
viewing of each object, and thus would be unaffected by the other measures. Moreover,
we assumed that valence or arousal evaluations would be affected negligibly by the
previous tasks.
The general structure of the study and the sequence and timing characteristics of the
stimulus events for the liking and valence ratings were the same as in part 1. To provide
the participants with sufficient time to use the 1^ 7 scale for the arousal ratings, the response
,
652 H Leder, P P L Tinio, M Bar
window was increased to 2916 ms. To make participants familiar with the trial structure,
they were given 12 practice trials (3 round positive; 3 round negative; 3 sharp positive;
and 3 sharp negative objects), which were identical in structure to the main trials. The
number of practice trials (12) in part 2 was much smaller than the number of practice
trials (60) in part 1. This lower number was chosen because it was commensurate
with the smaller number of stimuli in part 2. The stimuli used in the practice trials
were not included in the main trials. During the main trials, each participant viewed
one version of each object, either round or sharp (20 positive and 20 negative objects
for a total of 40 trials) counterbalanced across participants. Consequently, for each
participant, the same set of objects was used for all three blocks.
3.2 Results and discussion
Regarding the preferences (first block) a repeated-measures
ANOVA
was performed
with contour (round and sharp) and valence (positive and negative) as within-subjects
factors and the proportion of like responses to the total number of responses as
dependent variable. Means, sampled over participants, were 0.82 for round positive,
0.78 for sharp positive, 0.25 for round negative, and 0.28 for sharp negative objects.
The analysis revealed a main effect of valence: participants preferred positive over
negative objects (F
136
132:17,p50:01,Z
2
p
0:79). There was no main effect of
contour ( p0:63,Z
2
p
0:007), but there was a significant interaction between contour
and valence (F
136
5:34,p0:027,Z
2
p
0:13). t-Tests revealed that this interaction was
due to a lack of difference between negative round and negative sharp objects ( p0:32);
and a preference for positive round objects over positive sharp objects (t
36
2:83,
p0:01).
Regarding the valence ratings, an
ANOVA
with contour and valence as within-
subjects factors and valence ratings as the dependent variable revealed a significant
main effect of valence (F
136
202:461,p50:001,Z
2
p
0:849). The participants found
the positive stimuli (mean proportion 0.83) more pleasant than the negative stimuli
(mean proportion 0.19), and this was independent of contour. There were no other
,
,
,
0.95
0.85
0.75
0.65
0.55
0.45
0.35
0.25
Proportion of ``like'' responses
round sharp round sharp
positive negative
Stimulus type
Figure 1. Examples of stimuli and results of preference classifications.
Emotional valence and contour preference 653
significant effects. Planned t-tests did not reveal significant differences in valence ratings
between round positive (mean proportion 0.84) and sharp positive (mean propor-
tion 0.82) objects, and between round negative (mean proportion 0.18) and sharp
negative (mean proportion 0:20)objects(p0:09 and p0:67, respectively).
Regarding the arousal, an
ANOVA
with contour and valence as within-subjects
factors and arousal ratings as the dependent variable showed a significant main
effect of valence (F
136
53:36,p50:01,Z
2
p
0:60). Participants found the negative
objects more arousing (mean 4.80) than the positive objects (mean 3.30). There
were no other significant effects. Again, planned t-tests also did not reveal significant
differences in arousal ratings between round positive (mean 3.34) and sharp positive
(mean 3.26) objects, and between round negative (mean 4.86) and sharp negative
(mean 4.74) objects ( p0:337 and p0:326, respectively). These results support
the hypothesis that negative emotions might come along with higher arousal than
positive emotions (Ekman et al 1983). Regarding the interplay of valence and contour,
the findings of part 2 provide clear evidence that contour modulates preferences only
for objects that are positive but not negative in emotional valence.
4 General discussion
First, regarding objects neutral in valence, we found a preference for curved objects,
validating the findings of previous studies (Bar and Neta 2006). Thus, for objects with
neutral valence, preference for curved contours seems to be robust.When we tested objects
that varied clearly according to valence, we found that preference for curved objects was
modulated by emotional valence. Results showed that objects that were neutral and positive
in emotional valence revealed clear effects based on contour characteristics; the curved
versions of objects were liked more than the corresponding sharp versions of the same
objects. However, this contour-based liking bias was not found for objects that were clearly
negative in emotional valence.
These results indicate that the affective evaluation system involves some sort of
prioritisation scheme: when confronted with negative objects, inherent basic visual
features are not considered for the response as highly as the semantically based affec-
tive value of those objects. Consistent with assumptions in theories of aesthetics, the
pleasantness of a visual stimulus is restricted to non-threatening, positive, or at least
neutral objects (see Leder et al 2004). The results also indicate that aesthetic responses
are adaptive to the specific demands of the context. Such demands may involve some-
thing physical, such as the sharpness of an object (Bar and Neta 2006); social, such as
a person's suitability as a romantic partner (Leder et al 2010); or emotional, such as the
level of threat or negative semantic posed by an object, as was found in this study.
It remains to be shown what underlying mechanisms are involved in moderating
the effect. It is possible that positive valence leads to an interaction with objects in a
manner in which basic stimulus features are considered. In contrast, negative valence
might lead to a general avoidance response wherein object features are not considered.
Whether such processes are moderated by levels of consciousness should also be
examined. Moreover, in the present studies, the short presentation durations suggest
that, at least for very early perceptual representations, the two levels of valence
dissociate. In future studies, longer presentation durations would allow the testing of
further hypotheses regarding the time course of these affective evaluations. Future
studies could also use 3-D rendered objects to allow a more systematic examination
of view-dependent and systematically varied levels of curvature. Thus, we have shown
that curved objects are preferred when objects are characteristically neutral or positive
but when objects are inherently negative in emotional valence the attractiveness of
curvature cannot deceive the threatened perceiver.
,
654 H Leder, P P L Tinio, M Bar
Acknowledgments. We would like to thank Bianca Varga for her assistance with data collection.
We also thank Maital Neta for creating some of the stimuli, and Lisa Barrett and Gernot Gerger
and two reviewers for their insightful comments. This research was partially supported by project
FWF P18910 awarded to the first author.
References
Adolphs R, 2002 ``Neural systems for recognizing emotion'' Current Opinion in Neurobiology 12
169 ^ 177
Adolphs R, Russell J A, Tranel D, 1999a ``A role for the human amygdala in recognizing emotional
arousal from unpleasant stimuli'' Psychological Science 10 167^171
Adolphs R, Tranel D, Hamann S, Young A W, Calder A J, Phelps E A, Anderson A, Lee G P,
Damasio A R, 1999b ``Recognition of facial emotion in nine individuals with bilateral amygdala
damage'' Neuropsychologia 37 1111 ^ 1117
Bar M, Neta M, 2006 ``Humans prefer curved visual objects'' Psychological Science 17 645 ^ 648
Bar M, Neta M, 2007 ``Visual elements of subjective preference modulate amygdala activation''
Neuropsychologia 45 2191 ^ 2200
Bar M, Neta M, 2008 ``The proactive brain: Using rudimentary information to make predictive
judgments'' Journal of Consumer Behavior 7319 ^ 330
Carbon C C, 2010 ``The cycle of preference: long-term dynamics of aesthetic appreciation'' Acta
Psychologica 134 233 ^ 244
Cosmides L, Tooby J, 2000 ``Evolutionary psychology and the emotions'', in Handbook of Emotions
2nd edition, Eds M Lewis, J M Haviland-Jones (New York: Guilford) pp 91 ^ 115
Ekman P, Levenson R W, Friesen W V, 1983 ``Autonomic nervous system activity distinguishes
between emotions'' Science 221 1208 ^ 1210
Fechner G T, 1876 Vorschule der A
ësthetik (Leipzig: Breitkopf and Ha
«rtel)
Hevner K, 1935 ``Experimental studies of the affective value of colors and lines'' Journal of Applied
Psychology 19 385 ^ 398
Imamoglu C, 2000 ``Complexity, liking, and familiarity: Architecture and non-architecture Turkish
students'assessments of traditional and modern house facades'' Journal of Experimental Psychology
20 5^16
Jacobsen T, Ho
«fel L, 2002 ``Aesthetic judgments of novel graphic patterns: Analyses of individual
judgments'' Perceptual and Motor Skills 95 755^766
Leder H, Carbon C C, 2005 ``Dimensions in appreciation of car interior design'' Applied Cognitive
Psychology 19 603 ^ 618
Leder H, Belke B, Oeberst A, Augustin D, 2004 ``A model of aesthetic appreciation and aesthetic
judgments'' British Journal of Psychology 95 489 ^ 508
Leder H, Tinio P P L, Fuchs I M, Bohrn I, 2010 ``When attractiveness demands longer looks: The
effects of situation and gender'' Quarterly Journal of Experimental Psychology 63 1858^ 1871
Phillips F, Norman J F, Beers A M, 2010 ``Fechner's aesthetics revisited'' Seeing and Perceiving
23 263 ^ 271
Silvera D H, Josephs R A, Giesler R B, 2002 ``Bigger is better: The influence of physical size
on aesthetic preference judgments'' Journal of Behavioral Decision Making 15 189 ^ 202
Silvia P J, Barona C M, 2009 ``Do people prefer curved objects? Angularity, expertise, and aesthetic
preference'' Empirical Studies in the Arts 27 25 ^ 42
Tinio P P L, Leder H, 2009 ``Just how stable are stable aesthetic features? Symmetry, complexity,
and the jaws of massive familiarization'' Acta Psychologica 130 241 ^ 250
Tooby J, Cosmides L, 2005 ``Conceptual foundations of evolutionary psychology'', in The Handbook
of Evolutionary Psychology Ed. D M Buss (Hoboken, NJ: John Wiley) pp 5^ 67
Whalen P J, 1998 ``Fear, vigilance, and ambiguity: initial neuroimaging studies of the human
amygdala'' Current Directions in Psychological Science 7177 ^ 18 8
ß 2011 a Pion publication
Emotional valence and contour preference 655
Conditions of use. This article may be downloaded from the Perception website for personal research
by members of subscribing organisations. Authors are entitled to distribute their own article (in printed
form or by e-mail) to up to 50 people. This PDF may not be placed on any website (or other online
distribution system) without permission of the publisher.
www.perceptionweb.com
ISSN 0301-006 6 (print) ISSN 1468-423 3 (electronic)
