Content uploaded by Nicholas Epley
Author content
All content in this area was uploaded by Nicholas Epley on Nov 06, 2018
Content may be subject to copyright.
Mistaking Minds and Machines: How Speech Affects Dehumanization
and Anthropomorphism
Juliana Schroeder
University of California, Berkeley
Nicholas Epley
University of Chicago
Treating a human mind like a machine is an essential component of dehumanization, whereas attributing
a humanlike mind to a machine is an essential component of anthropomorphism. Here we tested how a
cue closely connected to a person’s actual mental experience—a humanlike voice—affects the likelihood
of mistaking a person for a machine, or a machine for a person. We predicted that paralinguistic cues in
speech are particularly likely to convey the presence of a humanlike mind, such that removing voice from
communication (leaving only text) would increase the likelihood of mistaking the text’s creator for a
machine. Conversely, adding voice to a computer-generated script (resulting in speech) would increase
the likelihood of mistaking the text’s creator for a human. Four experiments confirmed these hypotheses,
demonstrating that people are more likely to infer a human (vs. computer) creator when they hear a voice
expressing thoughts than when they read the same thoughts in text. Adding human visual cues to text
(i.e., seeing a person perform a script in a subtitled video clip), did not increase the likelihood of inferring
a human creator compared with only reading text, suggesting that defining features of personhood may
be conveyed more clearly in speech (Experiments 1 and 2). Removing the naturalistic paralinguistic cues
that convey humanlike capacity for thinking and feeling, such as varied pace and intonation, eliminates
the humanizing effect of speech (Experiment 4). We discuss implications for dehumanizing others
through text-based media, and for anthropomorphizing machines through speech-based media.
Keywords: communication, voice, mind perception, dehumanization, anthropomorphism
Alan Turing (1950) created a famous benchmark for determin-
ing whether a computer can “think:” when it can convince a
majority of people that they are interacting with a person instead of
a machine. Turing’s link between thinking and personhood is no
accident (Farah & Heberlein, 2007). Boethus, writing in the 6th
century, defined personhood as “an individual substance of ratio-
nal nature” (Singer, 1994). Centuries later, John Locke (1841/
1997) echoed this definition of a person as “an intelligent being
that has reason and reflection.” Immanuel Kant (1785/1993) used
this definition of personhood as the guiding light of morality,
noting that, “rational beings are called persons inasmuch as their
nature already marks them out as ends in themselves.” Recent
surveys of laypeople likewise identify the capacity for thinking as
a unique feature of humanity (Leyens et al., 2000,2001). A person
has a mind capable of thinking but a computer does not.
As clear as this reality may be, it may not be so clear psycho-
logically. People sometimes recognize a thoughtful mind in their
cars, computers, or other mindless gadgets (Guthrie, 1995;Naas,
2010). A robot that moves at a humanlike pace seems more
thoughtful than a relatively sluggish or frantic robot (Morewedge,
Preston, & Wegner, 2007). An autonomous automobile that inter-
acts with you using a human voice while driving itself seems
“smarter,” and therefore, more trustworthy, than a noninteractive
vehicle (Waytz, Heafner, & Epley, 2014). Attributing humanlike
mental capacities of thinking and feeling to nonhuman agents is
the essence of anthropomorphism (Epley, Waytz, & Cacioppo,
2007).
Inversely, people sometimes fail to recognize a thoughtful mind
in other human beings, treating them instead like relatively mind-
less animals or objects (Haslam, 2006). Failing to attribute a
humanlike mind to another person is the essence of dehumaniza-
tion. These twin phenomena of anthropomorphism and dehuman-
ization raise a fundamental question in social life that goes far
beyond Turing’s test: how does an agent convey the fundamentally
humanlike capacity to think or feel? Answering this question will
predict when machines might be treated like people, and when
people might be treated like machines. It also predicts when
machines might be mistaken for people, and people mistaken for
machines.
Existing theory predicts that anthropomorphism and dehuman-
ization are determined by features of the agent being perceived
(e.g., morphology, motion, and observed behavior) and by features
This article was published Online First August 11, 2016.
Juliana Schroeder, Haas School of Business, University of California,
Berkeley; Nicholas Epley, Booth School of Business, University of Chicago.
We thank the Neubauer Family Faculty Fellowship for financial support
of this research, and Daniel Gilbert for suggesting the main dependent
variable used in these experiments. We also thank Jasmine Kwong, Michal
Dzitko, Annette Felton, Shreya Kalva, Alex Kristal, Paul Lou, Adam
Picker, Megan Porter, Sunni Rogers, Jenna Rozelle, Max Snyder, and
Sherry Tseng for assistance conducting these experiments. Portions of this
research were presented at the 2013 Annual Meeting of the Society for
Personality and Social Psychology.
Correspondence concerning this article should be addressed to Juli-
ana Schroeder, Haas School of Business, University of California,
Berkeley, 2220 Piedmont Avenue, Berkeley, CA 94720. E-mail:
jschroeder@haas.berkeley.edu
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Experimental Psychology: General © 2016 American Psychological Association
2016, Vol. 145, No. 11, 1427–1437 0096-3445/16/$12.00 http://dx.doi.org/10.1037/xge0000214
1427
of the perceiving agent (e.g., group affiliation, motivation, and
social connection; Epley, Waytz, & Cacioppo, 2007). Here we test
what we believe is a particularly important feature of the agent
being perceived: a humanlike voice. Beyond the semantic content
in speech, a humanlike voice also conveys paralinguistic informa-
tion (e.g., volume, tone, and rate) that provides additional insight
into one’s thoughts and feelings. Indeed, voice evolved in large
part as a tool to communicate an agent’s mind to others through
speech (Pinker & Bloom, 1990), and people can more accurately
estimate others’ mental states when they hear someone speak than
when they read the same words in text (Hall & Schmid Mast, 2007;
Kruger, Epley, Parker, & Ng, 2005). Therefore, we predicted that
communicating with a humanlike voice would make an agent seem
more like a person (vs. machine) than communicating the same
content through other communication media (e.g., reading text,
observing body language, or speaking with a voice that lacks
critical paralinguistic cues). Adding a humanlike voice to a ma-
chine might make it seem more like a person (i.e., anthropomor-
phism). Likewise, removing voice from an actual person by com-
municating through text might make a person seem more like a
machine (i.e., dehumanization).
Several existing experiments suggest that adding a humanlike
voice to computerized agents increases anthropomorphism (Nass
& Brave, 2005;Takayama & Nass, 2008;Waytz, Heafner, &
Epley, 2014). Our experiments go beyond these results by provid-
ing a more precise understanding of the interpersonal conse-
quences of hearing a humanlike voice compared with observing
other inferential cues (e.g., visual cues such as seeing a human), by
identifying which cues in voice convey personhood, and by testing
the inverse possibility that removing voice from human interac-
tions might lead to dehumanized perceptions of a speaker. This
latter hypothesis is especially important as technology makes
text-based interactions increasingly common in everyday life.
We test how a humanlike speaking voice affects dehumaniza-
tion and anthropomorphism in four experiments. For each exper-
iment, we report how we determined our sample size, all data
exclusions, all manipulations, and all measures. Data for all ex-
periments can be retrieved at http://faculty.haas.berkeley.edu/
jschroeder/data.html. Experiments 1 and 3 remove voice and mea-
sure dehumanization (mistaking a human for a computer) whereas
Experiment 2 adds voice and measures anthropomorphism (mis-
taking a computer for a human). All experiments test the effect of
hearing speech versus reading the same words in text. To increase
generalizability and remove confounds, we generated text using
different methods in each experiment: transcriptions of human
speech (Exp. 1), computer-generated text (Experiment 2), or writ-
ten essays (Experiments 3 and 4). We added a third channel of
human visual cues in Experiments 1–3 to determine whether
speech is uniquely humanizing, and to test an alternative explana-
tion that any humanlike cue will reduce dehumanization and
increase anthropomorphism. We predicted a speaker’s voice would
be uniquely humanizing because it contains paralinguistic cues
that reveal active mental experiences of thinking and feeling, and
that these cues uniquely reveal the presence of a humanlike mind
(Schroeder & Epley, 2015).
Our final experiment tests why speech is humanizing. We sug-
gest that paralinguistic cues in a person’s voice can convey the
presence of a lively, thoughtful, and active mind, in a way that is
analogous to how visual cues convey the presence of biological
life. A person can tell whether another agent is alive or dead
because of variance in motion. A living person’s body moves in
naturalistic ways. A dead person’s body is still, with no variance in
motion. Likewise, a person’s voice also contains naturalistic vari-
ance through paralinguistic cues that may analogously convey the
presence of a lively and active mind. A speaker’s pitch rises and
falls over time, yielding variance in tone (intonation). A speaker’s
pace quickens and slows, producing variance in speech rate. The
presence of a lively and active mind—a humanlike mind— could
be reflected through these cues. A rising pitch, for instance, may
reflect confidence in judgment whereas a falling pitch may reflect
more careful deliberation. Consistent with this possibility, speak-
ers communicate mental states such as the valence of emotional
experience or intentions even when speech lacks any meaningful
semantic content (McAleer, Todorov, & Belin, 2014;Scherer,
Banse, & Wallbott, 2001;Weisbuch, Pauker, & Ambady, 2009).
This predicts that a person with a speaking voice lacking natural-
istic variance in paralinguistic cues would be evaluated similar to
a person being evaluated over text or another communication
media devoid of paralinguistic cues (such as silent video). A voice
lacking variance in paralinguistic cues might make another per-
son’s mind seem relatively dead or dull, more like a mindless
machine than like a mindful human being.
Experiment 1: Dehumanization
We examined our hypotheses by using the inverse of Alan
Turing’s (1950) famous test: are observers more likely to mistak-
enly believe genuine human speech was created by a computer
when they hear the speech than when they read the very same
words? To rule out the artifact that any human cue might lead
observers to think the script was created by a human, we also
added visual cues: seeing a human. If a person’s voice uniquely
conveys humanlike mental capacities, as we predict, then we
should find an effect of voice, and no unique effect of visual cues,
on judgments of a speech’s creator. To test our prediction, we
created a novel paradigm in which individuals evaluate whether
the creator of the content that they hear or read was a computer or
a human. Note that participants are not evaluating whether a voice
is that of a computer or a human, but rather whether the creator of
the content was a human or a machine. This is essential because
computerized voices could obviously sound different than a real
human voice. Our hypotheses are not testing how people evaluate
voices in communication, but rather about the inferences people
make about the agents who are actually communicating.”
Method
Participants. We decided to collect at least 20 speakers to
obtain a more ecologically valid range of stimuli than is common
in most psychology experiments (Fiedler, 2011;Kenny, 1985;
Wells & Windschitl, 1999). This stimulus sampling is an important
feature of all of our experiments because it enables us to assess
how people evaluate a range of naturalistic speakers that might be
encountered in everyday life (Brunswik, 1947). It also eliminates
concerns that any result might be produced by some idiosyncratic
feature of a single person’s voice, or by some artifact introduced
by creating artificial speech content. Naturalistic stimulus sam-
pling enables stronger inferences about the strength of a phenom-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1428 SCHROEDER AND EPLEY
enon in the midst of the perceptually rich and chaotic environment
of everyday life.
We initially recruited 33 people (M
age
⫽20.0, SD
age
⫽2.3,
58% female) from a Chicago research laboratory to serve as
speakers in exchange for $2.00. Each speaker created two videos:
one talking about a positive emotional experience and the other
talking about a negative emotional experience (in counterbalanced
order). We had no a priori prediction about the effect of emotional
valence; we manipulated valence only as a robustness check for
the magnitude of our predicted effect of speech (vs. text). To
obtain our target of 20 speakers, two independent raters subse-
quently coded the videos based on the extent to which the speaker
followed our instructions. Specifically, they evaluated “the extent
to which participants talked about a very emotional experience and
described all of their emotions in detail” on a scale of 0 (not at all),
1(somewhat), or 2 (very much), r⫽.73. We included the 20
speakers who followed our instructions the best on this measure
across their two speeches, giving us a final sample of 40 total
speeches (20 positive experiences and 20 negative experiences).
The 20 speakers included in the study did not differ significantly
from the 13 speakers not included based on their gender,
2
(1,
33) ⫽0.12, p⬎.10, or their age, t(31) ⫽1.23, p⬎.10.
We decided to collect at least 640 observers so that at least four
would watch each type of stimulus for each of the 40 videos in the
four experimental conditions (160 conditions total). In total, 652
observers (M
age
⫽32.0, SD
age
⫽10.6, 42% female, 7 missing
gender) from Amazon.com’s Mechanical Turk participated. These
observers completed the experiment in exchange for $0.30. We
removed five observers who did not complete our primary depen-
dent variable from analysis.
Speaker procedure. Participants described both a positive
and negative emotional experience in their life (in counterbalanced
order) on videotape. We selected this speech topic because it is
generally humanizing, to the extent that most people believe hu-
mans have more emotional experience than do machines. For the
positive emotional experience, the experimenter asked participants
to:
Think back on an important positive emotional experience that you
had. Talk about the entire experience and all of the positive emotions
that you felt during the experience. Describe your emotions from
beginning to end. This should be an experience that led you to feel
very emotional, such as feeling deep happiness. Be sure while telling
your story to explain all of your feelings and emotions throughout the
entire experience.
Instructions for the negative emotional experience were modi-
fied to refer to experiences of “deep sadness” instead of “deep
happiness.” The experimenter asked the speakers to repeat the
instructions to ensure that they understood. Speakers then sat in a
chair facing a video camera. To make the speeches as natural as
possible, the experimenter stood behind the camera and told speak-
ers to look at the camera and pretend like they were talking directly
to the experimenter. The speaker talked until he or she was
finished telling the story. Speech lengths varied from 1 to 3 min.
The experimenter then stopped the video camera and read the next
set of instructions for the negative emotional experience video
(order counterbalanced). Again, speakers repeated the instructions,
sat facing the video camera, and gave their speech. Finally, the
experimenter debriefed speakers.
One research assistant transcribed the speeches after the verba-
tim transcript procedure used in United States courtroom deposi-
tions. A second assistant checked the transcriptions for accuracy.
Observer procedure. We randomly assigned observers to one
of four experimental conditions: listening to a speaker (audio
condition), watching and listening to a speaker (audiovisual con-
dition), reading the speech (text condition), or reading and watch-
ing a speaker (subtitled video condition, with no sound included).
To make the presentation of stimuli as consistent as possible across
conditions, we presented all stimuli to observers as videos. Ob-
servers in the audio and text conditions, therefore, saw a black
video screen and either heard the words or read them on the screen,
respectively. This paradigm allows us to keep constant the amount
of time each observer spent on each stimulus.
Observers in the text condition read the following information
before observing the stimuli: “As you may know, computer tech-
nology is now attempting to mimic real human speech. Some
computer programs are good enough that they can convince some
observers that they are real people, whereas others are not as good.
For the next few minutes, you will read the text of a script. Your
job is to figure out whether the content of this script was originally
created by a computer trying to create a script that sounded like a
real human being or whether it was created by an actual human.”
Observers in the audio version read the same instructions, except
the third sentence read, “For the next few minutes, you will listen
to an actor (or actress) reciting a script.” Observers in the audio-
visual and subtitled video condition likewise read the same in-
structions, except the third sentence read, “For the next few min-
utes, you will watch an actor (or actress) reciting a memorized
script.”
To make sure that observers understood their task, those in the
audio condition received further clarification: “To be absolutely
clear, your job is not to determine whether the voice is of a real
person or not. You will hear a real human actor reading a script.
Your job is to determine whether the script itself was originally
written by a computer or an actual human.” Observers in the
audiovisual and subtitled video condition, in contrast, received this
clarification: “To be absolutely clear, your job is not to determine
whether the actor or actress is a real person or not. You will watch
a real human actor reciting a memorized script. Your job is to
determine whether the script itself was originally written by a
computer or an actual human.”
After reporting whether the script was originally created by a
human or computer, participants also reported how confident they
were that their answer was correct on an 11-point response scale
(0 ⫽not at all confident,5⫽moderately confident,10⫽abso-
lutely confident), and then explained “why they made the choice
they did” in a free-response box. We did not analyze the free
responses. Participants reported their confidence in this same way
in each of the following experiments as well (Experiments 2– 4),
but confidence did not vary reliably by condition in any experi-
ment reported in the manuscript and we, therefore, do not discuss
it further.
Results. Whether the speech was about a negative or positive
emotional experience did not affect the human versus computer
judgment, F(1, 639) ⫽0.01, nor did it interact with communica-
tion medium on this judgment, F(3, 639) ⫽0.82. Therefore, we
collapsed across this factor in the following analyses.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1429
MINDS AND MACHINES
Observers’ judgments of the script’s creator varied by experi-
mental condition, F(3, 643) ⫽10.76, p⬍.01,
2
⫽.05. Because
we used a nested experimental design (multiple observers for each
speaker), we analyzed the effect of each condition (fixed factors)
in a hierarchical regression controlling for the effect of speaker
(random factor). As shown in Figure 1, removing voice was
dehumanizing: observers who read the speeches were less likely to
believe it was created by a human (text condition; M⫽53.6%,
SD ⫽50.0%) than observers who listened to them (M⫽80.8%,
SD ⫽39.5%), t(627) ⫽5.29, p⬍.01, d⫽0.42. Adding individ-
uating visual cues to the voice in the audiovisual condition did not
increase the percentage who guessed the script was created by a
human (audiovisual condition; M⫽71.6%, SD ⫽45.2%) com-
pared with audio alone, t(628) ⫽⫺1.79, p⫽.07, d⫽0.14. In
contrast, stripping away the person’s voice while retaining text in
the subtitled video condition reduced the percentage who guessed
the script was created by a human (subtitled video condition; M⫽
60.7%, SD ⫽49.0%) compared with the audio condition alone,
t(628) ⫽3.70, p⬍.01, d⫽0.29. The text and subtitled video
conditions did not differ significantly from each other, t(628) ⫽
1.43, p⫽.15, d⫽0.11. An observer was most likely to believe a
script was created by a human when they heard the speaker’s
voice.
Discussion
Participants who heard a speech were more likely to believe its
content was created by a human (vs. computer) than participants
who read the very same speech. The cues in voice seem uniquely
humanizing, as visual cues— being able to see speakers in addition
to hearing their voices— did not increase the percentage of eval-
uators who believed the script was created by a human. Being able
to see a speaker without hearing his or her voice, however, sig-
nificantly decreased the percentage who believed the script was
created by a human. The presence of a mindful human creator was
most clearly conveyed through a person’s voice rather than by the
overall amount of individuating (humanlike) cues available.
Experiment 2: Anthropomorphism
The results of Experiment 1 suggest that removing a voice from
human-generated speech can lead people to believe its content was
created by a mindless machine. Experiment 2 tests the inverse: can
adding a human voice to computer-generated speech lead people to
believe its content was created by a mindful human being? This
experiment is therefore closer to an actual Turing test, examining
when human observers might mistake a mindless machine for a
mindful person.
Method
Participants. We predetermined a sample size of at least 10
speakers and 240 observers. These sample sizes were smaller than
Experiment 1 because speakers all read the same computer-
generated text and so we expected less variability in evaluations
across speakers. Ten people (50% female) from a Chicago research
laboratory served as speakers in exchange for $3.00. Subsequently,
243 people (M
age
⫽30.9, SD
age
⫽10.0, 38% female) from
Amazon Mechanical Turk served as observers in exchange for
$0.30.
Speaker procedure. We created the essay using a “Postmod-
ernism Generator” (http://www.elsewhere.org/pomo/) that uses a
computer system to generate random text from recursive grammars
(Bulhak, 1996). The full text was:
“Society is elitist,” says Derrida. It could be said that Porter suggests that
we have to choose between material nihilism and neocultural theory.
“Truth is intrinsically dead,” says Debord; however, according to
Reicher, it is not so much truth that is intrinsically dead, but rather the
paradigm, and subsequent stasis, of truth. Sartre promotes the use of
dialectic subconceptualist theory to deconstruct the status quo. How-
ever, Baudrillard uses the term ‘cultural feminism’ to denote the
common ground between sexual identity and class. Marx suggests the
use of material nihilism to analyze sexual identity. But Bataille uses
the term ‘cultural feminism’ to denote a self-supporting totality. The
example of material nihilism intrinsic to Gibson’s Count Zero
emerges again in Virtual Light, although in a more mythopoetical
sense. Therefore, Baudrillard promotes the use of subdeconstructivist
theory to challenge hierarchy. Debord’s model of semantic Marxism
states that culture is part of the futility of consciousness. It could be
said that the main theme of the works of Gibson is the failure, and
hence the dialectic, of postcultural society. Cultural feminism holds
that context comes from the masses. But Bataille suggests the use of
Lyotardist narrative to read and analyze art.
Just before giving their speech, speakers read the following
instructions that were intended to maintain the natural paralinguis-
tic cues present in actual human speech:
When you perform, please imagine that you are the person who wrote
the essay. Imbue your words with all of the thoughts, emotions, and
20%
40%
60%
80%
100%
Audio Audiovisual Text Subtitled Video
Likelihood of Choosing "Human"
(for Human-Generated Speech)
20%
40%
60%
80%
100%
Audio Audiovisual Text Subtitled Video
Likelihood of Choosing "Human"
(for Computer-Generated Text)
Figure 1. Percentage of observers in Experiment 1 (Panel 1; n⫽652;
stimuli created from human-generated speech) and Experiment 2 (Panel 2;
n⫽243; stimuli created from computer-generated text) who believed a
script had been created by a human (vs. computer) in the audio, audiovi-
sual, text, and subtitled video conditions. Errors bars represent the SEM.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1430 SCHROEDER AND EPLEY
substance that the writer him/herself felt. Read it as if you were
actually coming up with the lines naturally off the top of your head
rather than reading from an essay. Speak as naturally as you would if
you were in the midst of a real conversation.
Observer procedure. We created the same four types of
stimuli described in Experiment 1: (a) audio, (b) audiovisual, (c)
text, and (d) subtitled video. Observers watched one of these four
stimuli, then guessed whether the script was “originally created by
a human or computer.”
Results
Because observers were not fully nested within speakers (i.e.,
observers who read the text always read the same essay), we did
not analyze the effects using hierarchical models. Observers’ judg-
ments of the script’s creator varied as predicted by experimental
condition, F(3, 239) ⫽3.62, p⫽.01,
2
⫽.04. As shown in
Figure 1, adding voice increased the tendency to anthropomor-
phize computer-generated text. Observers who listened to the
speeches (audio condition) were more likely to guess the script
was created by a human (M⫽60.0%, SD ⫽49.4%) than did those
who read identical speeches (text condition; M⫽32.3%, SD ⫽
47.1%), t(239) ⫽3.14, p⬍.01, d⫽0.41. Adding visual human
cues to the voice in the audiovisual condition did not significantly
increase the percentage who believed the text was created by a
human (M⫽40.0%, SD ⫽49.4%) compared to the audio condi-
tion. In fact, it unexpectedly decreased it, t(239) ⫽⫺2.25, p⫽.03,
d⫽0.29. Stripping away a person’s voice while retaining text in
the subtitled video condition also significantly reduced the per-
centage who believed the script was created by a human (M⫽
39.3%, SD ⫽49.3%) compared with the audio condition, t(239) ⫽
2.33, p⫽.02, d⫽0.30. Human versus computer judgments did
not vary between the audiovisual, subtitled video, and text condi-
tions, ts⬍1.
Discussion
Participants in two experiments were more likely to believe that
the content of a speech was created by a mindful human being
rather than by a mindless computer when they heard a speech than
when they read the identical content in text. This result did not
simply reflect an increase in accuracy for identifying authentic
human content through a person’s voice, but rather a systematic
bias to infer that a speech was created by a human after hearing it
spoken.
Adding individuating cues to a person’s voice through video
again did not increase the tendency to guess that a script was
created by a human. If anything, watching someone recite
computer-generated text reduced the tendency to guess the text
was created by a person compared with audio alone. We did not
predict this reduction, do not observe it elsewhere, and do not
speculate on it further. We note simply that these results again
suggest that a humanlike voice may be uniquely equipped for
conveying the presence of a humanlike mind.
Experiment 3: Job Candidates
Experiments 1 and 2 suggest that text is subtly dehumanizing,
making a creator seem less likely to be a human regardless of
whether the creator was actually a human or a computer. Creating
these stimuli required either removing a human voice from speech
(creating a transcript), or adding it to computerized text. One
condition was therefore always derived from another, creating a
potentially problematic confound in our design. In Experiment 3,
human participants generated both spoken and written stimuli,
removing this confound. Because semantic content may systemat-
ically differ in spoken and written communication, we also asked
observers to evaluate transcriptions of participants’ speeches. This
enables two tests of our voice hypothesis, one comparing spoken
versus written content that may vary in semantic content, and
another comparing spoken content versus a transcript that contains
identical semantic content.
Experiment 3 also tests our hypotheses in a context of practical
relevance: short “elevator pitches” to a potential employer explain-
ing a candidate’s qualifications for the job. In prior research, we
found that hypothetical job candidates were judged to be less
thoughtful, less intelligent, and less worthy of being hired when
potential employers read a transcript or written elevator pitch
compared with hearing the pitch (Schroeder & Epley, 2015).
Because the capacity for thinking is a defining feature of person-
hood, we expected that observers would also believe that the
content of an elevator pitch was less likely to have been created by
a real person when they read what a candidate has to say than when
they hear it.
Method
Participants (speakers and observers). Speakers were 18
University of Chicago Booth School of Business students (M
age
⫽
28.2, SD
age
⫽2.07, 39% female) who responded to our request for
research assistance. Speakers received a $5 Starbucks gift card as
compensation. These stimuli are also used in Schroeder and Epley
(2015).
We decided to collect at least 270 observers so that at least five
would watch each type of stimulus for each of the 18 videos in the
three experimental conditions (54 conditions total). In total, 273
people (M
age
⫽34.8, SD
age
⫽13.5, 50% female) from the Mu-
seum of Science and Industry in Chicago served as speakers in
exchange for a small snack.
Speaker procedure. We recruited MBA students to partici-
pate as job candidates in a study on how people make hiring
decisions. Candidates first named the company for which they
would most like to work, and then considered the pitch they would
make to encourage this company to evaluate them positively and to
hire them. Candidates made both a spoken and a written pitch to
prospective employers (order counterbalanced). In the spoken
pitch condition, we told candidates we would videotape them as
they gave their pitch and that they should speak directly to the
camera. We told candidates they had 2 min to talk but allowed
them to reach the natural conclusion of their pitch (actual videos
times ranged from 49 s to 2 min and 30 s). In the written pitch
condition, we told candidates to compose a letter to a prospective
employer. Candidates had 10 min to type their letter on a com-
puter, after which we told them to finish their thought and stop
typing. We arrived at these suggested time limits by asking two
MBA students to create spoken and written pitches without any
time restrictions, and then timed how long it took for them to speak
or write their pitches. After finishing both their spoken and written
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1431
MINDS AND MACHINES
pitches, candidates completed a short survey. These items are
unrelated to the current hypotheses.
To create transcripts of the spoken job pitches, one research assis-
tant transcribed the spoken pitches and a second checked them for
accuracy. To make the files more readable, we removed verbal filler
words unless their exclusion changed the sentence’s meaning.
Observer procedure. Observers were visitors from the Mu-
seum of Science and Industry in Chicago who agreed to serve as
hypothetical employers. Observers consented to take a short sur-
vey in which they would decide whether a job pitch was created by
a human or computer. We randomly assigned observers to one of
three experimental conditions: listening to a spoken pitch (audio
condition), reading a transcript of a spoken pitch (transcript con-
dition), or reading the written pitch (writing condition). We gave
observers the same instructions described in Experiment 1. After
observing the stimuli, participants then reported whether the job
pitch was “originally created by a human or computer.”
Results
Observers’ judgments of the job pitch’s creator varied by ex-
perimental condition, F(2, 270) ⫽6.18, p⬍.01,
2
⫽.04.
Because we used a nested experimental design (multiple observers
for each speaker), we analyzed the effect of each condition (fixed
factors) in a hierarchical regression controlling for the effect of
speaker (random factor). As shown in Figure 2, voice was human-
izing: observers who listened to the pitches (audio condition) were
more likely to believe it was created by a human (M⫽79.1%,
SD ⫽40.9%) than did those who read the identical text from these
speeches (transcript condition; M⫽55.9%, SD ⫽49.9%),
t(270) ⫽3.34, p⬍.01, d⫽0.41, or those who read the written
pitches (written condition: M⫽59.6%, SD ⫽49.3%), t(270) ⫽2.74,
p⫽.01, d⫽0.33). There was no difference in beliefs about the
pitch’s creator in the transcript versus written conditions, t⬍1.
Discussion
Hypothetical employers who listened to a job candidate’s ele-
vator pitch explaining his or her qualifications believed the content
was more likely to have been created by a machine when they read
what the candidate had to say— either in a written pitch or a
transcript—than when they heard what the candidate had to say.
These results converge with prior research (Schroeder & Epley,
2015) to suggest that a person’s voice conveys the fundamental
human capacity for thinking and reasoning. Lacking those cues
from a person’s voice in text, readers do not seem to add them in
spontaneously while reading someone’s text, in this case decreas-
ing the perceived likelihood that a written elevator pitch was
created by an actual person.
Although Experiments 1–3 suggest that voice moderates anthro-
pomorphism and dehumanization, they do not explain which fea-
tures of a person’s voice do so. Earlier we hypothesized that a
person’s voice may contain cues to mental life in much the same
way an agent’s body contains cues to biological life. In particular,
biological life is revealed through bodily motion. Lacking any
motion or movement, an agent appears to be either dead or
sleeping. Likewise, a person’s voice also contains motion through
variance in pitch (intonation) and pace, among others, providing
cues to a mental life of conscious thought or emotional experience
(Gray & Wegner, 2008;Morewedge et al., 2007;Schroeder &
Epley, 2015). If these cues convey the mental capacity for thinking
and feeling, and hence personhood, then voice lacking these cues
should lead to the same inferences as observers reading text alone.
We tested this directly in Experiment 4.
Experiment 4: Mindless Voice
We examined the mechanism by which a person’s voice con-
veys the presence of a humanlike mind by manipulating the
paralinguistic cues (e.g., volume, pitch, and rate of speech) that a
human voice adds beyond the semantic content of language, and
then testing whether the objective qualities of these cues mediated
the influence of a speaker’s voice on observers’ inferences. Ob-
servers therefore listened to either a “humanlike” voice that con-
tained the natural paralinguistic cues that a person would typically
use when expressing himself or herself, or they listened to a voice
that lacked these cues. Our theory is that paralinguistic cues are
closely connected to actual conscious thought processes as they
occur, thereby serving as an “honest” cue to the presence of mental
life. Naturalistic intonation (typical human variance in pitch) may
reflect emotional experience or rational thought. Changes in vol-
ume may likewise reflect emotional experience or certainty in
judgment. If so, then removing these cues from a person’s voice
should make him or her seem less humanlike, just as we observed
with text-based communication.
Method
Participants. In line with the preceding experiments, we in-
tended to collect between 10 and 20 communicators. Fifteen peo-
ple (M
age
⫽21.8, SD
age
⫽3.4, 53% female) from a Chicago
research laboratory served as writers in exchange for $4.00. We
then recruited four actors from a University drama department (2
male, 2 female, M
age
⫽20) to serve as speakers in exchange for
$25.00. We predetermined a sample of 300 observers (at least 12
observers in each of 25 experimental conditions) to ensure suffi-
cient statistical power given the additional variability we expected
because of using different writers and different speakers. Adults
visiting the Museum of Science and Industry in Chicago served as
observers (n⫽300, M
age
⫽32.2, SD
age
⫽12.1, 48% female) in
exchange for a small snack.
20%
40%
60%
80%
100%
"namuH" gniso
h
C
fo doohileki
L
Audio
Writing
Transcript
Figure 2. Percentage of observers in Experiment 3 (n⫽273) who
believed a script had been created by a human (vs. computer) in the audio,
writing, and transcript conditions. Errors bars represent the SEM.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1432 SCHROEDER AND EPLEY
Writer procedure. Writers wrote a description (using a com-
puter) of an important life decision that ended poorly and one that
ended well in counterbalanced order. To select essays that could
conceivably be created by either a human or a computer (thereby
avoiding ceiling or floor effects on our dependent variable), we asked
325 Amazon Mechanical Turk workers to indicate whether they
believed the essay was created by a human or a computer. The ratings
were normally distributed, ranging from 30% to 92% who believed
the script was created by a human. We selected five essays closest to
the median estimate of this distribution (60%) to use as stimuli.
Speaker procedure. Actors read the selected essays in a sound
booth. Male actors read the three essays originally written by men,
whereas female actors read the two essays originally written by
women. To create stimuli that contained a more or less humanlike
voice, each actor read a given essay twice, first in a “mindful” voice
using the same instructions provided to speakers in Experiment 2, and
then in a “mindless” voice that asked speakers to “read the words
exactly as you see them on the page. Put little feeling or life into the
words.”
Observer procedure. Observers listened to an actor reading the
statement in a mindful voice (n⫽120) or a mindless voice (n⫽120),
or read the original essay (n⫽60). We collected double the number
of participants in the voice conditions because two actors read each
essay, providing an equal number of participants evaluating each
actor’s reading of the essay as in the text condition. We told observers
in the two audio conditions that an actor was reading from the original
written essay. After reading or listening to the essay, observers judged
whether the essay was originally created by a human or a computer.
Results
As shown in Figure 3, more observers guessed that the essay
was created by a human when they listened to the mindful voices
(M⫽65.0%, SE ⫽4.6%) than when they read the text (M⫽
47.2%, SE ⫽6.5%),
2
(1,180) ⫽5.55, p⫽.02, ⫽0.18, or
listened to mindless voices (M⫽50.3%, SE ⫽4.6%),
2
(1,240) ⫽
5.52, p⫽.02, ⫽0.15. The text and mindless voice conditions
did not differ from each other,
2
(1,180) ⫽0.18, p⬎.10. These
results demonstrate that altering the speaker’s voice altered per-
ceptions of the content’s creator.
Differences in paralinguistic cues must account for these results
because the semantic content was identical across conditions.
Based on prior literature examining how paralinguistic cues affect
trait-based impression (e.g., Addington, 1968;Collins & Missing,
2003;Gregory & Webster, 1996;Hughes, Mogilski, & Harrison,
2014;Jones, Feinberg, DeBruine, Little, & Vukovic, 2010;
Laplante & Ambady, 2003;Niedzielski, 1999), we extracted the
following cues using Praat software (Boersma & Weenink, 2016):
mean pitch, mean amplitude, SD of pitch, SD of amplitude, speech
length, and mean pause length for each of the 20 voice clips.
1
Our
purpose was to test whether any of these cues mediated the effect
of mindful versus mindless voice on perceptions of the content’s
creator.
As intended, the mindful voices significantly differed on almost
all of these paralinguistic cues compared to their mindless voices.
Mindful voices had higher mean pitch (M⫽165.5 Hz, SD ⫽50.5
vs. M⫽147.7 Hz, SD ⫽48.4), t(9) ⫽7.76, p⬍.01, d⫽2.45,
higher SD of pitch (M⫽38.8 Hz, SD ⫽6.6 vs. M⫽23.7 Hz,
SD ⫽8.7), t(9) ⫽9.51, p⬍.01, d⫽3.01, higher mean amplitude
(M⫽71.4 dB, SD ⫽5.7 vs. M⫽62.2 dB, SD ⫽7.0), t(9) ⫽3.39,
p⫽.01, d⫽1.07, higher SD of amplitude (M⫽9.3 dB, SD ⫽1.4
vs. M⫽7.6 dB, SD ⫽1.2), t(9) ⫽8.94, p⬍.01, d⫽2.83, and
marginally longer overall length (M⫽60.9 s, SD ⫽11.1 vs. M⫽
55.7 s, SD ⫽10.1), t(9) ⫽1.97, p⫽.08, d⫽0.62. We observed
no difference in the mean pause length, t(9) ⬍1.37, of mindful
versus mindless voices.
To identify which of these paralinguistic cues predicted observ-
ers’ estimates of the speech’s creator, we conducted a linear
regression using each of these paralinguistic cues as independent
variables. Only intonation (the SD of pitch) significantly predicted
evaluations in this regression, ⫽.24, p⬍.01. When we included
intonation in a bootstrapped regression model predicting the effect
of experimental condition on judgments, the effect of voice con-
dition became nonsignificant (from ⫽.62, SE ⫽.26, p⫽.02, to
⫽.20, SE ⫽.33, p⫽.54), but the effect of intonation remained
significant (⫽.03, SE ⫽.01, p⫽.03), demonstrating that
intonation fully mediated the effect of the audio condition on
observers’ judgments. A 5,000-sample bootstrap test estimated a
standardized indirect effect of .44 (SE ⫽.21, 95% bias-corrected
CI [.02, .86]), indicating a significant indirect effect (MacKinnon,
Fairchild, & Fritz, 2007).
Discussion
Observers who listened to an actor speaking mindfully, as if
actually experiencing the thoughts and feelings described in an
essay were more likely to believe the essay was created by a
human (vs. computer) compared with observers reading the essay
1
To compute each speaker’s pitch profile in Praat, we first set a fixed
time step of 0.01 s. We set the pitch range to 150 to 500 Hz for female
speakers and 75 to 500 Hz for male speakers. We used the autocorrelation
analysis method in the pitch settings. To export the pitch, we selected the
entire pitch profile and saved the pitch listing as a text file that we imported
into Excel. We then computed the average and SD of the pitch in Excel. To
compute amplitude, we used the default Praat settings for the intensity
analysis. We selected the intensity profile and imported it into Excel using
the same method as we did for pitch. The number of seconds provided in
these pitch and amplitude profiles composed our measure of speech length.
To compute pause length, we counted the number of blank cells in the pitch
profile (each Excel cell represented 0.01 s).
0%
20%
40%
60%
80%
100%
"namuH
"
gnisoohC fo doohile
kiL
Mindful Voice
Mindless Voice
Text
Figure 3. Percentage of observers in Experiment 4 (n⫽300) who
correctly guessed that a script had been created by a human (vs. computer)
in the mindful voice, mindless voice, and text conditions. Errors bars
represent the SEM.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1433
MINDS AND MACHINES
directly or those listening to the same actors speaking mindlessly,
without any “feeling or life” in their voices. Subsequent analyses
of paralinguistic cues suggested that intonation was the most
important vocal cue for revealing the presence of a human mind.
Altering these paralinguistic cues by asking actors to speak “as if
only reading words on a page,” resulted in significantly more
monotone speeches, and yielded evaluations that did not differ
from evaluations of the text alone.
This indirectly suggests that text may lead to dehumanized
perceptions in our experiments because observers do not sponta-
neously add cues to mental life, such as intonation, when reading
someone else’s speech. The voice in one’s head while reading may
sound more monotone, and hence more robotic, than the voice of
an actual human speaker.
General Discussion
A person has a mind capable of thinking and feeling but a
computer does not. Four experiments suggest that this defining
feature of personhood is communicated at least partly through the
paralinguistic cues contained in speech, such that adding a human-
like voice to computer-generated text increased the tendency to
infer that it was actually created by a real person (Experiment 2).
In contrast, removing a voice from human-generated speech in-
creased the tendency to presume the content was actually created
by a computer (Experiment 1). In prior research (Schroeder &
Epley, 2015), job candidates were judged to be more intelligent,
thoughtful, and hirable when potential employers heard their ele-
vator pitch than when they read the same text or read a written
pitch. In Experiment 3, observers inferred that these elevator
pitches were more likely to have been created by a real human
being when observers heard what the candidate had to say than
when they read what the candidate had to say. These results
suggest that speech can be humanizing and its absence, dehuman-
izing.
These results appear to stem from the paralinguistic cues present
in speech, especially intonation (i.e., variability in pitch, Experi-
ment 4). Just as variability in bodily movement (i.e., biological
motion) serves as a cue for biological life, our experiments suggest
that variability in pitch (i.e., intonation) can serve as a cue to
mental life, and hence the presence of a humanlike mind. Text
alone does not contain these paralinguistic cues. If readers do not
spontaneously imbue a passage of text with such cues as they are
reading it, then they may be less likely to recognize the humanlike
mind behind the words they read. Consistent with this possibility,
observers in Experiment 4 who listened to actors read text in a
lifeless way—without the intonation present in naturalistic human
speech— evaluated them the same as observers who read the
writer’s text alone.
Existing research suggests that speech is uniquely equipped to
convey a person’s mental states and capacities. Observers judge a
target’s thoughts and feelings more accurately when they hear
someone speak than when they read the same text alone (Kruger et
al., 2005), or when they watch a person (silently) speaking (Gesn
& Ickes, 1999;Hall & Schmid Mast, 2007;Kruger et al., 2005;
Zaki et al., 2009). Most relevant for our current findings, adding an
authentic humanlike voice to a mindless machine can increase the
tendency to anthropomorphize it (Nass & Brave, 2005;Takayama
& Nass, 2008;Waytz, Heafner, & Epley, 2014).
Our research advances existing research in two important
ways. First, we provide a more nuanced comparison among
different forms of communication, including both verbal and
visual cues. In two experiments that directly compared the
effect of adding a speaker’s voice to semantic content (i.e.,
speech) compared to adding visual content (i.e., video), we
observed that adding voice significantly affected observers’
evaluations but that adding visual content did not (Experiments
1 and 2). We believe these results are interesting, but also
inconclusive. On the one hand, these results could indicate an
interesting fact about social cognition: that another person’s
humanlike mental capacities are heard more easily than they are
seen. Nobody would attempt to teach the basic ideas of statistics
through their facial expressions, or wax nostalgic about a family
vacation through pantomime. Vocal cues might simply provide
a much stronger signal for sophisticated thoughts and feelings
than visual cues. On the other hand, these results could also
indicate an uninteresting artifact of our experimental designs:
subtitled video could have distracted viewers from nonverbal
behavior, or the speeches we asked participants to give may not
have allowed for much physical nonverbal behavior. Determin-
ing the relative importance of visual versus verbal cues in mind
perception would require using stimuli that vary visual cues as
clearly as our current experiments vary vocal cues. Understand-
ing how different cues reveal different aspects of an otherwise
hidden mind is a promising avenue for future research.
Second, we provide a more precise understanding of why speech
might be humanizing: intonation (i.e., pitch variance) reveals a
humanlike mind. Whereas other research has examined how mean
level pitch affects trait-based evaluations of others (Addington,
1968;Collins & Missing, 2003;Feinberg et al., 2008;Gregory &
Webster, 1996;Hughes et al., 2014;Jones, Feinberg, DeBruine,
Little, & Vukovic, 2010;Laplante & Ambady, 2003;Niedzielski,
1999;Ray & Ray, 1990;Tigue, Borak, O’Connor, Schandl, &
Feinberg, 2012), our results suggest that variability in pitch may
convey the existence of humanlike mental capacities, leading
observers to infer a human source. People naturally modulate their
pitch when expressing thoughts; pitch rises when asking a ques-
tion, falls when expressing sadness, and generally fluctuates as
people reason and express ideas with enthusiasm and emotion.
The influence of intonation also helps to explain why text
could be relatively dehumanizing: readers may not spontane-
ously infuse written content with the cues that would reveal a
humanlike mind. This does not mean that a skilled author would
be unable to humanize their writing, but rather that randomly
selected readers do not spontaneously humanize text as they are
reading it. Identifying precisely why voice is humanizing is
important because it demonstrates how different types of voices
might be evaluated systematically differently, regardless of
whether the voices are coming from a real human being or from
a machine. For computer scientists and engineers interested in
humanizing technology even further, Experiment 4 suggests
that accurately mimicking naturalistic intonation should be an
especially important goal.
These results also have interesting implications for how people
interact with both machines and human beings in the modern
world. We consider each in turn.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1434 SCHROEDER AND EPLEY
Anthropomorphizing With Voice
Many technological devices now come equipped with human-
like speech. GPS systems direct drivers using computer-generated
speech. Call centers are staffed by computers that talk to angry
customers. And cell phone “assistants” ask, “What can I do for
you?” Our experiments suggest that humanlike speech may be an
especially important attribute of anthropomorphized machines.
Our results are consistent with other correlational data linking the
presence of a human voice to greater anthropomorphism of ma-
chines (Nass & Brave, 2005;Takayama & Nass, 2008). They are
also consistent with a recent experiment testing the causal influ-
ence of an agent’s voice on anthropomorphism. In this experiment,
participants were randomly assigned to drive either a normal
vehicle, an autonomous vehicle, or an “anthropomorphized” au-
tonomous vehicle that had a real human voice rated the anthropo-
morphized vehicle as the most mindful and trusted it most to drive
safely (Waytz et al., 2014). Just as in our experiments, a car with
a voice of its own seems better able to think and feel than any
normal car, and is one you might be willing to trust with your own
life.
Neither this experiment, nor those we report in this paper,
specifically identify the importance of voice in these important
outcomes of anthropomorphism. The Waytz et al. (2014) experi-
ment necessarily confounded the presence of thoughtful content
with the presence of a human voice (because the voice actually
included intelligent content). We did not measure the outcomes of
anthropomorphizing machines in the present manuscript. Identify-
ing the relative importance of voice compared to other attributes,
such as content, is therefore an important direction for future
research.
Our experiments suggest that intonation—pitch variance—may
be a critical cue for humanizing these gadgets through their voices.
Current technology appears limited in this regard compared with
real voices. To examine the intonation in computerized voices, we
created Siri voices (using either the “Alex” and “Kathy” Siri voice
templates on Macintosh computers, matching genders) from the
transcripts of the speakers talking about their emotional life expe-
riences in Experiment 1, and compared those voices to the actual
participants describing their experiences. The Siri voices had sig-
nificantly less intonation than real human voices (M
SIRI
⫽24.34
Hz, SD ⫽3.80, vs. M
Human
⫽36.06 Hz, SD ⫽19.64), t(39) ⫽
3.97, p⬍.01, d⫽0.83. As people’s reliance on technology
continues to deepen in modern life, engineers might do well to
understand the psychological processes that guide social cognition.
Seemingly subtle cues such as intonation could have an effect on
how people think about, and treat, their technology.
Dehumanizing Without Voice
Adding voice might anthropomorphize a phone, but our exper-
iments demonstrate that removing the voice from a real person
might be subtly dehumanizing, making the person seem more like
a mindless machine. Our results have potential importance for
everyday human interactions. We end by considering two contexts
in particular: courtrooms and text-based social media.
In courtrooms, witnesses provide testimony that is transcribed
verbatim by a stenographer. When deliberating jurors ask for more
information about a person’s testimony, they do not hear a record-
ing of the testimony but instead get the verbatim transcript. Our
experiments suggest that this seemingly innocuous procedure
could have profound effects on how jurors evaluate a given piece
of testimony. Was the victim truly remorseful? Was the crime a
well-reasoned plan or an unintentional accident? How much did
the victim really suffer? Many legal judgments rest on inferences
about the minds of perpetrators and victims, inferences that depend
on the medium through which information is conveyed.
More widely in everyday life, social interactions are becoming
increasingly text-based, with people keeping in touch through
technology such as texting or emailing rather than through talking.
In 2012, a representative sample of 2,000 American adults re-
ported spending 39% more time socializing online than face-to-
face (Marketwired, 2012). Some researchers have suggested that
these changing interaction norms are partly responsible for cross-
generational changes, such as increases in narcissism, loneliness,
and social apathy (Cacioppo & Patrick, 2008;Konrath, 2012;
Turkle, 2012;Twenge, 2013). Our experiments suggest that these
voiceless media might remove an important human element from
these interactions. In one experiment, adolescent women facing a
stressful event became more relaxed when they were able to call
their mother on the phone and hear her voice directly, but became
no more relaxed when they were able to “chat” using text with
their mothers (Seltzer, Prososki, Ziegler, & Pollak, 2012). Al-
though people are generally much happier connecting with others
than being alone (Kahneman & Deaton, 2010), connecting with
others online (using Facebook) in one study significantly reduced
happiness over time (Kross et al., 2013). In another experiment,
those randomly assigned to reconnect with an old friend over the
telephone reported feeling more connected than those who recon-
nected over email (Kumar & Epley, 2016).
Technology enables connections between people at great dis-
tances, over the Internet, through text messages, via email, and
over the phone. Emerging text-based media provide great techno-
logical gains in efficiency over speech-based media. Our research
suggests these technological gains in efficiency, however, may
come with some surprising psychological costs.
References
Addington, D. W. (1968). The relationship of selected vocal characteristics
to personality perception. Speech Monographs, 35, 492–503. http://dx
.doi.org/10.1080/03637756809375599
Boersma, P., & Weenink, D. (2016). Praat: Doing phonetics by computer
[Computer program]. Retrieved from http://www.praat.org/
Brunswik, E. (1947). Systematic and representative design of psychologi-
cal experiments. Berkeley, CA: University of California Press.
Bulhak, A. C. (1996, April). On the simulation of postmodernism and
mental debility using recursive transition networks. Technical Report
No. 96/264, Department of Computer Science, Monash University,
Melbourne, Australia.
Cacioppo, J. T., & Patrick, B. (2008). Loneliness: Human nature and the
need for social connection. New York, NY: Norton.
Collins, S. A., & Missing, C. (2003). Vocal and visual attractiveness are
related in women. Animal Behaviour, 65, 997–1004. http://dx.doi.org/
10.1006/anbe.2003.2123
Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On seeing human: A
three-factor theory of anthropomorphism. Psychological Review, 114,
864 – 886. http://dx.doi.org/10.1037/0033-295X.114.4.864
Farah, M. J., & Heberlein, A. S. (2007). Personhood and neuroscience:
Naturalizing or nihilating? The American Journal of Bioethics, 7, 37– 48.
http://dx.doi.org/10.1080/15265160601064199
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1435
MINDS AND MACHINES
Feinberg, D. R., DeBruine, L. M., Jones, B. C., & Little, A. C. (2008).
Correlated preferences for men’s facial and vocal masculinity. Evolution
and Human Behavior, 29, 233–241. http://dx.doi.org/10.1016/j
.evolhumbehav.2007.12.008
Fiedler, K. (2011). Voodoo correlations are everywhere: Not only in
neuroscience. Perspectives on Psychological Science, 6, 163–171. http://
dx.doi.org/10.1177/1745691611400237
Gesn, P. R., & Ickes, W. (1999). The development of meaning contexts for
empathic accuracy: Channel and sequence effects. Journal of Personal-
ity and Social Psychology, 77, 746 –761. http://dx.doi.org/10.1037/0022-
3514.77.4.746
Gray, K., & Wegner, D. M. (2008). The sting of intentional pain. Psycho-
logical Science, 19, 1260 –1262. http://dx.doi.org/10.1111/j.1467-9280
.2008.02208.x
Gregory, S. W., Jr., & Webster, S. (1996). A nonverbal signal in voices of
interview partners effectively predicts communication accommodation
and social status perceptions. Journal of Personality and Social Psy-
chology, 70, 1231–1240. http://dx.doi.org/10.1037/0022-3514.70.6.1231
Guthrie, S. E. (1995). Faces in the clouds: A new theory of religion. New
York, NY: Oxford University Press.
Hall, J. A., & Schmid Mast, M. (2007). Sources of accuracy in the
empathic accuracy paradigm. Emotion, 7, 438 – 446. http://dx.doi.org/10
.1037/1528-3542.7.2.438
Haslam, N. (2006). Dehumanization: An integrative review. Personality
and Social Psychology Review, 10, 252–264. http://dx.doi.org/10.1207/
s15327957pspr1003_4
Hughes, S. M., Mogilski, J. K., & Harrison, M. A. (2014). The perception
and parameters of intentional voice manipulation. Journal of Nonverbal
Behavior, 38, 107–127. http://dx.doi.org/10.1007/s10919-013-0163-z
Jones, B. C., Feinberg, D. R., DeBruine, L. M., Little, A. C., & Vukovic,
J. (2010). A domain-specific opposite-sex bias in human preferences for
manipulated voice pitch. Animal Behaviour, 79, 57– 62. http://dx.doi
.org/10.1016/j.anbehav.2009.10.003
Kahneman, D., & Deaton, A. (2010). High income improves evaluation of
life but not emotional well-being. Proceedings of the National Academy
of Sciences of the United States of America, 107, 16489 –16493. http://
dx.doi.org/10.1073/pnas.1011492107
Kant, I. (1993). Grounding for the Metaphysics of Morals, 3rd Ed. (J. W.
Ellington, Trans.) Indianapolis, IN: Hackett Publishing Company.
Kenny, D. A. (1985). Quantitative methods for social psychology. In G.
Lindzey & E. Aronson (Eds.), Handbook of social psychology (3rd ed.,
Vol. 1, pp. 487–508). New York, NY: Random House.
Konrath, S. (2012). The empathy paradox: Increasing disconnection in the
age of increasing connection. In R. Luppicini (Ed.), Handbook of re-
search on technoself: Identity in a technological society (pp. 204 –228).
Hershey, PA: IGI Global.
Kross, E., Verduyn, P., Demiralp, E., Park, J., Lee, D. S., Lin, N.,...
Ybarra, O. (2013). Facebook use predicts declines in subjective well-
being in young adults. PLoS ONE, 8, e69841. http://dx.doi.org/10.1371/
journal.pone.0069841
Kruger, J., Epley, N., Parker, J., & Ng, Z. W. (2005). Egocentrism over
e-mail: Can we communicate as well as we think? Journal of Personality
and Social Psychology, 89, 925–936. http://dx.doi.org/10.1037/0022-
3514.89.6.925
Kumar, A., & Epley, N. (2016). It’s surprisingly nice to hear from you:
Talking increases social connection compared to typing. Unpublished
data, University of Chicago.
Laplante, D., & Ambady, N. (2003). On how things are said: Voice tone,
voice intensity, verbal content, and perceptions of politeness. Journal of
Language and Social Psychology, 22, 434 – 441. http://dx.doi.org/10
.1177/0261927X03258084
Leyens, J. P., Paladino, P. M., Rodriguez-Torres, R., Vaes, J., Demoulin,
S., Rodriguez, A. P., & Gaunt, R. (2000). The emotional side of
prejudice: The attribution of secondary emotions to ingroups and out-
groups. Personality and Social Psychology Review, 4, 186 –197. http://
dx.doi.org/10.1207/S15327957PSPR0402_06
Leyens, J.-P., Rodriguez-Perez, A., Rodriguez-Torres, R., Gaunt, R., Pal-
adino, M.-P., Vaes, J., & Demoulin, S. (2001). Psychological essential-
ism and the differential attribution of uniquely human emotions to
ingroups and outgroups. European Journal of Social Psychology, 31,
395– 411. http://dx.doi.org/10.1002/ejsp.50
Locke, J. (1997). An essay concerning human understanding. Harmond-
sworth, England: Penguin Books. (Original work published 1841)
MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation
analysis. Annual Review of Psychology, 58, 593– 614. http://dx.doi.org/
10.1146/annurev.psych.58.110405.085542
Marketwired. (2012). Generation lonely? 39 percent of Americans spend
more time socializing online than face-to-face. Retrieved from http://
www.marketwired.com/press-release/-1648444.htm
McAleer, P., Todorov, A., & Belin, P. (2014). How do you say ‘hello’?
Personality impressions from brief novel voices. PLoS ONE, 9, e90779.
http://dx.doi.org/10.1371/journal.pone.0090779
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. http://dx.doi.org/10.1037/0022-3514.93.1.1
Nass, C. (2010). The Man Who Lied to His Laptop: What Machines Teach
Us About Human Relationships. New York, NY: Penguin.
Nass, C., & Brave, S. (2005). Wired for speech: How voice activates and
advances the human-computer relationship. Cambridge, MA: MIT
Press.
Niedzielski, N. (1999). The effect of social information on the perception
of sociolinguistic variables. Journal of Language and Social Psychol-
ogy, 18, 62– 85. http://dx.doi.org/10.1177/0261927X99018001005
Pinker, S., & Bloom, P. (1990). Natural language and natural selection.
Behavioral and Brain Sciences, 13, 707–727. http://dx.doi.org/10.1017/
S0140525X00081061
Ray, E. B., & Ray, G. B. (1990). The relationship of paralinguistic cues to
impression formation and the recall of medical messages. Health Com-
munication, 2, 47–57. http://dx.doi.org/10.1207/s15327027hc0201_4
Scherer, K. R., Banse, R., & Wallbott, H. G. (2001). Emotion inferences
from vocal expression correlate across languages and cultures. Journal
of Cross-Cultural Psychology, 32, 76 –92. http://dx.doi.org/10.1177/
0022022101032001009
Schroeder, J., & Epley, N. (2015). The sound of intellect: Speech reveals
a thoughtful mind, increasing a job candidate’s appeal. Psychological
Science, 26, 877– 891. http://dx.doi.org/10.1177/0956797615572906
Seltzer, L. J., Prososki, A. R., Ziegler, T. E., & Pollak, S. D. (2012). Instant
messages vs. speech: Hormones and why we still need to hear each
other. Evolution and Human Behavior, 33, 42– 45. http://dx.doi.org/10
.1016/j.evolhumbehav.2011.05.004
Singer, P. (1994). Ethics. Oxford, United Kingdom: Oxford University
Press.
Takayama, L., & Nass, C. (2008). Driver safety and information from afar:
An experimental driving simulator study of wireless vs. in-car informa-
tion services. International Journal of Human-Computer Studies, 66,
173–184. http://dx.doi.org/10.1016/j.ijhcs.2006.06.005
Tigue, C. C., Borak, D. J., O’Connor, J. J. M., Schandl, C., & Feinberg, D. R.
(2012). Voice pitch influences voting behavior. Evolution and Human
Behavior, 33, 210 –216. http://dx.doi.org/10.1016/j.evolhumbehav.2011.09
.004
Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59,
433– 460. http://dx.doi.org/10.1093/mind/LIX.236.433
Turkle, S. (2012). Alone together: Why we expect more from technology
and less from each other. New York, NY: Basic Books.
Twenge, J. M. (2013). Does online social media lead to social connection
or social disconnection? Journal of College & Character, 14, 11–20.
http://dx.doi.org/10.1515/jcc-2013-0003
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1436 SCHROEDER AND EPLEY
Waytz, A., Heafner, J., & Epley, N. (2014). The mind in the machine:
Anthropomorphism increases trust in an autonomous vehicle. Journal of
Experimental Social Psychology, 52, 113–117. http://dx.doi.org/10
.1016/j.jesp.2014.01.005
Weisbuch, M., Pauker, K., & Ambady, N. (2009). The subtle transmission
of race bias via televised nonverbal behavior. Science, 326, 1711–1714.
http://dx.doi.org/10.1126/science.1178358
Wells, G. L., & Windschitl, P. D. (1999). Stimulus sampling and social
psychological experimentation. Personality and Social Psychology Bul-
letin, 25, 1115–1125. http://dx.doi.org/10.1177/01461672992512005
Zaki, J., Bolger, N., & Ochsner, K. (2009). Unpacking the informational
bases of empathic accuracy. Emotion, 9, 478 – 487. http://dx.doi.org/10
.1037/a0016551
Received May 5, 2016
Revision received June 7, 2016
Accepted July 6, 2016 䡲
Call for Nominations
The Publications and Communications (P&C) Board of the American Psychological Association
has opened nominations for the editorships of Clinician’s Research Digest: Adult Populations and
Child and Adolescent Populations;Journal of Experimental Psychology: Learning, Memory, and
Cognition; Professional Psychology: Research and Practice; Psychology and Aging; and Psychol-
ogy, Public Policy, and Law for the years 2019 to 2024. Thomas Joiner, PhD; Robert L. Greene,
PhD; Ronald T. Brown, PhD; Ulrich Mayr, PhD; and Michael E. Lamb, PhD, respectively, are the
incumbent editors.
Candidates should be members of APA and should be available to start receiving manuscripts in
early 2018 to prepare for issues published in 2019. Please note that the P&C Board encourages
participation by members of underrepresented groups in the publication process and would partic-
ularly welcome such nominees. Self-nominations are also encouraged.
Search chairs have been appointed as follows:
●Clinician’s Research Digest: Adult Populations and Child and Adolescent Populations,
Chair: Pamela Reid, PhD
●Journal of Experimental Psychology: Learning, Memory, and Cognition, Chair: Stephen
Rao, PhD
●Professional Psychology: Research and Practice, Chair: Kate Hays, PhD
●Psychology and Aging, Chair: Pamela Reid, PhD
●Psychology, Public Policy, and Law, Chair: David Dunning, PhD
Candidates should be nominated by accessing APA’s EditorQuest site on the Web. Using your browser,
go to http://editorquest.apa.org. On the Home menu on the left, find “Guests/Supporters.” Next, click on
the link “Submit a Nomination,” enter your nominee’s information, and click “Submit.”
Prepared statements of one page or less in support of a nominee can also be submitted by e-mail to Sarah
Wiederkehr, P&C Board Editor Search Liaison, at swiederkehr@apa.org.
Deadline for accepting nominations is Monday, January 9, 2017, after which phase one vetting will
begin.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
1437
MINDS AND MACHINES
