ArticlePDF Available

Abstract and Figures

It was recently shown that Austrians associate car front geometry with traits in a way that could be related to face shape geometry mapping to those same overall suites of traits. Yet, possible confounding effects of familiarity with the car models, media coverage and entertainment could not be ruled out. In order to address this, the current study uses a cross-cultural comparison. Adult subjects in two countries (Austria and Ethiopia, n=129) were asked to rate person characteristics of 46 standardized front views of automobiles on various trait scales. These two countries differ substantially with regard to their experience with car models and brands, as well as car marketing and media coverage. Geometric morphometrics was then used to assess the shape information underlying trait attribution. Car shapes for perceived maturity, maleness and dominance were highly similar in both countries, with patterns comparable to shape changes during facial growth in humans: Relative sizes of the forehead and windshield decrease with age/growth, eyes and headlights both become more slit-like, noses and grilles bigger, lips and air-intakes are wider. Austrian participants further attributed various degrees of some interpersonal attitudes and emotions, whereas neither Austrians nor Ethiopians congruently ascribed personalities. Morphological correlates of personal characteristics are discussed, as are person perception and its overgeneralization to inanimate objects. Cross-cultural similarities and differences are addressed, as well as implications for car styling, follow-up studies on driving and pedestrian behavior, and fundamental dimensions in inference from (human) faces.
Content may be subject to copyright.
Original Article
Cars have their own faces: cross-cultural ratings of car shapes
in biological (stereotypical) terms
Sonja Windhager
a,b,
, Fred L. Bookstein
a,c
, Karl Grammer
a
, Elisabeth Oberzaucher
a
,
Hasen Said
d
, Dennis E. Slice
e
, Truls Thorstensen
b
, Katrin Schaefer
a
a
Department of Anthropology, University of Vienna, Austria
b
EFS Unternehmensberatung GmbH, Vienna, Austria
c
Department of Statistics, University of Washington, USA
d
University of Addis Ababa, Museum of the Institute of Ethiopian Studies, Ethiopia
e
Department of Scientific Computing, Florida State University, USA
Initial receipt 20 July 2010; final revision received 15 June 2011
Abstract
It was recently shown that Austrians associate car front geometry with traits in a way that could be related to face shape geometry
mapping to those same overall suites of traits. Yet, possible confounding effects of familiarity with the car models, media coverage and
entertainment could not be ruled out. In order to address this, the current study uses a cross-cultural comparison. Adult subjects in two
countries (Austria and Ethiopia, n=129) were asked to rate person characteristics of 46 standardized front views of automobiles on various
trait scales. These two countries differ substantially with regard to their experience with car models and brands, as well as car marketing and
media coverage. Geometric morphometrics was then used to assess the shape information underlying trait attribution. Car shapes for
perceived maturity, maleness and dominance were highly similar in both countries, with patterns comparable to shape changes during facial
growth in humans: Relative sizes of the forehead and windshield decrease with age/growth, eyes and headlights both become more slit-like,
noses and grilles bigger, lips and air-intakes are wider. Austrian participants further attributed various degrees of some interpersonal attitudes
and emotions, whereas neither Austrians nor Ethiopians congruently ascribed personalities. Morphological correlates of personal
characteristics are discussed, as are person perception and its overgeneralization to inanimate objects. Cross-cultural similarities and
differences are addressed, as well as implications for car styling, follow-up studies on driving and pedestrian behavior, and fundamental
dimensions in inference from (human) faces.
© 2012 Elsevier Inc. All rights reserved.
Keywords: Behavioral anthropology; Geometric morphometrics; Cross-cultural overgeneralization; Facial shape
1. Introduction
Recently, Windhager et al. (2008) showed that people in
Austria attribute person-like traits to cars based on the shape
of cars and their constituent parts in a way that mirrors these
attributions to people and animals. Yet, the authors were not
able to rule out possible influences of familiarity with these
car models, with automobile advertisements, brand stereo-
types, anthropomorphic cars in movies and the like. The
current study extends this work to a cross-cultural perspec-
tive with Ethiopia as a reference country. In Ethiopian rural
areas, participants are not exposed to either the brands or car
models under study or even any kind of car marketing and
advertisement. Quite the opposite holds true for most parts of
Europe, the United States and large industrial centers around
the world.
Human and animal faces convey much essential
information in contexts ranging from predation to social
interaction. Organisms gain much from being right (e.g.,
identification of predator, prey, conspecifics) and lose little
when accidentally treating a nonagent as an agent. As an
example, taking a bear for a stone might be lethal, whereas
the opposite does not harm. This asymmetry in the costs of
errors might have led to a perception bias (error
management theory; Guthrie, 1993; Montepare & Zebrowitz,
1998; Nettle, 2004). Increasing sensitivity to the relevant
Evolution and Human Behavior 33 (2012) 109 120
Corresponding author. Department of Anthropology, Althanstrasse
14, 1090 Wien, Austria. Tel.: +43 1 4277 54713; fax: +43 1 4277 9547.
E-mail address: sonja.windhager@univie.ac.at (S. Windhager).
1090-5138/$ see front matter © 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.evolhumbehav.2011.06.003
features and configurations in the course of evolution might
have resulted in contemporary mechanisms for the interpre-
tation of faces that are activated not only by real faces and
their photographs, but also by abstract representations
(schematic faces: e.g., Keating, Mazur, & Segal,, 1977;
Senior et al., 1999) and that might extend even to car cartoons
(cf. Pittenger, Shaw, & Mark, 1979). This may be a version of
Sperber's (1994) process of extending an actual(evolu-
tionarily salient) domain into a proper(cognitively
conformal) domain.
Applying the rules of person perception to shapes with
the same general structure as a face or body has been
shown by the systematic manipulation of a Volkswagen
(VW) Beetle cartoon (Pittenger et al., 1979). The same
algorithm applied to other objects such as shoes and
armchairs did not lead to comparable trait attributions
(Mark, Shaw, & Pittenger, 1988). Thus, a certain degree of
schema congruity seems to be a prerequisite for the
overgeneralization from faces to objects (Aggarwal &
McGill, 2007). The level of congruity is the extent to
which features of an entity match those of a category
schema, in our case (human or mammal) faces. Both cars
and faces are bilaterally symmetric to a vertical axis with a
visually separated upper part (forehead, windshield), an
ellipsoid on each side of the main body (eyes, headlights)
and two extensions (ears, side-view mirrors), as well as
two features in the midline one above the other (nose/
grille, mouth/additional air-intake).
The analogy between automobiles and facial shape is
already suggested (and depicted) in Coss (2003) and in
Enlow's (1975) textbook on facial growth. Intuitively
assessed similarities confront us every day in car advertise-
ments, news coverage and entertainment media (e.g., Disney
Pixar's Carsor Disney's VW Beetle Herbiemovies)
exploiting our tendency to anthropomorphize nonhuman
agents the more similar they are to human physical
appearance and motion (Morewedge, Preston, & Wegner,
2007; Waytz, Epley, & Cacioppo, 2010).
Given the evolutionary significance of a correct interpre-
tation of another's biological state and intention, it is
apparent that morphological distinctions are the ones that
drive age attribution (Montepare & Zebrowitz, 1998, for a
review) and sex discrimination (Brown & Perrett, 1993), as
well as the attribution of femininity, masculinity (Bruce et
al., 1993) and associated traits such as strength, submissive-
ness and dominance (e.g., Grammer & Thornhill, 1994;
Todorov, Said, Engell, & Oosterhof, 2008). Children have a
relatively larger forehead, thin and arched brows, larger eyes,
shorter noses as well as a smaller mid and lower face
compared to adults (Bulygina, Mitteroecker, & Aiello, 2006;
Enlow & Hans, 1996; Trenouth & Joshi, 2006), and most
adult facial sexual dimorphism involves the same features
that constitute the difference between children and adults
(Schaefer, Mitteroecker, Gunz, Bernhard, & Bookstein,
2004; Weston, Friday, & Liò, 2007). There is recent and
fairly convincing evidence that perceptions of age and
perceptions of masculinity relate to very similar features in
the face (Boothroyd et al., 2005). Todorov and collaborators
(2008) describe dominance alongside trustworthiness as one
of two dimensions in the spontaneous characterization of
faces. The link of social dominance and (perceived) physical
body size has just been empirically confirmed (Marsh, Yu,
Schechter, & Blair, 2009). Williams and colleagues (1999)
found that women relative to men are regarded as soft and
submissive (Ashmore & Tumia, 1980; Hess, Adams, &
Kleck, 2005) in a pancultural study including 25 nations.
Artificially masculinized male and female faces are
perceived to be more dominant, masculine and older (Perrett
et al., 1998). Guthrie (1970) even speculates that the
protruding male chin in the human species was selected as a
sexual signal and thus exceeds the functional necessity of
food processing.
Ecological theory (summarized in Montepare & Zebro-
witz, 1998) provides a useful theoretical framework for the
investigation of accurate perception and behavior as well as
for overgeneralization as an evolutionarily beneficial
strategy. The three relevant tenets are as follows: (1) Social
perceptions serve biologically and socially adaptive func-
tions. Although this favors the expectation of increased
accuracy, it does not preclude errors, especially if there
might be a greater advantage in overdetection (cf. error
management theory; Guthrie, 1993; Montepare & Zebro-
witz, 1998; Nettle, 2004). (2) Social judgments are informed
by perceptible stimulus qualities, and overgeneralization
effects will occur when qualities typical of one context occur
in another. (3) People's physical qualities reveal affordances,
which are the opportunities for acting or interacting, and are
linked to behavioral tendencies.
Pittenger and colleagues (1979) modeled a drawing of a
VW Beetle front and a cartoon version including painted
facial features in line with a mathematical simulation of
growth the cardioidal strain algorithm and asked
undergraduates of an American university for relative age
attributions. The ratings were found to correlate positively
with expectations from the modeling. Windhager et al.
(2008) extended this approach in Austria by the use of
natural-looking, existing car models, a broader list of rated
traits (including sex, interpersonal attitudes, emotions and
personality) and the actual measurement of car shape and its
constituent parts. If there is a biologically based over-
generalization from faces to cars, then the biomorphic
aspects of such a visual stimulus should be broadly
recognized even if the actual object is not. We are saying
that to be recognized as a face may well be a biologically
based affordance of cars. Cross-cultural agreement would
therefore be a prerequisite for supporting theoretical
assumptions such as schema congruity and error manage-
ment theory. Thus, the primary purpose of the present study
was to test whether this shapetrait correspondence
generalizes across cultures (and especially to a culture that
is not exposed to car marketing, has little exposure to the film
industry, etc., such as found in regions of Ethiopia).
110 S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
Generalizability requires confirmation that:
the correlation between Austrian and Ethiopian mean
ratings is high, and
the predicted car shapes for a specific trait attribution
are similar.
From Pittenger et al. (1979) and from the structure of the
first principal component of trait attribution and the
corresponding car shape of the Windhager et al. 2008
study, we predict that (1) at least the biomorphic aspects of
maturity, sex and interpersonal attitudes should generalize
cross-culturally with (2) perceived maturity showing the
strongest signal and (3) that maturity, sex and dominance
attributions have a single factor in common. If there is such a
single factor, car shapes estimated from these three qualities
will be similar for the three scales and the two countries.
2. Materials and methods
2.1. Participants
The current study was based on three data sets (Austria,
prestudy Ethiopia 2008 and comparative sample Ethiopia
2009). Their characteristics are summarized in Table 1. The
Austrian sample was different from the study published in
2008 (Windhager et al., 2008). Data were collected in the
capital of Austria, Vienna, and in rural villages and
surrounding areas 30 to 45 km from the capital of Ethiopia,
Addis Ababa (in 2008: Dukem, Holeta Genet, Laga Tafo
and Sululta; in 2009: Holeta Genet, Sululta and two sites in
Debre Zeit). Participants in Austria were recruited through
online advertisements and in Ethiopia through local
assistants. In Ethiopia, taking both years together, 11 of
89 participants reported having a driver's license and 6
claimed to own a car. The abrupt rise in the amount of
compensation from 2008 to 2009 (Table 1) was necessary
due to high inflation.
2.2. Stimuli
As in the preceding study, all pictures stem from
computer renderings of high-resolution, realistic 3D digital
models. The major advantage of this stimulus source is that
materials, light and positioning can be better standardized
than in any real-life situation (in 3ds Max from Autodesk
Media & Entertainment, San Rafael, CA, USA). The digital
mockups were scaled to the size of the original cars and
colored silver, and the license plates were erased (Fig. 1). For
a realistic appearance, materials such as car paint, chrome,
rubber and glass as well as shadows were added. A virtual
sun was placed at a 45° angle right in front of the car. A
virtual camera with a 200-mm lens was positioned on the
midline in front of the car at 12-m distance and half the
height of the car (illustrated in Fig. 1).
We engaged in such standardization efforts because we
were interested in the response to the shape of the car front
and its constituent parts. All other differences such as
reflections, optical distortions (through different lenses and
distances), slight differences in positioning of the car to the
camera or different colors of the car body might obscure the
signal and hinder the interpretation. We used the stimuli of
the previous study and added eight new ones. A detailed list
of the models (all from 2004 to 2007) from 26 different
brands is given in Appendix A.
2.3. Questionnaire and procedure
Each participant was asked to rate each car on 19 trait
scales before the next car was presented. All 46 cars were
presented in random order for each person. As the human
face is a multisignal system from which humans can
immediately infer information on age, sex, attitudes,
personality traits and emotions, we wanted to cover this
variety of possible inferences when testing trait attributions
to car fronts. All rating scales were continuous and ranged
from 0 to 100 (values and ticks not visible). The first four
were bipolar [the biological features: childadult and male
female; the two main dimensions of interpersonal relation-
ships (Argyle, 2002): friendlyhostile and submissive
dominant]. In order to prevent artifacts from a possible
Table 1
Key data of the three samples
Data collection Participants
Location Date Researcher Women Men Age range Average (A)/
median (M)
age [years]
S.D./quartile
(Q) of age
Profession Compensation
Austria February and
March 2007
S.W. 20 20 2029 years A: 23.5 S.D.: 2.6 years 3/4 student 10 Euro
Ethiopia (prestudy) February
2008
K.S., S.W. 21 21 1945 years A: 28.1
M: 27.5
S.D.: 6.0 years
1st Q: 23 years
3rd Q: 32 years
N1/3 student
or teacher
3050
Ethiopian Birr
Ethiopia
(comparative sample)
April 2009 K.S., S.W. 23 24 1759 years A: 27.0
M: 23.0
S.D.: 10.4 years
1st Q: 20 years
3rd Q: 32 years
N1/3 student
or teacher
80
Ethiopian Birr
111S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
side bias, the direction of the sexscale, with male(0) on
the left and female(100) on the right side, was not in line
with the other scales (in the sense of a social signal where
perceptions of adulthood, maleness and dominance go
together). We reversed this scale prior to data analyses to
present the results more intuitively. The other scales were
unipolar, ranging from not at allto a lot/very much.
These remaining 15 items were (in the order on the
questionnaire): the six so-called basic emotions (Ekman,
1999)sad, angry, afraid, happy, disgusted, surprised; the
five cross-culturally valid personality factors (McCrae &
Costa, 1997)open, extroverted, agreeable, conscientious,
neurotic; two items that resulted from discussions with the
automotive industry contented, arrogant; as well as
arousedto account for the dimensional approach to
emotions by Russell (e.g., Russell, 1997) and, finally, I
like the car. Cross-cultural comparisons were of averages
for each dimension by car and country. Subjects who were
acquainted with the use of a computer mouse worked on a
computer interface (all participants in Austria, 17% in 2008
and 19% in 2009 in Ethiopia); the others responded on a
paper-and-pencil version. The color printouts of the cars, in
the latter case, were presented vertically to match the
presentation of the screen.
In Austria, we used the following instructions (translation
from German): This is a simple rating experiment. Please
describe your impression of the shown car front by moving
the slider to the left or the right. The experimenter stayed
with the subject until the first ratings were completed to
ensure that the task and the interface were understood. In the
Fig. 1. Preparation of the standardized car stimuli. Customary true-to-detail digital mockups of car models from 2004 to 2007 of 26 brands (example in the upper
left panel (A)) were edited in 3ds Max (Autodesk Media & Entertainment, San Rafael, CA, USA) to face a virtual camera (in blue) at half the height of the car
exactly frontally (B and C). A virtual sun (in yellow) produces symmetric reflections and shadows on the material to give the car a realistic appearance. The lower
right picture (D) shows the camera view after rendering. These are the pictures that were presented in the rating experiment.
112 S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
prestudy Ethiopia 2008, we used the following introductory
sentences in English and Amharic, the official language of
Ethiopia: Please tell us: How does this car look? There is no
right or wrong answer, just your opinion counts!As 71% of
participants in 2008 reported Oromo as their mother tongue,
we decided to add translations into this language. We also
felt the need for a more detailed introduction. The revised
version for 2009 was: You will see different car fronts. This
study is not about function of the car or whether you would
like to have it or not, but only about how you think it looks.
Please do not try to be polite, but mark your personal, honest
impression.The items on the 19 scales were presented in
German in Austria, in Amharic and English in Ethiopia 2008
and in Amharic, English and Oromo in Ethiopia 2009. Each
questionnaire was translated back and forth by three
Ethiopian native speakers fluent in at least two Ethiopian
languages, English and/or German. Much personal commu-
nication took place to ensure that not only the items but also
the concepts behind them translate well into the other
languages. The follow-up questionnaires included demo-
graphic questions as well as questions on car ownership, car
use and dreamcars. Finally, our subjects were debriefed,
paid and thanked for their cooperation.
2.4. Excluded and missing data
Each participant was asked to rate 46 car fronts on 18
traits and their degree of liking. From the Austrian sample,
we had to omit data from two men for reasons of obvious
problems with the rating scheme. Furthermore, the ratings of
four cars from another subject were not considered because
all values were defaulted. In the Ethiopian sample of 2008,
we excluded the data of three men and two women using
paper and pencil because their seriousness or comprehension
of the task was questionable from visual inspection (e.g.,
zigzag patterns across the page instead of rating the pictures).
Furthermore, in nine cases, all ratings of one car by a single
subject could not be taken into account due to missing data.
Of the remaining 32,167 data entries (meaning the rating of
one subject on one scale for one car), 173 were missing
(0.5%) and omitted. In the 2009 Ethiopian sample, one
woman left after rating 22 cars. From the remaining 36,252
data points, 327 were missing data (0.9%) because subjects
left the scale of a certain variable blank, made two markings
on the same scale or forgot to rate a specific car, and the like.
Again, such suspect or missing data were omitted from
subsequent analyses.
2.5. Statistical analyses and geometric morphometrics
Pearson correlation coefficients, bivariate linear regres-
sions, plots and significance tests, and descriptive statistics
were computed in SPSS 15. All tests were two-tailed.
Pearson correlation coefficients equal to or larger than 0.3
are statistically significant at the .05 level for our sample size
of 46 automobiles.
A geometric morphometrics approach was used to link
the ratings to car shape. Recent reviews of these methods
can be found in Mitteroecker and Gunz (2009) and, in
connection with perception data, in Schaefer and colleagues
(2009). The great advantage of geometric morphometrics is
that the results of statistical analyses are in terms of
coordinates so that corresponding shapes can be visualized.
The form of each car was captured by the two-dimensional
Cartesian coordinates of 34 landmarks that were described
in detail in the preceding publication (Windhager et al.,
2008). Shape information invariant to scale and position
was extracted by Generalized Procrustes Analysis (Rohlf &
Slice, 1990). The superimposed configurations were then
entered as dependent variables in a multivariate regression
with the averaged rated trait scores as the independent
variable (shape regression). The estimated shape changes
from the average car to a car that is rated as more childlike,
adult, female, male, etc. were visualized through deforma-
tion grids based on the thin-plate spline interpolation
function (Bookstein, 1991). For illustrative purposes,
outline landmarks were connected with straight lines.
Permutation tests based on 1000 permutations were used
as the test statistic (Good, 2000). In terms of computer
programs: landmark coordinates were digitized in tpsDig2
Version 2.12 (Rohlf, 2008), shape regressions and permu-
tation tests were computed in tpsRegr Version 1.36 (Rohlf,
2009), and thin-plate spline deformation grids were
generated using Mathematica 6. All figures were edited in
Adobe Illustrator CS3.
3. Results
3.1. Cross-cultural comparisons: correlations
Magnitudes of Pearson correlation coefficients were used
to assess the strength of association between the Austrian and
Ethiopian mean ratings per car for each trait. The sample size
equals the total number of stimuli, i.e., 46, for all subsequent
analyses. With the Ethiopian prestudy data from 2008, we
obtained a strong positive correlation of 0.75 for the child
adult dimension (see also Windhager et al., 2009), whereas
all other coefficients fell below 0.36.
When the Austrian and the Ethiopian sample of 2009
were considered, we observed strong, significant positive
correlations not only for childadult (r=0.85) but also for
femalemale (r=0.75) and submissivedominant (r=0.84)
ratings. All pvalues were below 2.7×10
9
so that we
abandoned explicit Bonferroni adjustments. The correlation
coefficients for the other scales remained below 0.57
(arrogant, see Fig. 3), followed by 0.43 (afraid) and 0.36
(sad), raising an issue that is considered later. Scatterplots
visualizing the association of Austrian and Ethiopian 2009
mean ratings for childadult, femalemale and submissive
dominant are presented in Fig. 2. Taking the Austrian
subjects as the reference sample, we found coefficients of
determination (R
2
) of 0.72, 0.56 and 0.71, respectively. As
113S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
Fig. 2. Intercultural consistency in the attribution of childadult, femalemale and submissivedominant ratings to car fronts. The three panels show the scatterplots and linear regression lines for the Austrian and the
2009 Ethiopian sample. The regressions were significant at the .001 level, as were the correlations (see main text). The numbers label the single car fronts, which are listed in Appendix A (n=46). This figure not only
shows intercultural consistency, but also hints at the close relation of the three scales, which is illustrated by highly similar relative positions of single cars in all three plots.
114 S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
the labels in Fig. 2, which identify the single car models,
suggest, the intercorrelations between these three dimensions
were high.
Within the Austrian sample, every pairwise comparison
had a correlation coefficient of more than 0.96. In the 2009
comparative Ethiopian sample, the correlation coefficients
ranged from 0.68 (femalemale compared to submissive
dominant) to 0.81 (childadult compared to femalemale).
All pvalues were below 2.2×10
7
.
The correlation coefficients for the other dimensions were
again rather low when the Austrian sample and the Ethiopian
sample of 2009 were considered. Basically, all items could
be allocated to one of two types. Plotting the Austrian car
ratings against the Ethiopians' for each trait separately, one
type was characterized by a relatively small circular scatter.
We use neuroticas an example in Fig. 3. This pattern was
also observed for agreeable, conscientious, content, disgust-
ed, extroverted, sad and surprised.
The second type is where the Austrians utilized a much
broader range than the Ethiopians. This pattern is illustrated
with the item arrogantin Fig. 3 (right panel). All cars were
rated as rather arrogant in Ethiopia, whereas we found
various degrees of attributed arrogance for the same cars in
Austria. A similar pattern was identified for afraid, angry,
aroused, happy, friendlyhostile and open. The average
Ethiopian ratings across all cars for these items were 62.5
(arrogant), 45.6 (afraid), 42.5 (angry), 44.5 (aroused), 66.6
(happy), 36.4 (friendlyhostile) and 65.6 (open) for the items
of this second type of low correlation.
Possible intervening variables such as familiarity with
questionnaires (which was investigated in the follow-up
questionnaire) and the use of a paper-and-pencil version
failed to explain these cross-cultural differences.
3.2. The association of trait attribution and car shape
Shape regressions were calculated to investigate the
relationship of the shape of a car front and its constituent
parts (headlights, windshield, etc.) with the three scales that
showed the greatest cross-cultural consistency: attributed
maturity, sex and dominance. For the assessment of shape
trait correspondence and for reasons of clarity, we used the
Ethiopian ratings of 2009 and the Austrian data. In all six
shape regressions (three items by two countries), none of the
1000 permutations achieved a better (lower Goodall F
values) result than the real data (thus, pb.001) in each case.
To optimize the depiction of the overall shape pattern that
corresponded to different trait attributions, we deformed the
average car to ±2 standard deviations (S.D.) of the
appearance variable for the Austrian ratings and to ±2.5
S.D. of the Ethiopian scores of the 2009 sample. As the
deformations of the average car towards an estimated car
that would be rated as child, female or submissive are
hardly visually distinguishable from each other and between
the two countries, we describe this common pattern of shape
change all at once. With regard to global differences, we
found a vertical stretching in cars of low attributed maturity
and dominance, but high femininity (Figs. 46on the left)
Fig. 3. Two phenomena of low correlation coefficients. Some variables were characterized by a low range of the averaged ratings in both countries and a circular
scatter; i.e., all 46 cars were rated similarly along these specific scales (exemplified with neuroticin the left panel). The unipolar scales are read from zero (not
at all) to 100 (a lot/very much). The second category was described by an elliptical scatter.
115S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
as opposed to a general compression in cars of high
maturity, masculinity and dominance (Figs. 46on the
right) in both countries. Concerning local features, the grids
show a deformation towards a relatively larger windshield
in the direction of low maturity, high femininity and low
dominance. Cars that led to associations of adultness,
maleness and dominance were characterized by a relatively
smaller windshield. As we standardized size during
Fig. 5. Estimated car shapes for increased femininity and masculinity attributions in Austria and Ethiopia. The geometry of the average car was deformed in the
directions of increased perceived femininity (on the left) and increased masculinity (on the right) for both Austrian ratings (upper panels) and Ethiopian ratings of
2009 (lower panels). The major shape changes were in the relative height-to-width ratio of the car, the relative size of the windshield, the shape of the headlights,
the height of the grille and the width of the additional air-intake slots.
Fig. 4. The shapes of car fronts that were rated as childand adultin Austria and Ethiopia. The upper panels visualize the shape regressions onto the Austrian
perception data; the lower ones, those upon the Ethiopian 2009 attributions on the childadult scale (n=46 car fronts). The undeformed grid with quadratic
squares in the middle in this and the next two figures corresponds to the average car shape of the sample. The configurations and deformation grids depict a
highly similar pattern for both countries and correspond to ±2 S.D. for the Austrian and ±2.5 S.D. for the Ethiopian sample from the estimate of the average car.
116 S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
Procrustes superimposition, we can only talk about relative
differences here. The grille becomes relatively wider and
taller with increasing attributed age, maleness and domi-
nance. Also, the headlights differed in shape. The extreme
upper and lower edges of the headlights were close to the
middle of the car in a vehicle that was likely to be rated as
childlike, feminine and submissive, whereas the headlights
were extended laterally and were more slit-like in the
estimated geometry of a vehicle with an adult, male,
dominant appearance. The additional air-intake became
wider and thinner with increasing attributed maturity,
masculinity and dominance. To repeat, the pattern looked
the same for Austrian and Ethiopian ratings (cf. upper and
lower panels in Figs. 46).
4. Discussion
Our approach of combining rating studies with geometric
morphometrics did not make any a priori psychological
assumptions on the similarity of faces (or bodies) and car
fronts. Actually, the results of the analyses (i.e., the thin-plate
spline deformation grids) could have associated any shape
change with the car ratings. Our findings local (e.g.,
angled vs. round headlights, relatively smaller vs. taller and
larger grille, narrow vs. wide additional air-intake) and
global (e.g., proportion of windshield to the rest of the car
body) shape differences according with changing impres-
sions from child to adult, female to male and submissive to
dominant unearthed the striking similarities between car
and face perception. There is much anecdotal evidence on
the parallelism of car fronts and (human) faces, yet whether
car fronts actually activate brain circuits for face processing
or whether our findings stem from a different kind of
generalization such as the (rational) judgment of proportions
will have to be the subject of follow-up studies.
The current study replicated the pattern of trait
attribution to car fronts from Windhager and colleagues
(2008) with another Austrian sample of people in their
twenties. Furthermore, and more importantly, we added a
cross-cultural perspective by collecting data in regions of
Ethiopia, where the car models and brands were unknown
and car marketing does not exist. There, street transporta-
tion is dominated by small trucks, off-road vehicles, some
taxi buses, horses and carts, donkeys and people carrying
their goods.
4.1. Cross-cultural similarities: inference of allometry
Even though street scenery could not be more
different, we found a high cross-cultural consistency in
childadult, femalemale and submissivedominant attri-
butions to cars, with correlation coefficients of 0.75 to
0.85. Thus, these two peoples may have drawn on a
common psychological mechanism. As there is hardly
any no car marketing in Ethiopia, our results likely are
due to the properties of the car. The estimated shape
patterns, which hardly vary at all (Figs. 46), reflect one
underlying dimension of facial proportion. We cannot
completely rule out carryover effects from the order of
Fig. 6. Visualization of shape regression onto the submissivedominant scale. The left panels show the predicted shapes for car fronts that are perceived as
submissive (abbreviated subm.) in Austria and Ethiopia. The right ones are the estimates that would elicit a dominant impression (n=46 cars). The upper panels
are based on the Austrian ratings; the lower ones, on the Ethiopian judgments of 2009. The car in the middle is the shape of the average car in the study and served
as the template for the thin-plate spline deformation grids.
117S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
items, but we tried to minimize them by counterbalancing
the sex stereotypes in the directions of our rating scales.
Moreover, friendlyhostile as the third scale on the
questionnaire (between malefemale and submissive
dominant) did not reflect the observed patterns for the
other three items. Therefore, we believe the biomorphic
aspects of the stimuli to be a likelier explanation of our
results than potential carryover effects. The most likely
candidate for this single dimension is allometry (i.e., the
change of facial shape along with proportions during
growth) and the corresponding social cues we reviewed
briefly in the introduction. A relatively large forehead
compared to a short lower face and large eyes leads to
increased babyishness attributions and perceived need for
protective aid in humans (Alley, 1981; Alley, 1983). The
attribution of femininity to the same morphological
features might relate to the extended growth period in
men (Schaefer et al., 2004; Weston et al., 2007). Along
the same lines, rating studies generally confirm the
positive association of perceived masculinity, dominance
and age in human faces (e.g., Boothroyd et al., 2005;
Hess, Adams, Grammer, & Kleck, 2009; Perrett et al.,
1998). The observed pattern of overgeneralization to car
fronts in this study adds support to the notion that these
traits are a single biological dimension of inference from
facial form. Of course, we cannot say how often this
percept would arise unevoked.
Our study supersedes the findings of Pittenger et al.
(1979) by using Cartesian coordinate landmark data instead
of a model of proportional gradients referring to just a center
and an outline curve that does not allow the detection of
regional effects (e.g., headlight shape). Even though our
figures and the overall proportions of windshield-to-car body
might look similar to those presented by Pittenger et al., the
geometry of the childadult gradient that we found is, in fact,
different: their geometry was cardioidal, but ours turned out
to be axial.
4.2. Another kind of cross-cultural agreement
We also observed a second, completely different category
of cross-cultural agreement: low rating variance in cars
around the middle of certain scales (illustrated on the left in
Fig. 3). A possible source for this phenomenon is a lack of
signal value of these traits in more abstract representations
such as cars, and consequently low interrater agreement
within each culture. Many studies have shown that some of
these traits are hard to judge even in humans (e.g., for
neuroticism: Borkenau & Liebler, 1992; disgust: Hess &
Blairy, 2001) or seem to correlate with other qualities that
our silver cars do not have, such as extroversion being
related to attractiveness and conscientiousness to dress style
(Albright, Kenny, & Malloy, 1988). Our earlier study
(Windhager et al., 2008) also found low interrater agreement
for perceived conscientiousness, disgust, extroversion,
neuroticism and sadness in car fronts.
4.3. Cross-cultural differences in trait attribution to
car fronts
There were differences between Austrian and Ethiopian
ratings in items that deal with (emotional) valence (with the
exception of openness). There was high differentiation
between cars in the Austrian mean ratings, while in Ethiopia,
all cars were judged as rather arrogant(perhaps in a
positive sense), happy,friendlyand open.One
possible reason is that Ethiopians are very polite and rather
restrained in sharing their personal opinion of somebody,
especially if it is negative. A second reason might be that
interactions with real cars in Ethiopia are generally only
positive (transport of water, animals, goods, etc.). Or, third,
the more elaborated introduction to the rating task in 2009
still did not overcome the tendency to judge cars in general
instead of particular car models (that all cars were judged as
arrogantin Ethiopia, Fig. 3, favors this conclusion).
Fourth, these traits may not be readily perceived in cars but
need the emphasis of marketing strategies. To a few
Ethiopian subjects, all the cars looked alike, a phenomenon
similar to other-race and other-age effects with regard to
human faces and their modulation by experience (e.g.,
Kuefner, Macchi Cassia, Picozzi, & Bricolo, 2008).
For afraid,angryand aroused, the mean ratings
ranged around the middle of the scale, perhaps owing to low
agreement between the participants. Also, Ethiopian subjects
in general tended to use simple subdivisions of the scale
(half, one third, two thirds).
4.4. Prospects for future research
The more detailed introduction and explanation of the
rating scale in 2009 extended cross-cultural agreement from
just childadult to include malefemale and submissive
dominant. The availability of an Oromo translation might
also have contributed to this development.
Cross-cultural comparisons would further benefit from
an extension of methodologies to assess different levels of
perception and cognition. It would help to repeat the eye
tracking study of Windhager and colleagues (2010) with
an Ethiopian sample and to further investigate the
preconscious processing of car fronts by electroencepha-
lography or functional magnetic resonance imaging (cf.
Erk, Spitzer, Wunderlich, Galley, & Walter, 2002). Also,
the addition of data from more countries and cultures
(Henrich, Heine, & Norenzayan, 2010) would certainly
help to distinguish between human universals and effects
of marketing or advertisement.
The similarities of car and face perception described
here might influence driving, pedestrian behavior and the
design of car fronts. Do we change lanes sooner when an
adult, dominant car appears in the rear-view mirror? Do
children perceive cars as agents with eyes (headlights) and
therefore assume they see them anyway when they try to
cross the street?
118 S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
4.5. Conclusions
Investigating the relationship between car front shape and
personal trait attribution, we obtained very similar results in
two different countries, Austria and Ethiopia, on childadult,
femalemale and submissivedominant scales. This finding
cannot be interpreted as a result of marketing, brand or owner
stereotypes. It likely reflects the perception and interpreta-
tion of proportions that shift between child and adult. Car
fronts today might address evolutionary mechanisms
originally designed for the perception of faces. There are
implications of a bias for driving and pedestrian behavior,
and for the automotive industry.
We end with the anecdote that inspired the title of this
article. Our very first Ethiopian participant in 2008,
completing the questionnaire on the side of a dusty road
amidst cattle led to the river for drinking, commented on a
question as to whether he generally associates a human face,
an animal face or no face with cars: I do not know what to
answer. Cars have their own faces!
Acknowledgments
We thank our subjects; our two local assistants in
Ethiopia, Kumelachew Alemu and Solomon Kebede; Desta
Alemu for the Amharic and Oromo translations; Asamerew
Dessie and his family for local support; and Horst Seidler for
overall support. We are also grateful to Philipp Gunz and
Philipp Mitteroecker for sharing their Mathematica code, as
well as to Editor Robert Kurzban and three anonymous
reviewers for their thoughtful comments, which substantially
improved the manuscript. Work was supported by EFS
Unternehmensberatung GmbH research grants to K.S. (FA
547005, FA 547009) and K.G. (Ludwig Boltzmann Institute
for Urban Ethology).
Appendix A
List of the car models under study.
The numbers correspond to the data labels in Fig. 2.
1. Alfa 147
2. Audi A6
3. BMW 3
4. BMW 5
5. BMW 645ci
6. Chrysler 300C
7. Chrysler Crossfire
8. Citroen C2
9. Citroen C4
10. Daihatsu Cuore
11. Fiat Stilo
12. Ford Focus
13. Honda Civic
14. Kia Picanto
15. Lexus GS
16. Maybach
17. Mazda 6
18. Mercedes A
19. Mercedes C
20. Mercedes E
21. Mercedes SLK
22. Mini Cooper
23. Mitsubishi Colt
24. Nissan New Micra
25. Opel Astra
26. Opel Signum
27. Peugeot 307
28. Peugeot 1007 Rc
29. Renault Modus
30. Saab 9-5
31. Seat Toledo
32. Smart Passion
33. Suzuki Swift
34. Toyota Prius
35. Toyota Yaris
36. VW Golf
37. VW New Beetle
38. VW Passat
39. Citroen C1
40. Peugeot 107
41. Toyota Aygo
42. Ford Galaxy
43. Seat Alhambra
44. VW Sharan
45. VW Touran 2005
46. VW Touran 2007
References
Aggarwal, P., & McGill, A. L. (2007). Is that car smiling at me? Schema
congruity as a basis for evaluating anthropomorphized products. Journal
of Consumer Research,34, 468479.
Albright, L., Kenny, D. A., & Malloy, T. E. (1988). Consensus in
personality judgements at zero acquaintance. Journal of Personality and
Social Psychology,55, 387395.
Alley, T. R. (1981). Head shape and the perception of cuteness. Develop-
mental Psychology,17, 650654.
Alley, T. R. (1983). Infantile head shape as an elicitor of adult protection.
MerrillPalmer Quarterly,29, 411427.
Argyle, M. (2002). Körpersprache & Kommunikation: Das Handbuch zur
nonverbalen Kommunikation. 8th ed. Paderborn: Junfermann [corresponds
to Argyle,M. (1988, 2nd ed.) Bodily communication. London: Routledge].
Ashmore, R. D., & Tumia, M. L. (1980). Sex stereotypes and implicit
personality theory. I. A personality description approach to the
assessment of sex stereotypes. Sex Roles,6, 501518.
Bookstein, F. L. (1991). Morphometric tools for landmark data: geometry
and biology. New York: Cambridge University Press.
Boothroyd, L. G., Jones, B. C., Burt, D. M., Cornwell, R. E., Little, A. C., &
Tiddeman, B. P., et al. (2005). Facial masculinity is related to perceived
age but not perceived health. Evolution and Human Behavior,26,
417431.
Borkenau, P., & Liebler, A. (1992). Trait inferences: sources of validity at
zero acquaintance. Journal of Personality and Social Psychology,62,
645657.
119S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
Brown, E., & Perrett, D. I. (1993). What gives a face its gender? Perception,
22, 829840.
Bruce, V., Burton, A. M., Hanna, E., Healey, P., Mason, O., & Coombes, A.,
et al. (1993). Sex discrimination: how do we tell the difference between
male and female faces? Perception,22, 131152.
Bulygina, E., Mitteroecker, P., & Aiello, L. (2006). Ontogeny of facial
dimorphism and patterns of individual development within one human
population. American Journal of Physical Anthropology,131,
432443.
Coss, R. G. (2003). The role of evolved perceptual biases in art and
design. In E. Voland, & K. Grammer (Eds.), Evolutionary aesthetics
(pp. 69-130). Berlin: Springer.
Ekman, P. (1999). Basic emotions. In T. Dalgleish, & M. Power (Eds.),
Handbook of cognition and emotion (pp. 45-60). Sussex, UK: John
Wiley & Sons, Ltd.
Enlow, D. H. (1975). Handbook of facial growth. Philadelphia: W.B.
Saunders.
Enlow, D. H., & Hans, M. G. (1996). Essentials of facial growth.
Philadelphia: Saunders.
Erk, S., Spitzer, M., Wunderlich, A. P., Galley, L., & Walter, H. (2002).
Cultural objects modulate reward circuitry. NeuroReport,13,
24992503.
Good, P. (2000). Permutation tests: a practical guide to resampling methods
for testing hypotheses. 2nd ed. New York: Springer.
Grammer, K., & Thornhill, R. (1994). Human (Homo sapiens) facial
attractiveness and sexual selection: the role of symmetry and
averageness. Journal of Comperative Psychology,108, 233242.
Guthrie, R. D. (1970). Evolution of human threat display organs. In T.
Dobzhansky, M. K. Hecht, & W. C. Steere (Eds.), Evolutionary biology,
Vol. 4. (pp. 257-302). New York: AppletonCenturyCrofts.
Guthrie, S. E. (1993). Faces in the clouds: a new theory of religion. New
York: Oxford University Press.
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in
the world. Behavioral and Brain Sciences,33,61135.
Hess, U., Adams, R. B., & Kleck, R. E. (2005). Who may frown and who
should smile? Dominance, affiliation, and the display of happiness and
anger. Cognition and Emotion,19, 515536.
Hess, U., Adams Jr, R. B., Grammer, K., & Kleck, R. E. (2009). Face gender
and emotion expression: are angry women more like men? Journal of
Vision,9,18.
Hess, U., & Blairy, S. (2001). Facial mimicry and emotional contagion to
dynamic emotional facial expressions and their influence on decoding
accuracy. International Journal of Psychophysiology,40, 129141.
Keating, C. F., Mazur, A., & Segal, M. H. (1977). Facial gestures which
influence the perception of status. Sociometry,40, 374378.
Kuefner, D., Macchi Cassia, V., Picozzi, M., & Bricolo, E. (2008). Do all kids look
alike? Evidence for an other-age effect in adults. Journal of Experimental
Psychology Human Perception and Performance,34, 811817.
Mark, L. S., Shaw, R. E., & Pittenger, J. B. (1988). Natural constraints,
scales of analysis, and information for the perception of growing
faces. In T. R. Alley, Ed. Social and applied aspects of perceiving
faces (pp. 11-49). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Marsh, A. A., Yu, H. H., Schechter, J. C., & Blair, R. J. R. (2009). Larger
than life: humans' nonverbal status cues alter perceived size. PLoS ONE,
4, e5707.
McCrae, R. R., & Costa, P. T. (1997). Personality trait structure as a human
universal. American Psychologist,52, 509516.
Mitteroecker, P., & Gunz, P. (2009). Advances in geometric morphometrics.
Evolutionary Biology,36, 235247.
Montepare, J. M., & Zebrowitz, L. A. (1998). Person perception comes
of age: the salience and significance of age in social judgments. In
M. Zanna, Ed. Advances in experimental social psychology, Vol. 30.
(pp. 93-163). New York: Academic Press.
Morewedge, C. K., Preston, J., & Wegner, D. M. (2007). Timescale bias in
the attribution of mind. Journal of Personality and Social Psychology,
93,111.
Nettle, D. (2004). Adaptive illusions: optimism, control and human
rationality. In D. Evans, & P. Cruse (Eds.), Emotion, evolution and
rationality (pp. 193-208). Oxford: Oxford University Press.
Perrett, D. I., Lee, K. J., Penton-Voak, I., Rowland, D., Yoshikawa, S., &
Burt, D. M., et al (1998). Effects of sexual dimorphism on facial
attractiveness. Nature,394, 884887.
Pittenger, J. B., Shaw, R. E., & Mark, L. S. (1979). Perceptual information
for the age level of faces as a higher order invariant of growth. Journal of
Experimental Psychology: Human Perception and Performance,5,
478493.
Rohlf, F. J. (2008). tpsDig2 (Version 2.12): Department of Ecology and
Evolution. New York: State University of New York at Stony Brook.
Rohlf, F. J. (2009). tpsRegr, shape regression (Version 1.36): Department of
Ecology and Evolution. New York: State University of New York at
Stony Brook.
F.J.Rohlf, F. J., & D.E.Slice, D. E. (1990). Extensions of the Procrustes
method for the optimal superimposition of landmarks. Systematic
Zoology,39,4059.
Russell, J. A. (1997). Reading emotions from and into faces:
resurrecting a dimensionalcontextual perspective. In J. A. Russell,
& J. M. Fernández-Dols (Eds.), The psychology of facial expression
(pp. 295-320). Paris: Cambridge University Press.
Schaefer, K., Mitteroecker, P., Fink, B., & Bookstein, F. L. (2009).
Psychomorphospace from biology to perception, and back: towards
an integrated quantification of facial form variation. Biological Theory,
4,98106.
Schaefer, K., Mitteroecker, P., Gunz, P., Bernhard, M., & Bookstein, F. L.
(2004). Craniofacial sexual dimorphism patterns and allometry among
extant hominids. Annals of Anatomy,186, 471478.
Senior, C., Barnes, J., Jenkins, R., Landau, S., Phillips, M. L., & David, A. S.
(1999). Attribution of social dominance and maleness to schematic faces.
Social Behavior and Personality,27, 331337.
Sperber, D. (1994). The modularity of thought and the epidemiology of
representations. In L. A. Hirschfeld, & S. A. Gelman (Eds.), Mapping
the mind: domain specificity in cognition and culture (pp. 39-67).
Cambridge, UK: Cambridge University Press.
Todorov, A., Said, C. P., Engell, A. D., & Oosterhof, N. N. (2008).
Understanding evaluation of faces on social dimensions. Trends in
Cognitive Sciences,12, 455460.
Trenouth, M. J., & Joshi, M. (2006). Proportional growth of craniofacial
regions. Journal of Orofacial Orthopedics,67,92104.
Waytz, A., Epley, N., & Cacioppo, J. T. (2010). Social cognition unbound:
insights into anthropomorphism and dehumanization. Current Directions
in Psychological Science,19,5862.
Weston, E. M., Friday, A. A., & Liò, P. (2007). Biometric evidence that
sexual selection has shaped the Hominin face. PloS ONE,2, e710.
Williams, J. E., Satterwhite, R. C., & Best, D. L. (1999). Pancultural
gender stereotypes revisited: the five factor model. Sex Roles,40,
513525.
Windhager, S., Bookstein, F.L., Grammer, K., Thorstensen, T., & Schaefer, K.
(2009). Principles of growth perception: does the human face processing
pathway generalize to cars? American Journal of Physical Anthropology,
138,274275.
Windhager, S., Hutzler, F., Carbon, C. C., Oberzaucher, E., Schaefer, K., &
Thorstensen, T., et al. (2010). Laying eyes on headlights: eye
movements suggest facial features in cars. Collegium Antropologicum,
34, 10751080.
Windhager, S., Slice, D. E., Schaefer, K., Oberzaucher, E., Thorstensen, T.,
& Grammer, K. (2008). Face to face the perception of automotive
designs. Human Nature,19, 331346.
120 S. Windhager et al. / Evolution and Human Behavior 33 (2012) 109120
... In 2016, in a survey among 6000 Americans, 8.4% reported to own patriotic decals [16]. The vehicle appearance leads to the formation of expectations on the driving behavior and even attribute person-like traits [64,65]. It was found that model [15], shape [2,64], and color [15] influence these expectations. ...
... The vehicle appearance leads to the formation of expectations on the driving behavior and even attribute person-like traits [64,65]. It was found that model [15], shape [2,64], and color [15] influence these expectations. However, people are accurate in estimating the real speed of vehicles despite these expectations [32]. ...
... Since then, geometric morphometrics has been continually refined and has found countless applications in biological, anthropological, paleontological, medical, psychological, archeological, and industrial fields (for reviews see, e.g., Adams & Otárola-Castillo, 2013;Bookstein, 2018;Cardini, 2020;Elewa, 2010;Halazonetis, 2004;Klingenberg, 2010;Lawing & Polly, 2010;MacLeod, 2018;Mitteroecker, 2020;Mitteroecker & Gunz, 2009;Schaefer et al., 2009;Slice, 2005;Wiley et al., 2005;Zelditch et al., 2012). The geometric morphometric toolkit has also been connected to other methodologies, including biomechanics (e.g., O'Higgins et al., 2019;Parr et al., 2012;Polly et al., 2016;Weber et al., 2011), systematics and phylogenetics (e.g., Adams, 2014;Klingenberg & Gidaszewski, 2010;Monteiro, 2013;Rohlf, 2002), image analysis (e.g., Mayer et al., 2014Mayer et al., , 2017, quantitative genetics (e.g., Adams, 2011;Baab, 2018;Martínez-Abadías et al., 2009;Pavličev et al., 2016;Schroeder & von Cramon-Taubadel, 2017), genetic mapping (e.g., Mitteroecker et al., 2016;Pallares et al., 2015;Var on-González et al., 2019), evolutionary psychology and brain imaging (e.g., Walla et al., 2020;Windhager et al., 2012Windhager et al., , 2018 as well as molecular and developmental biology (e.g., Arif et al., 2013;Buchberger et al., 2021;Hallgrimsson et al., 2015;Marchini et al., 2021;Martínez-Abadías et al., 2018). Recent implementations of geometric morphometric methods into R and Mathematica facilitated analyses (Adams & Otárola-Castillo, 2013;Dryden, 2021;Dryden & Mardia, 2016;Polly, 2017;Schlager, 2017). ...
Article
Full-text available
The foundations of geometric morphometrics were worked out about 30 years ago and have continually been refined and extended. What has remained as a central thrust and source of debate in the morphometrics community is the shared goal of meaningful biological inference through a tight connection between biological theory, measurement, multivariate biostatistics, and geometry. Here we review the building blocks of modern geometric morphometrics: the representation of organismal geometry by landmarks and semilandmarks, the computation of shape or form variables via superimposition, the visualization of statistical results as actual shapes or forms, the decomposition of shape variation into symmetric and asymmetric components and into different spatial scales, the interpretation of various geometries in shape or form space, and models of the association between shape or form and other variables, such as environmental, genetic, or behavioral data. We focus on recent developments and current methodological challenges, especially those arising from the increasing number of landmarks and semilandmarks, and emphasize the importance of thorough exploratory multivariate analyses rather than single scalar summary statistics. We outline promising directions for further research and for the evaluation of new developments, such as “landmark‐free” approaches. To illustrate these methods, we analyze three‐dimensional human face shape based on data from the Avon Longitudinal Study of Parents and Children (ALSPAC).
... Ichikawa et al. (2011) dealt with face-like objects, for example, electrical outlets, flowers, and houses. Some objects resemble human faces and the headlights of cars look like eyes (Windhager et al., 2012). Thus, matching cars to their owners might involve judging the resemblance between the face-like views of cars and the owners' face. ...
Article
Full-text available
Many often say that people resemble their pets or that the faces of manga characters and Buddha statues resemble those of their artists. Previous studies demonstrated that participants could match dogs with their owners, suggesting that pets resemble their owners. Other studies also demonstrated that people can match personal belongings, including inanimate objects, to their owners. However, it is unknown whether people tend to make objects that resemble themselves. In this study, we examined whether people tend to make objects that resemble themselves with dolls made of cloth as stimuli. The results demonstrated that people tend to project themselves into dolls, even in the case of amateur college students. The mere exposure effect or the algorithm "self seeks like" may be at play in not only people's selection of pets but also their making of objects.
... The eHMI communicates the vehicle's situational awareness (SA) with a visualization of 'eyes' as has been used in past concepts [2,12]. However, in this concept, we address some of the typical limitations of eye-based SA concepts: Eye based eHMIs are typically demonstrated on the headlights of a vehicle as in a vehicle the headlights echo the anthropomorphic representation of an eye in the 'face' of a vehicle [17,18]. This limits the ability to spot multiple (groups of) pedestrians in different relative positions around the vehicle. ...
Conference Paper
Full-text available
External Human-Machine Interfaces (eHMIs) are proposed to address the communication gap between automated vehicles (AVs) and other road users. However, existing concepts of eHMIs are often limited in their ability to communicate in a clear, unambiguous, and scalable way, which hinders their effectiveness. eHMIs typically communicate by showing the vehicle's driving intention, situational awareness, or its path/ trajectory; none of which is independently able to sufficiently achieve effective and unambiguous communication. We created a novel eHMI concept which combines these three communication elements in one unified design that (1) promises a more seamless, two-way interaction between AVs and other road users, and (2) is capable of a better scalable, high-resolution communication, i.e., facilitating communication beyond one-on-one communication while mitigating ambiguity regarding the specific recipient of the communication message. This addresses some of the limitations of existing eHMIs and paves a path towards a novel approach in eHMI design for future AVs.
... It would be interesting to understand why cars are so often given a name in comparison with other kinds of objects. The relatively high frequency of car naming could be due to the fact that certain facial traits (such as eyes, a mouth or a nose) tend to be perceived in car fronts (Windhager et al. 2008;2012). An eye-tracking study even showed that eye movements and fixations while watching car fronts were similar to those when looking faces (Windhager et al., 2010). ...
Article
Full-text available
Some people give a proper name to an owned individual object, such as a car or a computer. The study examined whether giving a proper name to a specific object is associated with object personification, and more specifically, whether object personification is a prerequisite to name giving. The latter question was assessed by asking 130 participants whether, in their adult life, they had ever given a personal name to an object, and if so, whether they had attributed psychological characteristics to that named object. The general relationship between personal name giving and personification was assessed by evaluating whether the scores from a questionnaire on anthropomorphism differed in participants who reported having given a specific name to at least one personal object, compared with those who reported not doing so (Mann-Whitney’s U test). Results showed that the scores from the questionnaire on anthropomorphism were significantly higher for participants who had given specific names to objects than for participants who had not done so. However, object personification was not found to be a prerequisite to name giving. Indeed, about 40 percent of people who reported giving personal names to objects did not attribute psychological qualities to these objects.
... In addition, as self-driving progresses, the amount of information exchanged between drivers and vehicles is increasing rapidly, and as a medium for information interaction, the car center fascia is becoming larger and more diverse in form. As the interior of the car is different by nationality [7] [8], it should be designed considering the preferences of each nationality, however affective research on the shape of the center fascia is still lacking. ...
Chapter
Successful dissemination of a new technology requires broad acceptance in society. This is no different with autonomous vehicles. Despite the benefits that autonomous driving promises to bring with it, many people remain skeptical about letting a computer control a car. In recent years, numerous studies have therefore investigated when, how and why people would (or would not) be willing to make use of autonomous vehicles. This chapter gives an overview of the relevant scientific literature. After a summary of essential theoretical frameworks, the variables that were found by empirical research to predict or correlate with the acceptance of autonomous vehicles are organized into three categories: user-specific determinants (e.g., socio-demographics and personality traits), car-specific determinants (e.g., perceived safety, predictability and appearance), and contextual determinants (e.g., road conditions). Based on this review, limitations of previous work, open research questions and practical implications are discussed, which leads to the ultimate conclusion that the obstacles on the road to autonomous mobility are not merely technical, but also psychological.
Article
Full-text available
Objective: It is of critical importance to develop socially sensitive vehicles that will enhance pedestrians’ sense of comfort and safety. The current study is the first to extend these effects to vehicles, by investigating individual comfort distance in virtual reality with regard to vehicles that vary in terms of size, viewing angle and anthropomorphized emotional expression. Furthermore, we investigate the effect of individual differences in terms of height, anxiety and aggression. Method: Forty-four individuals were presented with three-dimensional stimuli of vehicle models differing in size and viewing angle in virtual reality and positioned them at the distance they felt the most comfortable with. Results: Our results show that individuals are more comfortable standing further from larger vehicles and when presented with the front versus the rear view of a vehicle. Moreover, the distance from vehicles was negatively associated with the height of the individuals. Conclusion: This paper suggests that it is important for designing self-driving and autonomous vehicles to consider that vehicle size and direction as well as pedestrian’s height may impact the comfort distance felt by pedestrians. These data have clear implications for vehicle design, including self-driving and autonomous vehicles.
Article
Full-text available
Humans have evolved the cognitive ability to perceive and associate dominance, masculinity, and emotions to animals and humans of larger sizes. Previous research has shown that humans are able to anthropomorphize nonbiological objects such as cars and attribute different characteristics to them. We hypothesized that larger vehicles are perceived and processed in a similar fashion as other biological agents, and are therefore considered more aggressive, dominant, and masculine compared to smaller vehicles. This study investigated the effect of vehicles’ size on the perception of dominance, masculinity, anger, and hostility. A total of 221 individuals (139 men and 82 women) participated in the study and rated vehicles of large and small sizes on four dimensions of Submissive-Dominant, Angry-Happy, Masculine-Feminine, and Hostile-Friendly. Results showed that participants rated larger vehicles as more dominant, angry, hostile, and masculine than smaller vehicles, supporting the proposal that, similar to biological agents, large vehicles are more threatening and are associated with higher dominance. Moreover, while men and women responded to size in a similar manner, the ratings of anger, masculinity, and hostility for large vehicles increased with the age of participants. The implications of this work are considered, including ways to enhance the social acceptance of autonomous vehicles.
Article
This investigation aggregates intact or reconstructed Gahagan bifaces from the southern Caddo area and central Texas to test the hypothesis that Gahagan biface morphology differs between the regions. The Gahagan bifaces (n=102) were scanned, then analyzed using a novel landmarking protocol and the tools of geometric morphometrics. Results provide a preview of the significant differences in Gahagan biface morphology expressed between the southern Caddo area and central Texas regions. The size discrepancy represents an inversion of current theoretical constructs that posit a decrease in tool size thought to articulate with an increase in distance from the raw material source. It is posited that the contrasting morphologies represent two discrete communities of practice; one (emergent Caddo horticulturalists) where Gahagan bifaces were enlisted primarily for burial and ritualistic activities, and the other (central Texas hunter-gatherers) where Gahagan bifaces were utilized over a longer time span in more practical and utilitarian contexts.
Article
Full-text available
Differing positions of the eyebrows play a role in nonhuman primate displays of status (dominance and submission). Mazur and Stevens (1975) suggest that similar eyebrow gestures convey information about social status among humans, and this hypothesis was tested in the present experiment. We predicted that human models would be made to look more dominant by photographing them with lowered as compared to raised brows. College student observers were shown such portrait photographs of male and female models from various racial backgrounds and asked to judge dominance. Observers perceived each of twelve models as dominant significantly more often when models posed with lowered eyebrows than when they posed with raised eyebrows. Brows were more expressive of dominance than mouth gestures, which served as "controls." Results of status judgments made on cartoon face stimuli were consistent with those based on the photographs.
Article
We report findings which suggest perception of 'higher order' attributes such as gender and social dominance are perceived from a schematic face. To investigate a large population, the first two experiments were carried out in both the traditional manner and on the Internet. Results obtained from both were not significantly different so the data sets were combined. Lowered eyebrow position was a strong indicator of both social dominance and the male gender. A schematic face with a sad mouth resulted in the face's being viewed as less dominant and less male. Eyegaze direction also was investigated and discussed in terms of dyadic influence. Evidence supported the assumption that both social dominance and the male gender are perceived through similar facial configurations on a schematic face. Limitations include the use of schematic face pairs, and the presentation of single faces in research is discussed.
Chapter
The seminal ideas about the relationship of aesthetic appreciation and evolutionary theory emerged initially with the Darwinian construct of sexual selection that emphasized the importance of mate choice and physical attractiveness (Darwin 1885). Anthropomorphic linkage of processes of human intelligence and those of other species (Romanes 1886), coupled with emphasis on the role of natural selection in adjusting human intelligence (Spencer 1888), set the stage for describing behavior in terms of evolutionary history. The idea that innate knowledge accumulated over successive generations and influenced current behavior pervaded nineteenthcentury tomes that characterized behavioral relics as atavistic (Nietzsche 1909), especially the fearful behavior of children (Hall 1897). Jung (1916, 1972) continued the development of these ideas with his construct of the archetype as a species-typical pattern of thought that might account for cross-cultural similarities in mythology and graphical symbolism.
Article
This chapter considers the relationship between emotions and deliberative rationality. It argues that there are evolutionary reasons to think that emotions are in some cases likely systematically to skew rationality. Standard models of rationality assume that the mind is able to come to accurate assessments of the probability of future contingencies. However, robust evidence shows that people systematically overestimate the probability of positive future contingencies, and underestimate the probability of negative ones - only those who are depressed or dysphoric come to accurate assessments. The chapter argues that there are good evolutionary reasons why this should be the case, since there is an asymmetric pattern of costs and benefits from getting motivational judgements wrong.
Book
Morphometrics is the statistical study of biological shape and shape change. Its richest data are landmarks, points such as 'the bridge of the nose' that have biological names as well as geometric locations. This book is the first systematic survey of morphometric methods for landmark data. The methods presented here combine conventional multivariate statistical analysis with themes from plane and solid geometry and from biomathematics to support biological insights into the features of many different organs and organisms. This book will be of value to applied statisticians and geometers, as well as to all biological and biomedical researchers who need quantitative analyses of information from biomedical images.