Content uploaded by Nicholas Epley
Author content
All content in this area was uploaded by Nicholas Epley on Jul 28, 2014
Content may be subject to copyright.
http://pps.sagepub.com/
Science
Perspectives on Psychological
http://pps.sagepub.com/content/5/3/219
The online version of this article can be found at:
DOI: 10.1177/1745691610369336
2010 5: 219Perspectives on Psychological Science
Adam Waytz, John Cacioppo and Nicholas Epley
Who Sees Human? : The Stability and Importance of Individual Differences in Anthropomorphism
Published by:
http://www.sagepublications.com
On behalf of:
Association For Psychological Science
can be found at:Perspectives on Psychological ScienceAdditional services and information for
http://pps.sagepub.com/cgi/alertsEmail Alerts:
http://pps.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Who Sees Human? The Stability and
Importance of Individual Differences in
Anthropomorphism
Adam Waytz
1
, John Cacioppo
2
, and Nicholas Epley
3
1
Department of Psychology, Harvard University, Cambridge, MA;
2
Department of Psychology,
University of Chicago, IL; and
3
Booth School of Business, University of Chicago, IL
Abstract
Anthropomorphism is a far-reaching phenomenon that incorporates ideas from social psychology, cognitive psychology,
developmental psychology, and the neurosciences. Although commonly considered to be a relatively universal phenomenon with
only limited importance in modern industrialized societies—more cute than critical—our research suggests precisely the
opposite. In particular, we provide a measure of stable individual differences in anthropomorphism that predicts three important
consequences for everyday life. This research demonstrates that individual differences in anthropomorphism predict the degree
of moral care and concern afforded to an agent, the amount of responsibility and trust placed on an agent, and the extent to which
an agent serves as a source of social influence on the self. These consequences have implications for disciplines outside of
psychology including human–computer interaction, business (marketing and finance), and law. Concluding discussion addresses
how understanding anthropomorphism not only informs the burgeoning study of nonpersons, but how it informs classic issues
underlying person perception as well.
Keywords
anthropomorphism, social cognition, individual differences
General Motors (GM) ran an advertisement during the 2007
Super Bowl to demonstrate their commitment to manufacturing
quality. The advertisement, rated by viewers as the fourth most
popular ad shown during the game, capitalized on people’s
tendency to anthropomorphize by depicting a factory line robot
being fired from its job after it inadvertently dropped a screw it
was designed to install in a car. The ostensibly depressed robot
takes a series of low-level jobs until it becomes ‘‘distraught’’
enough to roll itself off a bridge. GM’s intended message was
clear—the slightest glitch in production would not meet their
quality standards—but so was their unintended message—that
depression had led the easily anthropomorphized robot to
commit suicide. The ad immediately incensed the American
Foundation for Suicide Prevention who said the ad ‘‘portrays
suicide as a viable option when someone fails or loses their
job’’ and that ‘‘research has also shown that graphic, sensatio-
nalized, or romanticized descriptions of suicide deaths in any
medium can contribute to suicide contagion, popularly referred
to as ‘copycat’ suicides’’’ (Associated Press, 2007). This exam-
ple seems to confirm David Hume’s (1757/1957, p. xix)
assertion that ‘‘there is a universal tendency among mankind
to conceive all beings like themselves.’’
Marketers appear to believe that anthropomorphism matters.
Hume appears to believe that anthropomorphism is a universal
tendency. We evaluate both of these claims by examining
whether there are stable individual differences in the tendency
to attribute humanlike attributes to nonhuman agents and
whether such differences map onto important judgments, deci-
sions, or behaviors. Our research suggests that the claim that
anthropomorphism is universal may be overstated; individual
differences in anthropomorphism exist, and we provide a
psychometrically valid measure of them (the Individual
Differences in Anthropomorphism Questionnaire, or IDAQ).
At the very least, anthropomorphism does not appear to be
Corresponding Author:
Adam Waytz, Department of Psychology, Harvard University, Northwest
Science Building, 52 Oxford Street, Cambridge, MA 02138
E-mail: waytz@wjh.harvard.edu
Perspectives on Psychological Science
5(3) 219–232
ª The Author(s) 2010
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1745691610369336
http://pps.sagepub.com
219
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
universal in the sense that it occurs to an equivalent degree
across all of ‘‘mankind.’’ Our research also suggests that mar-
keters are right to care about anthropomorphism; individual
differences in anthropomorphism matter for creating an
empathic connection with nonhuman agents, for judgments of
responsibility and culpability, and for creating social influence.
These consequences have implications for human–computer
interaction, business, law, and the inverse process of dehuma-
nization. Because of these broad implications, we argue that
psychologists should care more about understanding anthropo-
morphism as well.
Psychological Approaches to
Anthropomorphism
Within psychology, anthropomorphism has traditionally fig-
ured as a topic of debate over the accuracy of its use in studying
nonhuman animals rather than as a topic of scientific inquiry
aimed at understanding when people anthropomorphize nonhu-
man agents and when they do not (Cheney & Seyfarth, 1990;
Epley, Waytz, & Cacioppo, 2007; Hauser, 2000). Although
an interesting topic, whether dogs or cats or gadgets or gods
actually possess the humanlike attributes that people attribute
to them is orthogonal to understanding the psychological
mechanisms that lead people to attribute humanlike qualities
to these agents. It is important to avoid confusing questions
about anthropomorphism’s accuracy with questions about
anthropomorphism’s variability, frequency, and consequences.
Recent years, however, have seen a rapidly increasing inter-
est in understanding people’s propensity to turn nonhuman
agents into human ones (Bering, 2006; Epley et al., 2007;
Kwan & Fiske, 2008). Anthropomorphism touches on central
topics in virtually all major subfields within psychology, incor-
porating insights on the brain mechanisms underlying social
cognition in neuroscience, on reasoning and induction in cog-
nitive psychology, and on theory of mind in developmental
psychology. For example, neuroscientists have examined the
neural correlates of anthropomorphism (Gazzola, Rizzolatti,
Wicker, & Keysers, 2007; Harris & Fiske, 2008) and identified
deficits in anthropomorphism for amygdala-damaged patients
(Heberlein & Adolphs, 2004), as well as people diagnosed with
autism (Castelli, Frith, Happe´, & Frith, 2002). Cognitive
psychologists have examined anthropocentrism as a process
of inductive reasoning about biological kinds (Anggoro,
Waxman, & Medin, 2008; Waxman & Medin, 2007), and
developmental psychologists have assessed children’s capacity
to perceive humanlike intentions in nonhuman stimuli (Scholl
& Tremoulet, 2000) in the trajectory of learning to reason about
mental states.
Work in these subdomains has proceeded largely indepen-
dently, with a primary focus on the situational, developmental,
or cultural determinants of anthropomorphism (Epley et al.,
2007) rather than on its dispositional determinants or on the
potentially important consequences of anthropomorphism in
everyday life. We do the latter in this article by first presenting
a measure of individual differences in anthropomorphism and
then examining this measure’s predictive utility for the
evaluation, treatment, and social influence of a variety of
nonhuman agents. Finally, we explain why understanding
anthropomorphism is important for areas outside of psychol-
ogy, including human–computer interaction, business, and law.
Far more than just a cute and inconsequential response to
stuffed animals or marketing campaigns, anthropomorphism
is critical for understanding how people interact with an
increasingly wide variety of technological agents, how people
make decisions about seemingly agentic financial markets, and
how people decide who should be treated with the respect and
dignity afforded to other humans and who should not.
What Is Anthropomorphism?
Psychologists have used the term anthropomorphism rather
loosely to describe everything from mistaken inferences about
nonhuman agents to almost any kind of dispositional inference
about a nonhuman agent, definitions that do not fit with the
actual dictionary definition of attributing ‘‘human characteris-
tics or behavior to a god, animal, or object’’ (Soanes & Steven-
son, 2005). Xenophanes (6th Century B.C., as cited in Lesher,
1992) was the first to use the term anthropomorphism when
describing how gods and other supernatural agents tended to
bear a striking physical resemblance to their believers. Xeno-
phanes’s observation reflects one of two basic ways of anthro-
pomorphizing. The first involves attributing humanlike
physical features to nonhumans (e.g., a face, hands), and the
second involves attributing a humanlike mind to nonhumans
(e.g., intentions, conscious awareness, secondary emotions
such as shame or joy). Anthropomorphism therefore requires
going beyond purely behavioral or dispositional inferences
about a nonhuman agent and instead requires attributing human
form or a human mind to the agent. Regarding a fox as quick
does not necessarily denote anthropomorphic reasoning, but
regarding a fox as wily does. The former is simply a description
of an observable behavior, whereas the latter refers to a distinc-
tively mental quality. Anthropomorphism also goes beyond
animism—simply attributing life to a nonliving object. The
essence of anthropomorphism is therefore attributing capacities
that people tend to think of as distinctly human to nonhuman
agents, in particular humanlike mental capacities (e.g., inten-
tionality, emotion, cognition). The presence of mental states
constitutes both a necessary and sufficient condition for
humanness, as the presence of a humanlike face or humanlike
body movement generally implies the presence of humanlike
mental states as well (Johnson, Slaughter, & Carey, 1998;
Morewedge, Preston, & Wegner, 2007).
Although humans may not be the only agents with sophisti-
cated mental capacities, both philosophical and lay theories of
personhood focus on mental states as the defining feature that
distinguishes humans from other agents (Demoulin et al., 2004;
Haslam, Bain, Douge, Lee, & Bastian, 2005). Philosophical
definitions of personhood focus on the possession of higher
order mental capacities like self-reflection, metacognition,
conscious intention, and rational thought (Boethius, 6th
220 Waytz, Cacioppo, and Epley
220
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Century, cited in Farah & Heberlein, 2007; also see Dennett,
1978; Kant, 1785/1959; Locke, 1841/1997). People’s lay
theories of humanness also center on traits (e.g., imaginative)
and emotions (e.g., humiliation) that implicate higher order
mental states such as self-reflection, mental simulation, and
prospection (Demoulin et al., 2004; Haslam et al., 2005).
Anthropomorphism can therefore be operationalized as a
particular form of mental state attribution.
Measuring Stable Behavioral Tendencies:
The IDAQ
People may be readily able to think of a hurricane as vindictive
or an animated robot as depressed, but this does not mean that
such anthropomorphism is a tendency that all share in equal
degrees. Stable individual differences in the tendency to
anthropomorphize may arise from differences in culture,
norms, experience, education, cognitive reasoning styles, and
attachment to human and nonhuman agents (Epley et al.,
2007). We examined the possibility of such stable individual
differences by constructing the individual differences in
anthropomorphism questionnaire (IDAQ) and then examining
its factor structure.
Although many researchers have examined anthropomorph-
ism, no systematic measure of individual differences has
emerged. Measures have ranged from explicit questions about
how much an agent looks or acts human (Kiesler & Goetz,
2002) to implicit measures of memory mistakes in which peo-
ple recall supernatural agents behaving in humanlike ways
(Barrett & Keil, 1996). Research in other disciplines has devel-
oped scales to measure anthropomorphic tendencies (Chin
et al., 2005) or relationships with specific nonhuman targets
(e.g., parasocial characters, Auter, 1992; or God, Paloutzian
& Ellison, 1982), but these measures either do not measure
anthropomorphism efficiently (e.g., the number of items for
some measures ranges from 78 to 208), do not measure attribu-
tion of qualities that people perceive to be distinctively human
(and therefore do not qualify as anthropomorphism), or do not
assess anthropomorphism across a diverse array of nonhuman
targets. We aim to develop a single questionnaire-based
measure of anthropomorphism that predicts judgments across
targets and provides a common metric to promote research
on this topic.
In developing the IDAQ, we generated items by first iden-
tifying four classes of commonly anthropomorphized
agents—nonhuman animals, natural entities, spiritual agents,
and technological devices—and then pai ring each class of
agent with five anthropomorphic and five nonanthropo-
morphic traits (see Appendix). The 15 nonanthropomorphism
items (IDAQ-NA)
1
are not part of the IDAQ, but they are
included to dissociate anthropomorphism from dispositional
attribution more generally and to ensure that differences in
anthropomorphism do not merely reflect differences in scale
use. For purp ose s of theoretical focus, we have summarized
all subseque nt studi es in the text and provide acces s to
complete methods, materials, and addi tional analyses at
https://sites.google.com/site/idaqmaterials/.
In Study 1, 348 individuals from the University of Chicago
population completed the IDAQ. A preliminary exploratory
factor analysis on responses to all 40 items revealed three fac-
tors, one that captured anthropomorphism of animal stimuli, a
second that captured anthropomorphism of nonanimal stimuli
(e.g., technology and nature), and a third that captured all items
(both anthropomorphic and nonanthropomorphic) pertaining to
spiritual agents. This spiritual agent factor suggests that parti-
cipants did not discriminate between anthropomorphic and
nonanthropomorphic attributions of spiritual entities and that
this factor instead reflected more general beliefs about the exis-
tence of religious or spiritual agents.
Because the IDAQ should reflect anthropomorphism rather
than degree of religious belief, we excluded the spiritual target
items and assessed the factor structure for the 30 remaining
items pertaining to material agents (animals, technology, and
nature). This EFA identified a two-factor solution as optimal
(RMSEA ¼ 0.067, 90% CI ¼ 0.062–0.073), reflecting
anthropomorphism of animal stimuli and anthropomorphism
of nonanimal stimuli. Nonanthropomorphism items loaded dif-
fusely and insubstantially across both factors (see Table 1 for
all factor loadings).
Study 1 also revealed that these two factors are positively
correlated, differing only in terms of the target stimuli rather
than in the relative degree of anthropomorphic attributions.
Anthropomorphism may therefore be a more general beha-
vioral tendency that people engage in more or less across all
nonhuman targets (see Guthrie, 1993, and Mithen, 1996, for
similar suggestions). In Study 2, we investigated this possibil-
ity that a single superordinate factor can account for anthropo-
morphism of both classes of stimuli, and also examined the
reliability of the factor structure observed in Study 1 by draw-
ing participants from a different population.
In Study 2, 609 individuals from the general population
completed the revised 15-item IDAQ (accompanied by the
15 nonanthropomorphism items). A confirmatory factor analy-
sis (CFA) of all 30 items specified the 10 anthropomorphic
items pertaining to nonanimals as a first factor and the 5 anthro-
pomorphic items pertaining to animals as a second anthropo-
morphism factor, permitting the nonanthropomorphism items
to load on both factors. This model provided good fit
(RMSEA ¼ .073, 90% CI ¼ .070–.077; see Box 1 for addi-
tional measures of fit), and revealed a significant correlation
between the two first-order anthropomorphism factors (r ¼
.52, p < .001). To determine whether this reflects a single
superordinate factor, we applied a second-order CFA (using
only the 15 IDAQ items), which specified factors assessing
anthropomorphism of animals and anthropomorphism of nona-
nimals and indicated ‘‘general anthropomorphism’’ as the
superordinate factor. This model provided good fit (RMSEA
¼ .077, 90% CI ¼ .070–.085; see Fig. 1), with the superordi-
nate factor of general anthropomorphism loading highly on the
animal anthropomorphism factor (0.88), and moderately highly
on the nonanimal anthropomorphism factor (0.57; see Table 2).
Who Sees Human 221
221
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
The 15 items assessing anthropomorphism across agents are
also highly intercorrelated (a .82 in all studies). Anthropo-
morphism of both animate and inanimate stimuli therefore
appear to be manifestations of a more general tendency to
anthropomorphize nonhuman agents.
A third study examined the temporal stability of the IDAQ
by having participants from Study 1 complete the measure a
second time, 12 to 19 weeks after the initial study. This yielded
evidence of reasonable temporal stability, r(67) ¼ .55, p <
.0001. Together, these findings demonstrate a reliable tendency
for some people to anthropomorphize more than others, and
they provide a psychometrically valid measure of this ten-
dency. Of course, measuring stable tendencies is worthwhile
only to the extent that they predict judgments or behaviors that
matter in everyday life. Having developed a reliable measure of
anthropomorphism, we now use the IDAQ to examine why it
matters.
Consequences of Anthropomorphism
Perceiving an agent to have a humanlike mind has at least three
major consequences for both the perceiver and the agent per-
ceived (Epley & Waytz, 2009). First, perceiving an agent to
have a mind means that agent is capable of conscious experi-
ence and should therefore be treated as a moral agent worthy
of care and concern (Gray, Gray, & Wegner, 2007). Second,
perceiving an agent to have a mind means that the agent is
capable of intentional action and can therefore be held respon-
sible for its actions (Gray et al., 2007). Finally, perceiving an
agent to have a mind means that the agent is capable of obser-
ving, evaluating, and judging a perceiver, thereby serving as a
source of normative social influence on the perceiver. In this
section, we review previous research on these consequences
and provide novel data using the IDAQ to illustrate each one.
Moral Care and Concern
One of the most widely hypothesized consequences of anthro-
pomorphism is that it grants nonhuman agents moral regard,
conferring rights such as freedom and autonomy. ‘‘Anthropo-
morphizing nature allows it to be moralized’’ (Gebhard,
Nevers, & Billman-Mahecha, 2003, pp. 97–98), presumably
because of a general sentiment that ‘‘when moral worth is in
question, it is not a matter of actions which one sees but their
inner principles which one does not see’’ (Kant, 1785/1959,
p. 23). Bentham appeared to agree with Kant when he argued
that the key question for animal rights was not whether animals
were capable of certain behaviors (e.g., talking), but rather
‘‘can they suffer?’’ (Bentham & Browning, 1843, p. 143).
Indeed, emerging psychological evidence of higher order
mental experiences—such as sadness and depression—in
chimpanzees has fueled causes like the Great Ape Project
(Cavalieri & Singer, 1993), an organization that advocates for
the extension of basic legal rights to great apes. Our first set of
studies on the consequences of anthropomorphism therefore
examined whether those who tend to anthropomorphize nonhu-
man agents are also more likely to treat them as moral agents
worthy of empathic care and concern.
Table 1. Items From Study 1 and Factor Loadings
Item
Factor 1:
Pattern
coefficients
Factor 2:
Pattern
coefficients
Factor 1:
Structure
coefficients
Factor 2:
Structure
coefficients
tmind .615 .023 .623 .228
twill .756 .064 .735 .189
tintent .512 .047 .528 .218
tcon .523 .007 .521 .168
temo .596 .085 .568 .114
tdur .039 .166 .016 .153
tuse .205 .249 .122 .181
tgoodl .133 .155 .185 .199
tact .038 .222 .112 .235
tleth .261 .072 .285 .159
amind .029 .740 .276 .750
awill .002 .787 .265 .788
aintent .011 .695 .243 .699
acon .053 .746 .196 .728
aemo .040 .653 .258 .666
adur .026 .310 .130 .319
ause .008 .329 .118 .332
agoodl .068 .318 .174 .341
aact .029 .412 .109 .402
aleth .068 .124 .109 .147
nmind .708 .137 .754 .373
nwill .762 .018 .756 .237
nintent .775 .049 .759 .210
ncon .761 .027 .770 .281
nemo .736 .026 .745 .272
ndur .007 .390 .137 .392
nuse .165 .290 .068 .235
ngoodl .190 .254 .275 .317
nact .251 .287 .347 .371
nleth .292 .084 .320 .182
Note: Loadings >.45 are in bold. For each item, first letter indicates type of
agent (a ¼ animal, n ¼ nature, t ¼ technology) with attribute indicated by the
following code: mind ¼ mind, will ¼ free will, intent ¼ intentions, con ¼
consciousness, emo ¼ emotions, act ¼ active, leth ¼ lethargic, goodl ¼ good
looking, dur ¼ durable, and use ¼ useful.
Box 1. Alternate Measures of Fit for Two-Factor Model (Preliminary
Analysis) in Study 2
Sample discrepancy function value 4.12
Population discrepancy function value, Fo
bias adjusted point estimate
3.43 (90% CI: 3.18, .3.70)
Expected cross-validation index
point estimate (modified AIC)
4.27 (90% CI: 4.02, 4.53)
CVI (modified AIC) for the saturated model 1.53
Test statistic 2505.202
Degrees of freedom 420
Effective number of parameters 45
Note: AIC ¼ Akaike’s Information Criteria, CVI ¼ cross-validation index.
222 Waytz, Cacioppo, and Epley
222
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Sample discrepancy function value
Population discrepancy function value, Fo
2.98
1.53
Expected cross-validation index
CVI (modified AIC) for the saturated model
1661.23Test statistic
389modeerffoseergeD
2.732
2.092
76sretemarapforebmunevitceffE
awillaintentaconaemonmind
nwill
nintentncon
tmind twill tintent tcon temo amind
nemo
tmind 1
.41
.35
.44
.25
1
1
1
.18
1
1
1
1
1
1
1
1
1
1
1
.31
.25
.28
.48
.28
.21 .16 .07.18.14
.61.16 .15 .13.20.16
.52.45.12 .15 .05.12.14
.51.46.51.13 .15 .05.10.12
.46.55.46.39.09 .10 .08.14.12
.32.30.32.33.40.27 .28 .23.35.29
.54.22.18.24.29.21.37 .27 .28.36.33
.58.57.29.24.28.30.30.36 .30 .32.43.37
.56.53.59.25.27.26.25.25.33 .35 .22.30.29
.56.53.49.51.33.22.25.28.25.31 .27 .27.34.24
twill
tintent
tcon
temo
amind
awill
aintent
acon
aemo
nmind
nwill
nintent
ncon
nemo
Fig. 1. Alternate measures of fit and covariance matrix for two-factor model (secondary analysis) in Study 2.
AIC ¼ Akaike’s Information Criteria, CVI ¼ cross-validation index. For each item, first letter indicates type of
agent (a ¼ animal, n ¼ nature, t ¼ technology) with attribute indicated by the following code: mind ¼ mind,
will ¼ free will, intent ¼ intentions, con ¼ consciousness, emo ¼ emotions, act ¼ active, leth ¼ lethargic,
goodl ¼ good looking, dur ¼ durable, and use ¼ useful.
Who Sees Human 223
223
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Secondary emotions. We first examined whether anthropo-
morphism (measured by the IDAQ) predicts the attribution of
secondary emotions, a set of emotions that people commonly
consider to be uniquely human (Demoulin et al., 2004; Leyens
et al., 2003). Attributing secondary emotions to victims of a
natural disaster increases the desire to help those victims
(Cuddy, Rock, & Norton, 2007). People are also more likely
to attribute secondary emotions to ingroup members than to
commonly dehumanized outgroup members (e.g., Leyens
et al., 2003).
In this study, 40 individuals from the University of Chicago
population completed the IDAQ and then watched two short
videos: one of three kittens playing together and one of two
snakes fighting with each other (order counterbalanced). Parti-
cipants then rated the extent to which each animal experienced
10 primary emotions (pain, fear, panic, fright, surprise, suffer-
ing, anger, affection, attraction, and pleasure) and 10 secondary
emotions (admiration, resentment, shame, remorse, embarrass-
ment, guilt, hope, nostalgia, humiliation, and optimism). Prior
research has shown that people perceive these primary emo-
tions to be the least uniquely human emotions, and these sec-
ondary emotions to be the most uniquely human emotions
(Demoulin et al., 2004). Participants rated the experience of
these emotions on scales ranging from 1 (not at all)to7(very
much).
The IDAQ significantly predicted the attribution of
secondary emotions to the nonhuman animals, b ¼ .61, t(38)
¼ 4.71, p < .0001. This relationship held when controlling
for participants’ attribution of primary emotions to the stimuli,
b ¼ .49, t(37) ¼ 3.91, p < .0001.
Moral judgments. Given that the IDAQ predicts attributions
of complex emotional experience to nonhuman agents, we next
examined whether the IDAQ predicts moral judgments about
the treatment of these agents as well (Study 5). Kant (1785/
1959) described this basic moral principle of autonomy when
arguing that ‘‘every rational being exists as an end in himself
and not merely as a means to be arbitrarily used by this or that
will ... rational beings are called persons inasmuch as their
nature already marks them out as ends in themselves’’ (quoted
in Farah & Heberlein, 2007, p. 6). Dennett (1996) also noted
the centrality of mental states to ethical debates:
Some think it’s obvious that a ten-week-old fetus has a mind,
and others think it’s obvious that it does not. If it does not, then
the path is open to argue that it has no more interest than, say, a
gangrenous leg or an abscessed tooth—it can be destroyed to
save the life of (or just to suit the interests) of the mind-haver
of which it is a part. If it does already have a mind, then, what-
ever we decide, we obviously have to consider its interests
along with the interests of its temporary host. (p. 6)
If anthropomorphism involves attributing humanlike mental
states to nonhuman agents, then it should also predict the extent
to which people consider and respect a nonhuman agent’s inter-
ests and wellbeing. We examined this prediction in Study 5.
Fifty visitors to the Museum of Science and Industry in
Chicago volunteered to complete this study. They completed
the IDAQ (a ¼ .88) and then read a series of vignettes about
nonhuman stimuli (based on materials by Greene, Sommer-
ville, Nystrom, Darley, & Cohen, 2001). In the three moral
dilemmas, participants made judgments about the morality of
destroying IBM’s legendary chess-playing computer, ‘‘Deep
Blue’’ (3 ¼ absolutely morally wrong to þ3 ¼ absolutely
morally right), the appropriateness of leaving a bed of rare
flowers to be demolished (3 ¼ absolutely not to þ3 ¼ abso-
lutely yes), and the appropriateness of destroying a prized
motorcycle to save a human life (3 ¼ absolutely not to þ3
¼ absolutely yes). In two nonmoral dilemmas, participants
evaluated the ‘‘morality’’ of waiting to purchase a computer
at a lower price and replacing an ingredient of a cookie recipe.
As predicted, the IDAQ significantly predicted how wrong
participants reported it was to harm the computer, b ¼.47,
t(48) ¼ 3.64, p ¼ .001, the motorcycle, b ¼.38, t(48) ¼
2.84, p < .01, and the flowers, b ¼.33, t(48) ¼ 2.42, p <
.05. The IDAQ also significantly predicted evaluations of
wrongdoing for all moral scenarios when controlling indepen-
dently for judgment of the two nonmoral scenarios (all ps<
.025). Anthropomorphism did not significantly predict judg-
ments of the nonmoral scenarios (both ps > .50).
Environmental concern. The relationship between anthropo-
morphism and moral care may be especially clear and increas-
ingly important in how people view nature. Research has
demonstrated that empathizing with nature increases concern
for the environment (Gebhard et al., 2003; Schultz, 2000) and
that taking the perspective of a harmed animal increases envi-
ronmental concern (Sevillano, Aragones, & Schultz, 2007).
Cultures that anthropomorphize nature, such as the Guatemala
Itza Maya community that ascribes ‘‘spirits’’ to their rainforest
habitat, follow more sustainable ecological practices than do
other groups inhabiting the same area (Atran & Medin, 2008;
Table 2. Factor Loading Matrix for Secondary Factor Analysis in
Study 2
Item First-order factor Point estimate
tmind inanimate 0.499
twill inanimate 0.474
tintent inanimate 0.512
tcon inanimate 0.461
temo inanimate 0.378
amind animate 0.710
awill animate 0.746
aintent animate 0.713
acon animate 0.674
aemo animate 0.654
nmind inanimate 0.733
nwill inanimate 0.722
nintent inanimate 0.773
ncon inanimate 0.734
nemo inanimate 0.685
Note: For each item, the first letter indicates type of agent (a ¼ animal, n ¼ nature,
t ¼ technology) with attribute indicated by the following code:mind ¼ mind,will¼
free will, intent ¼ intentions, con ¼ consciousness, emo ¼ emotions, act ¼ active,
leth ¼ lethargic, goodl ¼ good looking, dur ¼ durable, and use ¼ useful.
224 Waytz, Cacioppo, and Epley
224
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Atran et al., 2002). And it is surely no accident that environ-
mentalists frequently refer to the planet as ‘‘mother earth.’’
In Study 6, we examined whether anthropomorphism,
(measured by the IDAQ) predicted environmental concern. In
this study, 52 adults completed the IDAQ in an online study
alongside a four-item measure of concern for the environment
(‘‘It upsets me when I hear about a forest being destroyed,’’
‘‘I am not very concerned with the well-being of nature,’’ ‘‘The
government should do more to prevent pollution of the environ-
ment,’’ ‘‘The protection of plants and trees is not very impor-
tant’’). Participants rated these items on a scale of 1 (strongly
disagree)to7(strongly agree) and we computed an environ-
mental concern score (a ¼ .67) from the mean of these items
(reverse-scored where appropriate). The IDAQ predicted
environmental concern, b ¼ .28, t(50) ¼ 2.12, p < .05, again
demonstrating a relationship between anthropomorphism and
moral care toward nonhuman agents in nature. Promoting
anthropomorphism of nature, as many have suggested, may
indeed be an effective way to increase concern for environmen-
tal issues such as global warming, air pollution, and water
contamination, whereas reducing anthropomorphism may
diminish concern.
Responsibility and Trust
Granting an agent mental capacities also means that the agent is
capable of autonomous self-directed behavior and can there-
fore be held responsible for its actions. Existing research
demonstrates that people more willingly punish an agent they
consider mindful (Gray et al., 2007), and corporations repre-
sented as single, personified agents may be held more legally
responsible for moral violations than corporations that are rep-
resented as collectives of disparate individuals (see French,
1986). In centuries past, legal practices even allowed for crim-
inal prosecution of nonhuman agents such as rodents and sta-
tues based on the belief that these agents were conscious
intentional actors (Berman, 1994; see also Sunstein &
Nussbaum, 2004). If the presence of a thoughtful humanlike
mind renders agents worthy of blame, then it may also render
agents worthy of trust when their competence is required. In
an age where technology is increasingly used to make critical
life or death decisions in medical settings, to make investment
decisions in stock market settings, or to catch liars in legal
settings, the extent to which people trust such technology is
becoming increasingly relevant. We predicted that those who are
especially likely to anthropomorphize nonhuman agents would
also be more likely to trust technology with important tasks.
To test this hypothesis, we asked 54 adults in Study 7 to
complete the IDAQ and then indicate whether they would trust
a human or a technological agent to predict heart attack risk,
detect when a person is lying, determine the best college foot-
ball team in the country, wash a fragile set of dishes, calculate
the cost of preventing air pollution, and select individuals to
admit to a university. For each decision, participants read a sce-
nario explaining the situation and then indicated whether they
would trust a human or technology with completing a particular
task. For instance, participants read that a person had been
accused of murder, read the details about the case, and then
reported whether they would trust a trained psychologist or a
polygraph machine to detect whether or not this suspect was
lying. Regressing both age and the IDAQ on a composite mea-
sure of participants’ trust in technology revealed a significant
predictive effect for the IDAQ, b ¼ .30, t(51) ¼ 2.29, p <
.05. Those more likely to anthropomorphize nonhuman agents
were also more likely to report that they would trust technolo-
gical agents to make important decisions. These findings are
consistent with an existing experiment in which people work-
ing collaboratively with a robot attributed more responsibility
for the overall work to the robot when they were led to anthro-
pomorphize the robot (Hinds, Roberts, & Jones, 2004) and
other experiments in which people rated anthropomorphized
agents as more credible and capable than nonanthropomor-
phized agents (Burgoon et al., 2000; Nowak & Rauh, 2005).
We believe these findings raise at least three very interesting
questions for future research. First, the data above come from
hypothetical scenarios without any real consequences for parti-
cipants’ responses. It is critical to examine whether these sce-
nario results can replicate in real and consequential decisions.
Second, if anthropomorphizing technology makes them appear
more competent and capable, then it may increase social loaf-
ing among people on tasks that require collaboration between
humans and nonhumans. Finally, if anthropomorphizing non-
human agents makes them appear more responsible for their
actions, then the humans controlling those agents may appear
less responsible themselves. Modern warfare, for instance, is
increasingly becoming a battle of technology in which harm
is done indirectly between humans through robots or other mil-
itary technology. Ron Arkin, a robotics expert, has noted that
‘‘it appears inevitable that increasing levels of autonomy will
be moved onto unmanned and robotic systems ...there are a
range of effects [that can occur]: difficulty of responsibility-
attribution in the event of war crimes, the potential lowering
of the threshold of entry into war, proliferation of the technol-
ogy into terrorist organizations, and many more’’ (Bennett,
2008). If robots in war, computers in admissions decisions, or
automobiles in accidents appear humanlike, does this decrease
the perceived responsibility of the people who programmed the
robots, wrote the computer algorithm, or drove the car during
possible instances of war crimes, racial discrimination, or vehi-
cular manslaughter?
Social Surveillance
Agents with humanlike minds may appear able to feel, think,
and control their own actions, but these mindful agents may
also evaluate, judge, and form impressions. In fact, these
anthropomorphized agents may be able to form impressions
of us. People are more likely to follow social norms—typically
behaving more desirably—when watched by other people than
when alone, in large part because people care deeply about
what others think of them and do their best to make a good
impression (Leary, 1995). Other mindful agents therefore
Who Sees Human 225
225
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
serve as sources of social influence. Does anthropomorphizing
a nonhuman agent—whether it be a robot, a pet, or a
god—increase adherence to socially desirable norms?
Some existing research is consistent with this possibility.
People with anthropomorphic representations of God believe
God to be more judgmental than those with less anthropo-
morphic representations (Morewedge & Clear, 2008). And reli-
gious systems that propose an omnipresent and judgmental
God appear better able to enhance cooperation between group
members, possibly because of the capacity for these gods to
watch people’s behavior at all times and serve as a constant
source of social surveillance (Norenzayan & Shariff, 2008).
People also present themselves more desirably to a computer
interface that has a human face than to one that is purely
text-based (Sproull, Subramani, Kiesler, Walker, & Waters,
1996), and they behave more cooperatively in an economic
game when humanlike eyes are presented on the computer
screen (Haley & Fessler, 2005). Those who are more likely
to anthropomorphize nonhuman agents may therefore behave
more desirably (or normatively) in the presence of those agents
than people who are less likely to anthropomorphize.
We tested this hypothesis (Study 8) by asking 38 partici-
pants to complete the IDAQ and then answer an eight-item ver-
sion of the Marlowe–Crowne scale of socially desirable
responding (Crowne & Marlowe, 1964; Ray, 1984), asked over
a computer interface by an easily anthropomorphized robot
named Kismet. For example, participants would see the robot
appear on the screen asking, ‘‘Have there been occasions when
you have taken advantage of someone?’’ and would then
respond ‘‘yes’’ or ‘‘no.’’ As expected, the IDAQ significantly
predicted socially desirable responding, b ¼ .42, t(36) ¼
2.77, p < .01. These results suggest that anthropomorphism
may increase the social influence of nonhuman agents. Being
watched by others matters, perhaps especially when others
have a mind like one’s own.
Anthropomorphism: A Central Concept
Within a Hub Discipline
We have thus far provided a reliable measure of anthropo-
morphism and provided evidence that this measure matters for
some behaviors that psychologists care a great deal about,
including judgments about the emotions and mental capacities
of other agents, the degree of trust placed in these agents, and
the potential influence of these agents on one’s own behavior.
We believe anthropomorphism matters, however, not simply
for psychology but for disciplines far beyond psychology as
well. Over the past decade, psychology has emerged as a hub
discipline, functioning as one of several core academic
domains through which other disciplines communicate and
connect (Cacioppo, 2007). Rather than operating as an insular
field of study, psychology is highly interdisciplinary and
capable of informing these multiple other fields, reflected in
the degree to which these other fields cite psychological
research and theory (Boyack, Klavans, & Bo¨rner, 2005). Given
anthropomorphism’s consequences for moral concern,
perceived responsibility, and social surveillance, we believe
understanding it can provide insight into adjacent domains that
care about these topics as well. Here, we focus on three in
particular: human–computer interaction, business, and law.
Human–Computer Interaction
Anthropomorphism is directly relevant to human–computer
interaction, a domain that encompasses artificial intelligence,
computer science, and engineering. Recent work in artificial
intelligence has produced robots with traces of the most sophis-
ticated of human capacities, with further advances in creating
humanlike technology becoming increasingly dependent on
psychology. In turn, psychologists have now begun to speculate
about the challenges of increased human–android interaction in
the next 50 years (Roese & Amir, 2009). Within the past
decade alone, engineers have developed robots that can express
emotion (Breazeal & Aryananda, 2002), recognize emotional
and social cues (Breazeal, 2002), and even imitate human
action and behave interdependently (Breazeal & Scassellatti,
2002).
Although people anthropomorphize in varying degrees,
these humanlike agents seem to induce at least some anthropo-
morphism quite readily in most people. One recent neuroima-
ging study demonstrated the same neural circuitry underlying
the perception of human behavior and that of an anthropomor-
phized robot (Gazzola et al., 2007). Not only do people per-
ceive robots to be humanlike, but people appear to behave
toward technological agents following the same social conven-
tions and rules as when interacting with other humans (Nass &
Moon, 2000). Capitalizing on this tendency, engineers now
routinely design the front side of motorcycles and automobiles
to resemble ‘‘faces’’ in order to convey particular impressions
(Taylor, 2008).
As computer scientists, robotics developers, and engineers
have begun to identify anthropomorphism’s effects on human
interaction with technology, understanding the determinants
of anthropomorphism can identify the conditions under which
these effects will be most potent. In many cases, anthropo-
morphism appears to enhance human–computer interaction.
One study has demonstrated that anthropomorphizing an alarm
clock and a robot (as well as a dog and a series of shapes)
makes these agents appear more understandable and predict-
able (Waytz et al., 2009). Other studies demonstrate that
anthropomorphic avatars appear more intelligent (Koda &
Maes, 1996) and more credible (Nowak & Rauh, 2005) than
nonanthropomorphic ones. Anthropomorphic computer inter-
faces tend to increase engagement (Nass, Moon, Fogg, Reeves,
& Dryer, 1995), and appear more effective in collaborative
decision-making tasks (Burgoon et al., 2000). People also
like robots more when they express emotions in a more human-
like fashion (Siino, Chung, & Hinds, 2008). Anthropomorphic
companion robots also provide social support for the
elderly, improving both physical and mental health (Banks,
Willoughby, & Banks, 2008; Melson, Kahn, Beck, &
Friedman, 2009).
226 Waytz, Cacioppo, and Epley
226
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Although anthropomorphized technology increases
engagement and perceived intelligence, these advances can
have some undesirable side effects as well. Certain anthropo-
morphic computer ‘‘assistants,’’ such as the Microsoft Word
paperclip, are strongly disliked because they seem very
distracting, much like a insensitive colleague who pops in to
one’s office far too often (Shneiderman, 1995; Swartz, 2003).
The uncanny valley hypothesis (Mori, 1970) also suggests that
robots that look too humanlike actually repulse and discomfort
users (MacDorman, Green, Ho, & Koch, 2009). Beyond
appearance, the enhanced degree of responsibility afforded to
anthropomorphic agents presents some problems as well.
People are more likely to treat anthropomorphic interfaces as
scapegoats when the technology malfunctions (Serenko,
2007), and they feel less responsible for success on tasks that
use humanlike interfaces (Quintanar, Crowell, & Pryor,
1982). Anthropomorphism can also generate inappropriate
expectations for how computers and robotics are capable of
behaving (DiSalvo & Gemperle, 2003; Shneiderman, 1980).
Some research has attempted to address these concerns by
proposing an optimal level of anthropomorphism for robotics
design (Duffy, 2003). The present research does not necessarily
offer prescriptive claims for the anthropomorphism of
technology, but it does help determine when and for whom
anthropomorphism’s effects are most likely to occur. Computer
scientists, robotics developers, and engineers can use this
research in their efforts to optimize technology by focusing
on the consequences of anthropomorphism and also identifying
the people that are most prone to these consequences.
Business: Marketing and Finance
Just as engineers humanize technology, advertisers continue to
humanize a wide array of products, and marketing is one of two
business-related domains (along with financial decision
making) that anthropomorphism can inform. Marketers have
long provided anthropomorphic representations of products
ranging from Kool-Aid to condoms to car parts with consider-
able success (Aggarwal & McGill, 2007; Arnheim, 1969; Biel,
2000). Brand ‘‘personalities’’ influence consumer decision
making because individuals often attempt to utilize these per-
sonalities to express their own self-concepts (Aaker, 1997).
Specific humanlike cues, such as an apparent smile in the grill
of a car, can also enhance product evaluations if consumers are
already primed with an anthropomorphic schema (Aggarwal &
McGill, 2007). The anthropomorphic appearance of a product
(such as a watch that appears to be smiling when its hands are
set to 10:10) can increase liking of that product as well (Labroo,
Dhar, & Schwarz, 2008). Given people’s natural attentiveness
to humanlike cues, anthropomorphism provides an effective
way to increase attention to advertising. Studying variation in
anthropomorphism can determine who is likely to be influ-
enced by these campaigns and how to make them more (or less)
effective (for better or worse).
Equally powerful is the effect of anthropomorphism on
the interpretation of the complex and unpredictable working
of financial markets. In one study, for instance, the
anthropomorphic emotions evoked by particular market sectors
predicted investors’ willingness to invest in those sectors
(MacGregor, Slovic, Dreman, & Berry, 2000). In another,
describing the stock market in anthropomorphic terms (as
opposed to mechanistic terms) increased predictions that price
trends would continue (Morris, Sheldon, Ames, & Young,
2007). In a third, the higher people scored on the IDAQ, the
more they predicted stock market trends to continue, as if
guided by the stable intentions or goals of a mindful agent
(Caruso, Waytz, & Epley, 2010). Adam Smith’s metaphor of
the ‘‘invisible hand’’ may have more literal consequences for
investor decision making than he would have guessed. Practi-
tioners and researchers working at the intersection of psychol-
ogy and economics—generally called behavioral economics—
can benefit from understanding how the anthropomorphic
depictions of financial systems interact with the presentation
of more objective data (e.g., stock prices) to affect economic
behavior.
Law
Anthropomorphism’s implications for an agent’s moral status
have immediate relevance to legal practice. Not only do judg-
ments of guilt or innocence center on whether the agent in
question is capable of intentional action, but legal decisions
about an agent’s rights rest on that agent’s perceived mental
capabilities as well. Animal rights is perhaps the most obvious
legal issue relevant to anthropomorphism. Debates over
whether animals can be used in biomedical research, whether
animals should be treated as property, and whether it is
acceptable to eat certain animals all center on these agents’
mental similarity to humans (Hauser, Cushman, & Kamen,
2006; Morton, Burghardt, & Smith, 1990). Recently, Spain’s
lower house of parliament supported a manifesto granting
human rights—‘‘life, liberty, and freedom from physical and
psychological torture’’—to great apes. Spanish congress-
woman Joan Herrera justified this decision by noting that these
animals are ‘‘capable of recognizing themselves, and have cog-
nitive capabilities.’’ Marta Tafalla, a law professor specializing
in animal rights, added, ‘‘They are animals with highly devel-
oped intelligence and emotional capacity’’ (Abend, 2008). Psy-
chological research on anthropomorphism may not to be able to
make such definitive claims about the humanness of various
agents, but it can determine the conditions under which people
are most likely to represent these agents as humanlike, and
what consequences such inferences might have on people’s
behavior toward those agents.
Equally complex legal decisions concern the rights of
humans with ambiguous or incomplete capacities such as a
12-week old fetus, a brain-damaged individual, or a diagnosed
sociopath. Topics ranging from abortion, to capital punish-
ment, to euthanasia, to torture center on the humanness of a
particular agent to determine whether the agent deserves
fundamental human rights. Anthropomorphism may power-
fully influence people’s judgments on these critical issues.
Who Sees Human 227
227
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Psychological research on anthropomorphism can contribute to
the domain of law by identifying when the attribution of human
rights is most likely to occur and identifying the critical precon-
ditions for perceiving humanlike mental capacities in other
agents.
Concluding Thoughts: Implications for
Person Perception
The exponential increase in natural, biological, and manufac-
tured nonhuman agents in the 21st century makes it increas-
ingly important to study how people understand and treat
these agents. Anthropomorphism provides a far-reaching con-
struct for studying how people interact with agents ranging
from pets that provide social companionship, to typhoons that
decimate entire cities, to robots that perform open-heart sur-
gery. The research presented here examines the existence of
stable individual differences in anthropomorphism and uses
those differences to identify important consequences of anthro-
pomorphism for psychology and related disciplines. This
research complements the most recent theoretical treatment
of anthropomorphism (Epley et al., 2007), and expands on
empirical demonstrations of predictable variability in anthro-
pomorphism across situations (e.g., Epley, Akalis, Waytz, &
Cacioppo, 2008), personality types (e.g., Epley, Waytz, Akalis,
& Cacioppo, 2008), developmental stages (e.g., Carey, 1985),
and cultures (e.g., Asquith, 1986; Medin & Atran, 2004; Wax-
man & Medin, 2007). Understanding how these situational,
biological, and cultural factors work in concert to create reli-
able individual differences in anthropomorphism is a very
interesting and relatively unexplored topic for future research.
Another interesting topic is the relation between the explicit
measure of anthropomorphism we have provided here and
more implicit manifestations of anthropomorphism that may
be reflected in people’s behavior but that may not be con-
sciously accessible.
Although the present article focuses on anthropomorphism’s
effects on perceptions of nonhuman agents, the tendency to
anthropomorphize should also influence evaluations of other
humans. Humanness exists on a continuum such that individu-
als can attribute humanlike capacities to nonhuman agents
through anthropomorphism and can also fail to attribute these
same capacities to other people through dehumanization. The
antecedents and consequences of anthropomorphism and dehu-
manization may be closely linked (Epley et al., 2007; Kwan &
Fiske, 2008; Waytz, Epley, & Cacioppo, 2010), and recent
empirical work suggests that the same factors that increase
anthropomorphism may likewise influence dehumanization.
For example, just as an agent’s similarity to humans increases
anthropomorphism (Morewedge et al., 2007), those who seem
very different from the prototypical human are also the most
likely to be dehumanized (Harris & Fiske, 2006). Those who
are socially connected are less likely than those who are lonely
to anthropomorphize nonhuman agents (Epley et al., 2008), and
those who are socially connected also appear more likely to
dehumanize other humans (Waytz & Epley, 2009). Even the
moral rights and responsibilities granted to humanized agents
may be the same ones that are denied to people who are dehu-
manized (Waytz et al., 2010). Understanding individual differ-
ences in anthropomorphism not only seems important for
identifying who is likely to treat nonhuman agents as human-
like, but also for identifying who is likely to treat other humans
as animals or objects.
Dehumanization has equivalent and opposite implications
of anthropomorphism for moral treatment of an agent. Anthro-
pomorphism increases moral concern, whereas dehumaniza-
tion increases moral disengagement that can license immoral
action toward others (Bandura, Barbaranelli, Caprara, &
Pastorelli, 1996). For instance, dehumanization increases aggres-
sion toward individuals and groups (Bandura, Underwood, &
Fromson, 1975; Struch & Schwarz, 1989), endorsement of
discrimination toward racial outgroups (Goff, Eberhardt,
Williams, & Jackson, 2008), general negative attitudes toward
outgroups (Hodson & Costello, 2007), and justification for past
wrongdoing toward outgroups (Castano & Giner-Sorolla,
2006). Dehumanization may similarly decrease attributions of
responsibility and trust or diminish perceptions of social surveil-
lance, and these domains are ripe for future research to address. In
identifying the structure of individual differences in anthropo-
morphism and consequences of the tendency to ‘‘see human,’’ the
present research should contribute to an understanding of these
well-established topics within person perception just as it contri-
butes to the burgeoning study of nonperson perception.
Notes
1. Across studies, IDAQ-NA did not consistently constitute an intern-
ally reliable measure and we thus do not report internal consistency
for this measure. This lack of internal reliability is expected
because these items were developed simply to measure a diffuse set
of nonanthropomorphic attributions rather than a single coherent
construct. Analyses involving the IDAQ-NA thus appear as ancil-
lary results at https://sites.google.com/site/idaqmaterials/
Acknowledgments
We thank Ashley Angulo, Adrianna Guerrero, Mina Kang, Jasmine
Kwong, Ye Li, Paul Thomas, Rebecca White, and Louise Hawkley for
their assistance, and the Templeton Foundation and Booth School of
Business for financial support.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect
to their authorship or the publication of this article.
228 Waytz, Cacioppo, and Epley
228
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Appendix
Scale Development
We generated items for a preliminary version of the IDAQ
by first identifying four classes of commonly anthropomor-
phized agents—nonhuman animals, natural entities, spiritual
agents, and technological devices—and pairing each class of
agent with five anthropomorphic and five nonanthropomorphic
traits. The nonanthropomorphic traits consisted of qualities
related to clearly observable or functional features of a
stimulus (the extent to which a stimulus is ‘‘durable,’’ ‘‘use-
ful,’’ ‘‘good-looking,’’ ‘‘active,’’ and ‘‘lethargic’’). The nonan-
thropomorphic items are not part of the IDAQ, and we simply
included these items to dissociate anthropomorphism from
dispositional attribution more generally and to ensure that
differences in anthropomorphism did not merely reflect
differences in scale use.
The mental state attributes used in items for scale develop-
ment (the extent to which a stimulus has ‘‘a mind of its own,’’
‘‘has free will,’’ ‘‘has consciousness,’’ ‘‘has intentions,’’ and
‘‘can experience emotions’’) reflect properties captured in pre-
viously used measures of attribution of human uniqueness and
higher order cognition to human targets (e.g., Demoulin et al.,
2004; Haslam et al., 2005; Kozak, Marsh, & Wegner, 2006).
This method of item generation yielded 40 items—20 assessing
anthropomorphism and 20 unrelated to anthropomorphism—
that we then reduced to 30 items, as shown in Box A1. The
10 items related to spiritual agents ultimately were not used
as part of the IDAQ based on the results of Study 1.
References
Aaker, J.L. (1997). Dimensions of brand personality. Journal of Mar-
keting Research, 34, 347–356.
Abend, L. (2008, Ju ly 18). In Spain, human rights for apes [Elec tronic ver-
sion]. Time. Retrieved August 27, 2008, from http://www.time.com/
time/world/article/0,8599,1824206,00 .html?imw¼Y
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, 468–479.
Anggoro, F.K., Waxman, S.R., & Medin, D.L. (2008). Naming prac-
tices and the acquisition of key biological concepts: Evidence from
English and Indonesian. Psychological Science, 19, 314–319.
Arnheim, R. (1969). Visual thinking. Berkeley: University of
California Press.
Asquith, P.J. (1986). Anthropomorphism and the Japanese and West-
ern traditions in primatology. In J.G. Else & P.C. Lee (Eds.), Pri-
mate ontogeny, cognition, and social behavior (pp. 61–71).
Cambridge, United Kingdom: Cambridge University Press.
Associated Press. (2007, February 9). GM changes robot suicide ad.
CNN. Retrieved August 27, 2008, from http://money.cnn.com/
2007/02/09/news/companies/gm_robotad/
Atran, S., & Medin, D. L. (2008). The native mind and the cultural
construction of nature. Cambridge, MA: MIT Press.
Atran, S., Medin, D., Vapnarsky, V., Ucan Ek’, E., Coley, J.D.,
Timura, C., & Baran, M. (2002). Folkecology, cultural epidemiol-
ogy, and the spirit of the commons: A garden experiment in the
Maya lowlands, 1995–2000. Current Anthropology, 43, 421–450.
Auter, P.J. (1992). TV that talks back: An experimental validation of a
parasocial interaction scale. Journal of Broadcasting & Electronic
Media, 36, 173–181.
Box A1. All IDAQ Items
1. To what extent is the desert lethargic?
2. To what extent is the average computer active?
3. To what extent does technology—devices and machines
for manufacturing, entertainment, and productive pro-
cesses (e.g., cars, computers, television sets)—have
intentions?
4. To what extent does the average fish have free will?
5. To what extent is the average cloud good-looking?
6. To what extent are pets useful?
7. To what extent does the average mountain have free
will?
8. To what extent is the average amphibian lethargic?
9. To what extent does a television set experience
emotions?
10. To what extent is the average robot good-looking?
11. To what extent does the average robot have
consciousness?
12. To what extent do cows have intentions?
13. To what extent does a car have free will?
14. To what extent does the ocean have consciousness?
15. To what extent is the average camera lethargic?
16. To what extent is a river useful?
17. To what extent does the average computer have a mind
of its own?
18. To what extent is a tree active?
19. To what extent is the average kitchen appliance useful?
20. To what extent does a cheetah experience emotions?
21. To what extent does the environment experience
emotions?
22. To what extent does the average insect have a mind of
its own?
23. To what extent does a tree have a mind of its own?
24. To what extent is technology—devices and machines for
manufacturing, entertainment, and productive processes (e.g.,
cars, computers, television sets)—durable?
25. To what extent is the average cat active?
26. To what extent does the wind have intentions?
27. To what extent is the forest durable?
28. To what extent is a tortoise durable?
29. To what extent does the average reptile have
consciousness?
30. To what extent is the average dog good-looking?
Note: IDAQ items are bolded. All items are rated on a 0 (not at all)to10(very
much) scale. To compute the IDAQ response score, sum items 3, 4, 7, 9, 11, 12,
13, 14, 17, 20, 21, 22, 23, 26, 29. To compute IDAQ-NA, sum items 1, 2, 5, 6, 8,
10, 15, 16, 18, 19, 24, 25, 27, 28, 30. Unused items with spiritual agents are: ‘‘To
what extent does a spirit (or spirits) have a mind of its own?’’, ‘‘To what extent
does a ghost have free will?’’, ‘‘To what extent do supernatural beings have
intentions?’’, ‘‘To what extent does the average spiritual agent have conscious-
ness?’’, ‘‘To what extent does a god experience emotions?’’, ‘‘To what extent
are deities durable?’’, ‘‘To what extent is a god useful?’’, ‘‘To what extent is the
average supernatural being good-looking?’’, ‘‘To what extent is a spirit (or
spirits) active?’’, and ‘‘To what extent is the average spiritual agent lethargic?’’
Who Sees Human 229
229
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Bandura, A., Barbaranelli, C., Caprara, G.V., & Pastorelli, C. (1996).
Mechanisms of moral disengagement in the exercise of moral
agency. Journal of Personality and Social Psychology, 71,
364–374.
Bandura, A., Underwood, B., & Fromson, M.E. (1975). Disinhibition
of aggression through diffusion of responsibility and dehumaniza-
tion of victims. Journal of Research in Personality, 9, 253–269.
Banks, M.R., Willoughby, L.M., & Banks, W.A. (2008). Animal–
assisted therapy and loneliness in nursing homes: Use of robotic
versus living dogs. Journal of the American Medical Directors
Association, 9, 173–177.
Barrett, J.L., & Keil, F.C. (1996). Conceptualizing a non–natural
entity: Anthropomorphism in God concepts. Cognitive Psychol-
ogy, 31, 219–247.
Bennett, J.R. (2008, October 1). Warbots: Latest in military strategy
[Electronic version]. International Relations and Security Net-
work. Retrieved February 14, 2009, from http://www.isn.ethz.ch/
isn/layout/set/print/content/view/full/73?id¼92220&lng¼en&
ots591¼4888CAA0-B3DB-1461-98B9-E20E7B9C13D4
Bentham, J., & Browning, J. (1843). The works of Jeremy Bentham.
London: Simpkin, Marshall, & Co.
Bering, J. (2006). The folk psychology of souls. Behavioral and Brain
Sciences, 29, 453–462.
Berman, P.S. (1994). Rats, pigs, and statues on trial: The creation of
cultural narratives in the prosecution of animals and inanimate
objects. NYU Law Review, 69, 288–326.
Biel, A. (2000). Converting image into equity. In D.A. Aaker &
A.L. Biel (Eds.), Brand equity and advertising: Advertising’s role
in building strong brands (pp. 67–82). Hillsdale, NJ: Erlbaum.
Boyack, K.W., Klavans, R., & Bo¨rner, K. (2005). Mapping the back-
bone of science. Scientometrics, 64, 351–374.
Breazeal, C. (2002). Regulation and entrainment in human-robot
interaction. International Journal of Robotics Research, 21,
1–20.
Breazeal, C., & Aryananda, L. (2002). Recognizing affective intent in
robot directed speech. Autonomous Robots, 12, 83–104.
Breazeal, C., & Scassellatti, B. (2002). Robots that imitate humans.
Trends in Cognitive Science, 6, 481–487.
Burgoon, J.K., Bonito, J.A., Bengtsson, B., Cederberg, C.,
Lundeberg, M., & Allspach, L. (2000). Interactivity in human–
computer interaction: A study of credibility, understanding, and
influence. Computers in Human Behavior, 16, 553–574.
Cacioppo, J.T. (2007, September). Psychology is a hub science.
Observer, 20
(8), pp. 5 & 42.
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA:
Bradford Books, MIT Press.
Caruso, E., Waytz, A., & Epley, N. (2010). Perceiving intentions
makes streaks seem likely to continue. Manuscript submitted for
publication.
Castano, E., & Giner-Sorolla, R. (2006). Not quite human:
Infra-humanization as a response to collective responsibility for
intergroup killing. Journal of Personality and Social Psychology,
90, 804–818.
Castelli, F., Frith, C., Happe´, F., & Frith, U. (2002). Autism and brain
mechanisms for the attribution of mental states to animated shapes.
Brain, 125, 1839–1849.
Cavalieri, P., & Singer, P. (Eds.). (1993). The great ape project:
Equality beyond humanity. New York: St. Martin’s Press.
Cheney, D., & Seyfarth, R. (1990). How monkeys see the world.
Chicago, IL: University of Chicago Press.
Chin, M.G., Sims, V.K., Ellis, L.U., Yordon, R.E., Clark, B.R.,
Ballion, T., et al. (2005). Developing an anthropomorphic
tendencies scale. Human Factors and Ergonomics Society Annual
Meeting Proceedings: Individual Differences in Performance, 3,
1266–1268.
Crowne, D.P., & Marlowe, D. (1964). The approval motive: Studies in
evaluative dependence. New York: Wiley.
Cuddy, A.J.C., Rock, M., & Norton, M.I. (2007). Aid in the aftermath
of Hurricane Katrina: Inferences of secondary emotions and inter-
group helping. Group Processes and Intergroup Relations, 10,
107–118.
Demoulin, S., Leyens, J.P., Paladino, M.P., Rodriguez, R.T.,
Rodriguez, A.P., & Dovidio, J.F. (2004). Dimensions of
‘‘uniquely’’ and ‘‘nonuniquely’’ human emotions. Cognition &
Emotion, 18, 71–96.
Dennett, D.C. (1978). Brainstorms: Philosophical essays on mind and
psychology. Cambridge, MA: MIT Press.
Dennett, D.C. (1996). Kinds of minds. New York: Basic Books.
DiSalvo, C., & Gemperle, F. (2003). From seduction to fulfillment:
The use of anthropomorphic form in design. In B. Hanington &
J. Forlizzi (Eds.), Proceedings of the Designing Pleasurable
Products and Interfaces Conference (pp. 67–72). Pittsburgh, PA:
Carnegie Mellon.
Duffy, B.R. (2003). Anthropomorphism and the social robot. Robotics
and Autonomous Systems, 42, 177–190.
Epley, N., Akalis, S., Waytz, A., & Cacioppo, J.T. (2008). Creating
social connection through inferential reproduction: Loneliness and
perceived agency in gadgets, gods, and greyhounds. Psychological
Science, 19, 114–120.
Epley, N., & Waytz, A. (2009). Mind perception. In S.T. Fiske, D.T.
Gilbert, & G. Lindzey (Eds.), The handbook of social psychology
(5th ed., pp. 498–541). New York: Wiley.
Epley, N., Waytz, A., Akalis, S., & Cacioppo, J.T. (2008). When I
need a human: Motivational determinants of anthropomorphism.
Social Cognition, 26, 143–155.
Epley, N., Waytz, A., & Cacioppo, J.T. (2007). On seeing human: A
three-factor theory of anthropomorphism. Psychological Review,
114, 864–886.
Farah, M. J., & Heberlein, A. S. (2007). Personhood and neuroscience:
Naturalizing or nihilating? American Journal of Bioethics
, 7,
37–48.
French, P.A. (1986). Principles of responsibility, shame, and the cor-
poration. In H. Curtler (Ed.), Shame, responsibility, and the corpo-
ration (pp. 17–55). New York: Haven.
Gazzola, V., Rizzolatti, G., Wicker, B., & Keysers, C. (2007). The
anthropomorphic brain: The mirror neuron system responds to
human and robotic actions. NeuroImage, 35, 1674–1684.
Gebhard, U., Nevers, P., & Billman-Mahecha, E. (2003). Moralizing
trees: Anthropomorphism and identity in children’s relationships
to nature. In L. Warden Clayton, S. Clayton, & S. Opotow (Eds.),
Identity and the natural environment (pp. 91–112). Cambridge,
MA: MIT Press.
230 Waytz, Cacioppo, and Epley
230
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
Goff, P.A., Eberhardt, J.L., Williams, M., & Jackson, M.C. (2008).
Not yet human: Implicit knowledge, historical dehumanization,
and contemporary consequences. Journal of Personality and
Social Psychology, 94, 292–306.
Gray, H.M., Gray, K., & Wegner, D.M. (2007). Dimensions of mind
perception. Science, 315, 619.
Greene, J.D., Sommerville, R.B., Nystrom, L.E., Darley, J.M., &
Cohen, J.D. (2001). An fMRI investigation of emotional
engagement in moral judgment. Science, 293, 2105–2108.
Guthrie, S. (1993). Faces in the clouds. Oxford, United Kingdom:
Oxford University Press.
Haley, K.J., & Fessler, D.M.T. (2005). Nobody’s watching? Subtle
cues affect generosity in an anonymous economic game. Evolution
and Human Behavior, 26, 245–256.
Harris, L.T., & Fiske, S.T. (2006). Dehumanizing the lowest of the
low: Neuroimaging responses to extreme out-groups. Psychologi-
cal Science, 17, 847–853.
Harris, L.T., & Fiske, S.T. (2008). Brooms in Fantasia: Neural corre-
lates of anthropomorphizing objects. Social Cognition, 26,
209–222.
Haslam, N., Bain, P., Douge, L., Lee, M., & Bastian, B. (2005). More
human than you: Attributing humanness to self and others. Journal
of Personality and Social Psychology, 89, 937–950.
Hauser, M.D. (2000). Wild minds: What animals really think. New
York: Holt.
Hauser, M.D., Cushman, F., & Kamen, M. (2006). People, pets, or
property? Lafayette, IN: Purdue University Press.
Heberlein, A.S., & Adolphs, R. (2004). Impaired spontaneous
anthropomorphizing despite intact perception and social knowl-
edge. Proceedings of the National Academy of Sciences, USA,
101, 7487–7491.
Hinds, P.J., Roberts, T.L., & Jones, H. (2004). Whose job is it any-
way? A study of human–robot interaction in a collaborative task.
Human–Computer Interaction, 19, 151–181.
Hodson, G., & Costello, K. (2007). Interpersonal disgust, ideological
orientations, and dehumanization as predictors of intergroup
attitudes. Psychological Science, 18, 691–698.
Hume, D. (1957). The natural history of religion. Stanford, CA:
Stanford University Press. (Original work published 1757).
Johnson, S.C., Slaughter, V., & Carey, S. (1998). Whose gaze
will infants follow? The elicitation of gaze following in
12-month-olds. Developmental Science, 1, 233–238.
Kant, I. (1959). Foundations of the metaphysics of morals
(L.W. Beck, Trans.). New York: Macmillan. (Originally published
1785).
Kiesler, S., & Goetz, J. (2002, April). Mental models and cooperation
with robotic assistants. Proceedings of the SIGCHI conference on
human factors in computing systems (pp. 576–577). Minneapolis,
MN: ACM Press.
Koda, T., & Maes, P. (1996). Agents with faces: The effect of perso-
nification. Proceedings of IEEE. Workshop on Robot and Human
Communication, 5, 189–194.
Kozak, M.J., Marsh, A.A., & Wegner, D.M. (2006). What do I think
you’re doing? Action identification and mind attribution. Journal
of Personality and Social Psychology, 90, 543–555.
Kwan, V.S.Y., & Fiske, S.T. (2008). Missing links in social cognition:
The continuum from nonhuman agents to dehumanized humans.
Social Cognition, 26, 125–128.
Labroo, A.A., Dhar, R., & Schwarz, N. (2008). Of frowning watches
and frog wines: Semantic priming, perceptual fluency, and brand
evaluation. Journal of Consumer Research, 34, 819–31.
Leary, M.R. (1995). Self-presentation: Impression management and
interpersonal behavior. Madison, WI: Brown & Benchmark.
Lesher, J.H. (1992). Xenophanes of Colophon: Fragments. Toronto,
Canada: University of Toronto Press.
Leyens, J.P., Cortes, B.P., Demoulin, S., Dovidio, J., Fiske, S.T., &
Gaunt, R., et al. (2003). Emotional prejudice, essentialism, and
nationalism. European Journal of Social Psychology, 33, 703–717.
Locke, J. (1997). An essay concerning human understanding. Har-
mondsworth, United Kingdom: Penguin Books. (Original work
published 1841).
MacDorman, K.F., Green, R.D., Ho, C.-C., & Koch, C. (2009). Too
real for comfort: Uncanny responses to computer generated faces.
Computers in Human Behavior, 25, 695–710.
MacGregor, D.G., Slovic, P., Dreman, D., & Berry, M. (2000). Ima-
gery, affect, and financial judgment. Journal of Psychology and
Financial Markets, 1, 104–110.
Medin, D.L., & Atran, S. (2004). The native mind: Biological categor-
ization, reasoning and decision making in development and across
cultures. Psychological Review, 111, 960–983.
Melson, G.F., Kahn, P.H., Jr., Beck, A.M., & Friedman, B. (2009).
Robotic pets in human lives: Implications for the human-animal
bond and for human relationships with personified technologies.
Journal of Social Issues, 65 , 545–567.
Mithen, S. (1996). The prehistory of the mind. London: Thames &
Hudson.
Morewedge, C.K., & Clear, M.E. (2008). Anthropomorphic God con-
cepts engender moral judgment. Social Cognition, 26, 181–188.
Morewedge, C.K., Preston, J., & Wegner, D.M. (2007). Timescale
bias in the attribution of mind. Journal of Personality and Social
Psychology, 93, 1–11.
Mori, M. (1970). The uncanny valley. Energy, 7, 33–35.
Morris, M.W., Sheldon, O.J., Ames, D.R., & Young, M.J. (2007).
Metaphors and the market: Consequences and preconditions
of agent and object metaphors in stock market commentary.
Organizational Behavior and Human Decision Processes, 102,
174–192.
Morton, D.B., Burghardt, G.M., & Smith, J.A. (1990). Critical anthro-
pomorphism, animal suffering, and the ecological context. In
S. Donnelley & K. Nolan (Eds.), Animals, science, and ethics
(pp. 13–19). Garrison, NY: Hastings Center.
Nass, C., & Moon, Y. (2000). Machines and mindlessness: Social
responses to computers. Journal of Social Issues, 56, 81–103.
Nass, C., Moon, Y., Fogg, B.J., Reeves, B., & Dryer, D.C. (1995). Can
computer personalities be human personalities? International
Journal of Human–Computer Studies, 43, 223–239.
Norenzayan, A., & Shariff, A.F. (2008). The origin and evolution of
religious prosociality. Science, 322, 58–62.
Nowak, K.L., & Rauh, C. (2005). The influence of the avatar on online
perceptions of anthropomorphism, androgyny, credibility,
Who Sees Human 231
231
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from
homophily, and attraction. Journal of Computer-Mediated Com-
munication, 11(1), article 8.
Paloutzian, R., & Ellison, C. (1982). Loneliness, spiritual well-being
and the quality of life. In L. Peplau & D. Perlman (Eds.), Loneli-
ness: A sourcebook of current theory, research and therapy (pp.
224–237). New York: John Wiley and Sons.
Preacher, K.J., & MacCallum, R.C. (2003). Repairing Tom Swift’s
electric factor analysis machine. Understanding Statistics, 2, 13–43.
Quintanar, L.R., Crowell, C.R., & Pryor, J.B. (1982). Human-
computer interaction: A preliminary social psychological analysis.
Behavior Research Methods and Instrumentation, 14, 210–220.
Ray, J.J. (1984). The reliability of short social desirability scales.
Journal of Social Psychology, 123, 133–134.
Roese, N.J., & Amir, E. (2009). Speculations on human–android
interaction in the near and distant future. Perspectives on
Psychological Science, 4, 429–434.
Scholl, B.J., & Tremoulet, P.D. (2000). Perceptual causality and
animacy. Trends in Cognitive Sciences, 4, 299–309.
Schultz, P.W. (2000). Empathizing with nature: The effects of
perspective taking on concern for environmental issues. Journal
of Social Issues, 56, 391–406.
Serenko, A. (2007). Are interface agents scapegoats? Attributions of
responsibility in human-agent interaction. Interacting With
Computers, 19, 293–303.
Sevillano, V., Aragones, J., & Schultz, P.W. (2007). Perspective
taking, environmental concern, and the moderating role of disposi-
tional empathy. Environment and Behavior , 35 , 685–705.
Shneiderman, B. (1980). Software psychology: Human factors in
computer and information systems. Cambridge, MA: Winthrop
Publishers.
Shneiderman, B. (1995). Looking for the bright side of user interface
agents. Interactions, 2, 13–15.
Siino, R., Chung, J., & Hinds, P. (2008). Colleague vs. tool: Effects of
disclosure in human-robot collaboration. Proceedings of the IEEE
International Symposium on Robot and Human Interactive
Communication, 1, 558–562.
Soanes, C., & Stevenson, A. (Eds.). (2005). Oxford dictionary of
English. (2nd ed.). New York: Oxford University Press.
Sproull, L., Subramani, R., Kiesler, S., Walker, J., & Waters, K.
(1996). When the interface is a face. Human-Computer Interac-
tion, 11, 97–124.
Struch, N., & Schwartz, S.H. (1989). Intergroup aggression: Its
predictors and distinctness from in-group bias. Journal of
Personality and Social Psychology, 56, 364–373.
Sunstein, C.R., & Nussbaum, M.C. (2004).
Animal rights:
Current debates and new directions. New York: Oxford
University Press.
Swartz, L. (2003). Why people hate the paperclip: Labels,
appearance, behavior and social responses to user
interface agents. Unpublished manuscript. Stanford University,
Palo Alto, CA.
Taylor, A. (2008, March 7). Inside Honda’s brain [Electronic version].
CNN Money. Retrieved September 8, 2009, from http://money.cnn.
com/2008/03/03/news/companies/taylor_honda.fortune/index.htm
Waxman, S.R., & Medin, D.L. (2007). Experience and cultural models
matter: Placing firm limits on anthropocentrism. Human Develop-
ment, 50, 23–30.
Waytz, A., & Epley, N. (2009). Social connection enables
dehumanization. Unpublished manuscript, University of Chicago,
Chicago, IL.
Waytz, A., Epley, N., & Cacioppo, J.T. (2010). Social cognition
unbound: Psychological insights into anthropomorphism and dehu-
manization. Current Directions in Psychological Science, 19,
58–62.
Waytz, A., Morewedge, C.K., Epley, N., Monteleone, G., Gao, J.-H.,
& Cacioppo, J.T. (2009). Making sense by making sentient: Unpre-
dictability increases anthropomorphism. Manuscript submitted for
publication.
232 Waytz, Cacioppo, and Epley
232
at Harvard Libraries on May 18, 2010pps.sagepub.comDownloaded from