Content uploaded by José-Miguel Fernandez-Dols
Author content
All content in this area was uploaded by José-Miguel Fernandez-Dols
Content may be subject to copyright.
Russell, Bachorowski, & Fernandez Dols 1
Running Title: Emotional Expressions
Annual Review of Psychology, 2003
Facial and Vocal Expressions of Emotion
James A. Russell
Department of Psychology
Boston College
james.russell@bc.edu
Jo-Anne Bachorowski
Department of Psychology
Vanderbilt University
j.a.bachorowski@vanderbilt.edu
José-Miguel Fernández-Dols
Departmento de Psicología Social y Metodología
Universidad Autónoma de Madrid
jose.dols@uam.es
Key words: affect, display rule, perception, nonverbal, communication
Corresponding author:
James A. Russell
Department of Psychology
McGuinn Hall
140 Commonwealth Avenue
Boston College
Chestnut Hill, MA 02467
email: james.russell@bc.edu
telephone: 617 552-4546
fax: 617 552-0523
Russell, Bachorowski, & Fernandez Dols 2
Abstract
A flurry of theoretical and empirical work concerning the production of and response to facial
and vocal expressions has occurred in the past decade. That emotional expressions express
emotions is a tautology, but may not be a fact. Debates have centered on universality, the nature
of emotion, and the link between emotions and expressions. Modern evolutionary theory is
informing more models, emphasizing that expressions are directed at a receiver, that the interests
of sender and receiver can conflict, that there are many determinants of sending an expression in
addition to emotion, that expressions influence the receiver in a variety of ways, and that the
receiver’s response is more than simply decoding a message.
Russell, Bachorowski, & Fernandez Dols 3
Outline
Introduction: Definition and scope
History
Key Theoretical Advances
The Receiver’s Response
Recognition of Discrete Emotions
Alternative Views of the Receiver
Still Other Effects
The Sender
Emotions as causes of facial expressions
Measurement
Positive Emotions
Negative Emotions
Surprise
Emotions as causes of vocal expressions
Measurement
Nonlinguistic vocalizations
Speech
Future Directions
Conclusions
Russell, Bachorowski, & Fernandez Dols 4
Introduction: Definition and Scope
Smiles, chuckles, guffaws, smirks, frowns, and sobs — these and their milder cousins
occurring in the fleeting changes in the countenance of a face and in the tone of a voice are
essential aspects of human social interaction. Indeed, expressionless faces and voices are
considered to be indicators of mental illness, expressive faces and voices to be windows to the
soul. The last chapter in the Annual Review of Psychology devoted entirely to this topic (Ekman
& Oster 1979) summarized a rich research tradition that was predominant in the study of
emotion at that time. Since their chapter, much has changed.
Both scientists and nonscientists traditionally considered smiles, chuckles, and the rest to
be “expressions of emotion” (EEs). Ekman and Oster (1979) continued this tradition, but newer
work questions the assumptions in both key words, expression and emotion. Signals might be a
better term for some cases of EE, although signal, symptom, symbol, manifestation, display,
sign, expression and other terms are often used interchangeably, without clear definitions or
distinctions. The relation of EEs to emotion (and the nature of emotion) remains unclear. Further,
the class of EEs is probably heterogeneous and so any one name will prove misleading. For
instance, some EEs are, to use Goffman’s (1959) terms, given (produced for the purpose of
communication) and others are given off (side-effects of movements produced for other
purposes). The boundaries encircling the class of EEs are not self-evident, leaving us pointing to
examples and leaving the category EE conceptually undefined. Indeed, we doubt that it is a
scientifically viable unitary category.
History
Traditionally, senders have been thought to “express” or “encode” – that is, emit veridical
information about – their internal state, much as a lighthouse broadcasts its visual and auditory
Russell, Bachorowski, & Fernandez Dols 5
warning to any and all who happen to perceive it. In turn, receivers “recognize” or “decode” the
message and benefit thereby. This image of honest and altruistic broadcasting has deep historical
roots. Thought of as a God-given and universal language, EEs revealed passions (such as love
and hate), virtues (courage), and vices (sloth). These ideas were evident in philosophical,
religious, and artistic theories from ancient times to the 19th century, and continued to appear in
later work by anatomists, physiologists, and other scientists (Montagu 1994). Among those
scientists was Charles Darwin (1872). Although he relied on traditional assumptions about
expression and emotion, Darwin substituted natural selection for God and made important
observations about cross-species and cross-cultural similarities in EEs to bolster his argument for
that substitution.
The modern era of the study of EEs began in 1962 with a theory proposed by Sylvan
Tomkins. Like Darwin, Tomkins and those he inspired (Izard 1971, Ekman et al 1972)
perpetuated many of the traditional assumptions about expression. To these, Tomkins added
another ancient idea, that of a small, fixed number of discrete (“basic”) emotions. On Tomkins’
theory, each basic emotion can vary in intensity and consists of a single brain process (an “affect
program”), whose triggering produces all the various manifestations (components) of the
emotion, including its facial and vocal expression, changes in peripheral physiology, subjective
experience, and instrumental action. Because they have a single cause, these components tightly
cohere in time and are intercorrelated in intensity. Emotions are sharply distinguished from
cognitions. The set of theories, methods, and assumptions inspired by Tomkins guided the study
of emotion for over a quarter century.
Another assumption found in Darwin and continued by Tomkins – that the same message
is encoded and decoded – guided much of the research on EEs: If (except in cases of deliberate,
Russell, Bachorowski, & Fernandez Dols 6
socially induced deception) EEs broadcast veridical information which the receiver recognizes,
then researchers can focus on either the encoding (sending) or decoding (receiving) side. Either
could establish which of the small number of basic emotions was expressed by a specific facial
or vocal pattern. For practical reasons, most research therefore relied on decoding (judgments by
observers) to establish just what emotion a specific EE represents. The actual emotional state of
the sender was typically neglected.
Some Key Theoretical Advances
Modern evolutionary theory renders obsolete Darwin’s specific analysis of EEs, which
relied on a Lamarkian inheritance of acquired characteristics, on group selection pressures, and
on a characterization of EEs as vestiges. Modern theory instead emphasizes natural selection, the
interests of the individual, adaptation, and function (Dawkins & Krebs 1978, Fridlund 1994,
Owren & Rendall 2001). A pivotal recognition in modern theories is that EEs, even when
“given” in Goffman’s sense, are not broadcast to any and all but are directed at a receiver and
evolved to influence that receiver in ways beneficial to the sender. As the interests of sender and
receiver only sometimes coincide, it is not always in the sender’s interest to provide veridical
information. EEs are thus are as capable of being deceptive as honest.
A second key recognition was that the receiving side is more than a reflex-like decoding
of a message. If EEs evolved to alter the receiver, then a variety of effects can occur. For
instance, vocal stimuli can capture the receiver’s attention and alter his or her affective state
without any emotion being encoded or decoded (Owren et al 2002). Furthermore, receiving
mechanisms were subject to their own course of evolution. The receiver’s interest lies not only in
detecting cues but also in distinguishing veridical cues from deceptive ones. Receivers also
benefit by using cues given off to anticipate the sender’s subsequent actions. This last point is
Russell, Bachorowski, & Fernandez Dols 7
underscored by inadvertent communication, such as when a predator uses the prey’s EEs to
locate the prey (Seyforth & Cheney 2003).
The theory of basic emotions has also been cogently criticized (Turner & Ortony 1992),
and new conceptions of emotion have emerged (Russell 2002). These conceptions include an
emphasis on multi-component dynamic processes laced with cognition (Scherer 2001, Smith &
Kirby 2001), with a looser, more malleable and context-dependent relation among the
components (Bradley & Lang 2000b), and with a role for broad primitive affective dimensions
such as pleasure-displeasure and activation (Davidson 2000, Russell & Feldman Barrett 1999).
These theoretical advances led us to separate the study of EEs into two topics: (a) the
receiver’s response to an EE (including but not limited to an attribution of emotion to the sender)
and (b) the sender’s production of an EE (there may be a variety of factors influencing the
production of a given EE, some of which have little to do with emotion). Evidence on one of
these topics cannot be taken as evidence on the other. Evidence on the universality of one cannot
be taken as evidence on the universality of the other. Rather than judging emotion attributions as
correct or incorrect, we suggest a more descriptive approach on how these two processes work,
on what natural selection has bequeathed to the newborn regarding these processes, and on how
they develop over the lifetime. Our discussion centers on a psychological analysis of (non-
clinical) human adults.
The Response of the Receiver
Recognition of Discrete Emotions
Much research was and still is inspired by the theory that certain EEs signal specific
emotions, which receivers “decode.” Receivers include human infants (Nelson & de Haan 1997)
and nonhuman species (Marler & Evans 1997), although Seyforth and Cheney (2003) observed
Russell, Bachorowski, & Fernandez Dols 8
that, with the possible exception of the chimpanzee, no other species is currently thought to
decode EEs in the way that humans are theorized to attribute emotions to the sender. The process
of “decoding” has not been specified, but has been characterized as innate (Izard, 1994), easy
(Ekman, 1975), categorical (Calder et al 1996), and immediate: “The initial translation of an
expression into some meaning is likely to be so immediate that we are not aware of the process
we go through” (Ekman 1997, p.334).
In the typical study, a facial or vocal EE is presented to a receiver who then indicates
which emotion it signals. The impressive empirical foundation for this theory is the repeated
finding that, despite differences in culture, age, or background, receivers agree on the emotion
signaled more often than could be achieved by chance (for facial EEs, see reviews by Elfenbein
& Ambady 2002 and Russell 1994; for vocal, Johnstone & Scherer 2000). Agreement is typically
higher for facial than vocal EEs (Hess et al 1988, Wallbott & Scherer 1986).
Nevertheless, key problems remain unresolved. One problem concerns the facial or vocal
signals chosen for study. The “correct” signal for each specific emotion in these studies was not
specified on theoretical grounds, although Darwin’s speculations along these lines are sometimes
alluded to. Nor was the signal empirically specified by recording the EEs emitted by senders in
known emotional states (more on this shortly). Instead, instances were typically obtained from
actors asked to convey emotions through their face or voice. Through an iterative process, those
portrayals that achieved highest agreement on the emotion conveyed were selected as the correct
signals. One technical question is how to interpret the significance of agreement obtained in this
way. A deeper theoretical question also arises because this iterative process has not yielded what
might have been expected, namely a specific physically characterizable signal for each emotion.
Russell, Bachorowski, & Fernandez Dols 9
Instead, for each emotion, there is a range of signals that achieve varying degrees of
agreement. For example, Ekman and Friesen (1978, Table 11.1) specified 65 different facial
patterns that they consider to be signals for anger. Comparable difficulties arose in attempts to
specify a vocal signature for each basic emotion (Banse & Scherer 1996). No theoretical
rationale for this variety has been offered. Further, the sets of “correct signals” resulting from
this iterative process have dubious ecological validity – given that we know of no evidence that
acted stimuli used in this research correspond to what persons in the specified emotional states
produce spontaneously (and some indirect evidence to the contrary, Russell & Carroll 1997).
Indeed, when spontaneous rather than acted EEs are presented to receivers, the amount of
agreement on a specific emotion drops or disappears (Motley & Camden 1988 for facial;
Johnson et al 1986, Exp. 1, Pakosz 1983 for vocal).
The typical decoding study is also compromised by the task given the receiver. Forcing
the receiver to choose one from a short list of emotions can inflate agreement and even produce
blatant artifacts (Russell 1994). Providing the receiver with more options lowers agreement
(Banse & Scherer 1996). Allowing the receiver to specify any emotion (free labeling) lowers
agreement still further (Russell 1994). Some of the artifacts can be eliminated by providing
“none of the above” as a response option (Frank & Stennett 1999), and future studies should do
so.
A lively discussion centered on the question of universality (Ekman 1994, Izard 1994,
Russell 1994, 1995). In an empirical response to that debate, Haidt and Keltner (1999) obtained
evidence in the USA and India that was consistent with both proponents and critics. One
interesting finding was a “gradient of recognition:” Some emotions are more “recognizable” than
others, and the gradient is steep enough that the recognizable fades into the unrecognizable.
Russell, Bachorowski, & Fernandez Dols 10
(Because the term “recognition” presupposes that the emotion is present in the stimulus to be
recognized, a neutral term such as “attribution” would be preferable.) Attribution depends on the
similarity between the sender's and the receiver's language and culture; see Russell (1991, 1994)
and Elfenbein & Ambady (2002) for facial EEs, Scherer et al (2001) for vocal ones. Attribution
of the specific emotion predicted by Tomkins’ theory also declines as one moves further from a
Western cultural background. With participants isolated from Western ways, agreement that
smiles indicate something positive is high, but agreement on what emotion to attribute to other
facial expressions is low and may or may not exceed chance when method artifacts are
eliminated (Russell 1994).
Russell and Fernandez Dols (1997) summarized the available evidence as consistent with
“minimal universality:” (a) Facial and vocal changes occur everywhere and are coordinated with
the sender’s psychological state; (b) Most people everywhere can infer something of the sender’s
psychological state from those facial and vocal changes. The challenge for those who would
maintain any stronger version of universality (such as the existence of universal signals for
specific emotions) is to find evidence that goes beyond what can be accounted for by minimal
universality. The implication for everyone is that it is time to pursue other conceptualizations of
the response of the receiver. (Kappas et al 1991, Owren et al 2002, and Wierzbicka 1999 arrived
at similar conclusions, although for different reasons.)
Alternative Views of the Receiver
Receivers do, sometimes, interpret an EE in terms of a specific emotion, but the nature of
the interpretive process remains to be determined. There is evidence that attributing a specific
emotion to the sender is more complex than the simple, easy, immediate detection of a signal.
For example, even when the stimuli are the hypothesized prototypical facial expressions of
Russell, Bachorowski, & Fernandez Dols 11
emotion, the emotion attributed to the face depends on the context within which the expression
occurs (Carroll & Russell 1996), on the gender of the sender (Widen & Russell 2002), and on the
receiver’s current affective state (Niedenthal et al 2000). Long ago, Hebb (1946) reported that
observers learned how to predict the emotions of chimpanzees, not by decoding emotion signals
from their faces or voices, but by learning how the individual chimp’s current expressive and
other behavior fit into a temporal pattern idiosyncratic to that chimp. Longitudinal studies of
clinical samples yielded similar conclusions (Ellgring 1986).
A receiver’s typical response might include much more than attribution of an emotion or
even be different from that. One well-supported possibility is that the receiver perceives the
internal state of the sender in terms of broad bipolar dimensions such as valence (pleasure –
displeasure) and activation (sleepy – hyperactivated). Receivers agree with one another in
judging EEs along these dimensions. For faces, both dimensions are readily apparent (Russell
1997), even when the receivers are 2-year olds (Russell & Bullock 1986). For voice, activation
dominates (Pittam et al 1990); valence is weak (Bachorowski 1999, Pereira 2000). Analyses of
confusions among emotions inferred from EEs supports this hypothesis (Russell & Bullock 1986
for faces, Pakosz 1983 for voice). Even in Schröder’s (2000) promising study of vocal outburts
(“yuck!”), which yield high agreement as to specific emotion, an analysis of confusions among
them suggests the presence of valence and activation dimensions.
In addition, faces and voices provide non-emotional information: The receiver notes
whether the sender is staring or looking away, laughing with or at someone, shouting because
background noise requires it or not. From a facial expression, receivers agree on the sender’s
situation (“she looks as if she is looking at a small child playing” Frijda 1969, p.169) and likely
future action (Frijda & Tchersasoff 1997). Receivers agree that the sender may be conveying a
Russell, Bachorowski, & Fernandez Dols 12
social message such as “back off”or “hello” to someone (Yik & Russell 1999). And we
anticipate that receivers will agree on aspects of the sender’s cognitive state, including attention,
uncertainty, puzzlement, determination, anticipated effort, registration of novelty, and sense of
control (Smith & Scott 1997).
In short, the receiver probably obtains from an EE information on the sender’s valence,
activation, quasi-physical actions (such as staring or talking), current situation, future actions,
social attitude, and cognitive state. If so, then the receiver might use this information to infer the
sender’s emotion (Russell 1997). Obviously, the reverse is also possible: Decoding a specific
emotion from the EE, the receiver could then infer the other information. Clearly, research is
needed on what information the receiver extracts first, easily, automatically, at a younger age, or
spontaneously from an EE and what information requires effort, training, or measures that guide
or channel the receiver’s response into something close to the researcher’s a priori hypothesis.
The cross-cultural study of such questions is especially needed.
Still Other Effects
EEs produce a variety of effects other than getting the receiver to think “lo, anger” or
some other emotion. Laughs elicit laughter; yawns elicit yawns (Provine 1997); and more
generally, receivers “mimic” the EEs of senders (Hatfield et al 1992). A receiver’s facial
musculature mirrors a face presented nonconsciously (Dimberg et al 2000). EEs alter the
receiver’s physiological state (Dimberg & Öhman 1996, Levenson 1996). Vocal EEs alter the
receiver’s self-reported affect (Bachorowski & Owren 2001) – which is not surprising given that
EEs are visual and auditory stimuli, which are known to influence affect along the dimensions of
valence and activation (Bradley & Lang 2000). Indeed, everyday experience shows that sounds
alter the hearer’s affect, as illustrated by sirens, thunder, and an infant’s cry. Receiver’s self-
Russell, Bachorowski, & Fernandez Dols 13
reported affect reflects the affective tone of a passage heard (Neumann & Strack 2000) or the
affective demeanor of a face seen (Wild et al 2001), even when the face is presented
nonconsciously (Dimberg et al 2000).
Such evidence is consistent with the theory that EEs function to alter the receiver’s state,
especially affect. Owren and colleagues (1997, 2001, 2002) questioned the exclusive focus
traditionally placed on the receiver’s cognitive representation of the message of the EE,
suggesting that in addition EEs alter the receiver’s state and thereby serve the larger goal of
social influence. And indeed, EEs do alter the course of social interaction. Sender’s
embarrassment (appeasement) elicits self-reported positive feelings in the receiver (Keltner &
Buswell 1997). EEs influence the degree of cooperation, dominance/submission, or antagonism
in subsequent interaction (Zivin 1977, Tiedens 2001). Norm violators who smile are treated more
leniently than those who do not (LaFrance & Hecht 1995). People whose facial expression is
imitated feel that they were better understood and that their interaction was smoother (Chartrand
& Bargh 1999). Complementary evidence for the same theme comes from the finding that
various EEs are differentially susceptible to serving as a conditional stimulus in a Pavlovian
conditioning paradigm (Dimberg & Öhman 1996). In this case, EEs would function to alter long-
term interaction.
Owren et al’s (2002) perspective is nicely illustrated by thinking of EEs as being like
infant-directed (ID) speech. ID-speech (baby talk) has known acoustic characteristics (Fernald
1991, Katz et al 1996) and accompanying facial behavior (Chong et al 2002). ID speech is
preferred by infants (Cooper & Aslin 1990, Fernald & Kuhl 1987), elicits their attention, alters
their emotional behavior, helps direct their attention to a specific stimulus, and facilitates their
learning of associations (Kaplan et al 1997). Infants deprived of ID speech (e.g., when their
Russell, Bachorowski, & Fernandez Dols 14
caregiver is depressed) show significant learning and developmental deficits (Kaplan et al 2002,
Murray et al 1996). Perhaps EEs function in a comparable fashion. According to this account,
EEs operate on at least two levels. The simple acoustics of the sound elicit attention and alter
core affect (valence and activation) directly, but the affect and meaning attributed to the sound
also depend on context and prior experience. Thus, hearing laughter is a generally pleasant
experience, with voiced more pleasant than unvoiced laughter (Bachorowski & Owren 2001):
sound acoustics have a direct effect. On the other hand, the affect-altering effect of EEs also
depends on context (J-A Bachorowski et al unpublished manuscript, Vanderbilt Univ., Hess &
Kirouac 2000, Kappas et al 1991). For instance, hearing a high pitched shriek might be pleasant
during a party but unpleasant when alone in a dark street.
The Sender
Surprisingly few studies have tested the basic claim of EEs: Emotions cause them.
Perhaps the claim was simply taken as obviously true, perhaps studies on the receiver’s decoding
of EEs was mistakenly believed to be an adequate test, perhaps practical and ethical concerns
hindered research, or perhaps needed measurement techniques were slow in coming and difficult
to use. The studies that have been done almost always focused on either facial or vocal changes
rather than their combination (Hess et al 1988 is an excellent counterexample). In this section,
we therefore review these two literatures separately. We also consider some alternatives.
Emotions as Causes of Facial Expressions
Measurement. Techniques for facial measurement were slow to develop. Some systems
provide not an objective description of facial movement, but a description in terms of the
emotion (Izard 1979) or affective dimension (AM Kring & D Sloan, unpublished manuscript,
Russell, Bachorowski, & Fernandez Dols 15
Univ. of California, Berkeley) inferred, typically from clusters of physically different
movements. An objective (but intrusive) technique is electromyography, especially useful for
brief or small muscular movement (Fridlund & Cacioppo 1986, Tassinary & Cacioppo 1992). An
objective and unobtrusive technique useful for visible movements was developed by Hjortsjö
(1969) based on facial anatomy. This technique was subsequently revised and renamed the
Facial Action Coding System (FACS) by Ekman and Friesen (1978). An updated version of
FACS was recently announced (Ekman et al 2002). H Oster and D Rosenstein (unpublished
manuscript, Adelphi Univ.) developed a version of FACS for infant faces. Still another system is
Katsikitis and Pilowsky’s (1988) FACEM, which assesses facial movement in terms of 12
distances between key points on the face.
Positive Emotions. Does happiness produce a smile? There is a clear association between
pleasant feelings and zygomatic activity (smiling) (Davidson et al 1990, Lang et al 1993,
Winkielman & Cacioppo 2001). Yet, the relation is far from simple, and happiness is neither
necessary nor sufficient for smiling. Kraut and Johnson (1979) found surprisingly few smiling
faces among bowlers and hockey fans during happy events – unless they were simultaneously
engaged in social interaction (replicated in Spain by Ruiz Belda et al 2002). Fernandez Dols and
Ruiz Belda (1995) similarly found smiles limited to social circumstances even for ecstatically
happy persons: gold-medal winners at the Olympics. Even in children, smiling is more
associated with the particular social interaction in which they are engaged than with their own
happiness (Schneider & Unzner 1992, Soussignan & Schaal 1996): Children smile as much after
failure as after success, but smiling is coordinated with eye contact (Schneider & Josephs 1991).
Smiling also occurs during humorous films in proportion to self-reported amusement (Ekman et
al 1990), but this same study found little smiling during another pleasant but non-humorous film.
Russell, Bachorowski, & Fernandez Dols 16
The proposal that a Duchenne smile (in which zygomatic activity is combined with
contraction of the orbicularis oculi) is the index of happiness has yielded mixed results. This
distinction has not clarified the causes of smiling or laughter (Keltner & Bonnano 1997,
Rosenberg et al 2001) and does not nullify the general finding reviewed in the last paragraph of
the social nature of smiles. For instance, Duchenne smiles too occur as much after failure as after
success (Schneider & Josephs 1991) and may simply be more intense smiles (Schneider &
Unzner 1992).
Negative Emotions. Tomkins’ theory predicts that negative basic emotions – fear, anger,
sadness, disgust, and, possibly, contempt, shame, and embarrassment – each produce a distinct
signal. One interesting examination of this prediction was Camras’s (1991) year-long
observational study of her own daughter. Camras found “(1) situations in which we believe an
emotion is present yet the facial expression is not seen, and (2) situations in which an expression
is observed but does not appear to be best described using the discrete emotion categories of
differential emotion theory” (Camras, 1991, p.26). In another study, 30 babies were subjected to
an arm restraint procedure, to which each baby reacted with distress (Camras et al 1992). Rather
than one common pattern of facial response, however, there were many different patterns, few of
which fit the criteria for a discrete emotion signal. Clearly, there is great need for ecological
research on what facial activity occurs and under what circumstances.
In the laboratory, researchers have tested Tomkins’ predictions by using films, slides, and
remembered or imagined events to induce emotion, but with similarly weak results (e.g.,
Fernandez Dols et al 1997). Rosenberg and Ekman (1994) criticized prior laboratory research but
also offered new supporting evidence. While participants viewed four films selected to induce
intense negative emotions, their faces were surreptitiously recorded. Participants then watched
Russell, Bachorowski, & Fernandez Dols 17
the films again. This time, they reported each emotion they had experienced and pinpointed the
time of its occurrence. What was the proportion of hits? That is, consider all occasions on which
participants reported a negative emotion. On what proportion of these occasions did they show
the facial expression predicted for that emotion? This figure was not given. Instead, the
proportion of hits within a selected subset of these occasions was given, namely, those occasions
in which a negative emotion was reported and some facial expression had occurred. For these
selected occasions, the proportion of hits was .42 (p > .10) for one film and .50 (p < .05) for
another; no figures were given for the remaining two films. Clearly these figures are inflated
because the excluded occasions (on which an emotion was reported but no facial expression
occurred) were all misses. (Also ignored were occasions on which a facial expression occurred
but no emotion was reported.) In addition, it was not clear exactly which facial expressions were
considered correct and which incorrect for a given emotion.
A more promising approach capitalized on the fact that some subjects become
embarrassed in the laboratory when asked to pose facial expressions (Keltner 1995). In these
cases, embarrassment was associated not with a single static configuration (something that could
be captured well in a photograph or painting) but with a complex sequence of face and body
movements.
Surprise. In the most sophisticated set of laboratory studies on this topic to date,
Reisenzein (2000) addressed prior technical criticisms and examined the coherence among four
components of surprise: cognitive appraisal of the stimulus as unexpected, self report of surprise,
reaction time, and facial expression. Reisenzein found, “Even with an optimal data analysis
design (raw data, within-subjects), the average linear correlations between the different surprise
components were – with the important exception of the correlation between [a cognitive
Russell, Bachorowski, & Fernandez Dols 18
appraisal of] unexpectedness and [self-reported] surprise (r = .78) – only low to moderate,
ranging from .19 ([reaction time]-expression) to .46 (surprise feeling-expression)” (p. 28).
Emotions as Causes of Vocal Expressions
Measurement. Measurement of vocal acoustics is guided by the source-filter model
developed in the 1950s (Fant 1960, Stevens & House 1955, Titze 1994). The “source” refers to
the vocal folds, which vibrate in a quasi-periodic fashion during phonation. The rate of vibration
directly corresponds to the fundamental frequency (F0), and is highly correlated with the
perception of pitch. Mean F0 and measures of F0 variability have been the most commonly
studied acoustic cues in research on vocal EEs. More recent advances in digital cue-extraction
and modeling techniques have made it increasingly feasible to measure these and other sound
properties of interest, including minute perturbations in the amplitude and frequency of vocal-
fold vibration (Bachorowski & Owren 1995, Protopapas & Lieberman 1997) and glottal airflow
characteristics (Cummings & Clemments 1995).
The resonance properties of the various cavities and articulators in the supralaryngeal
vocal tract contribute to “filter” effects, which are typically indexed by formant frequencies (see
Johnson 1997, Lieberman & Blumstein 1988). Recently, emphasis has also been given to the
long-term average spectrum (LTAS), which represents the average distribution of energy over
the course of continuous speech (Pittam & Scherer 1993). LTAS assessment has the advantage
over most other measures of being quick and less susceptible to measurement error. A significant
drawback, however, is that LTAS does not directly correspond to sound production at any given
moment.
Nonlinguistic Vocalizations. Laughs, cries, sighs, yawns, and other such vocal outbursts
seem at first to be good examples of expressions of discrete (although not necessarily basic)
Russell, Bachorowski, & Fernandez Dols 19
emotions: A funny joke elicits amusement, which produces a laugh; a loss elicits sadness, which
produces crying; an uninspired lecture elicits boredom, which produces a yawn. Mounting
evidence, however, questions whether such vocalizations are each linked to a specific, discrete
state.
Infant crying is a good illustration of this conclusion. Different cries were once thought
associated with different states, such as frustration, fear, hunger, cold, pain, fatigue, or a soiled
diaper (Berry 1975). The evidence instead is that the cry more simply indexes the degree of the
infant’s distress (Barr et al 2000). The cry’s typical acoustic features (abrupt onset, high F0, high
amplitude, and characteristic pulsing) serve to attract the attention of and to cause negative affect
in the receiver. The marked variability in these acoustic features serves not to mark different
states (frustration, etc) but to lessen the chances of the receiver habituating. The receiver then
infers the infant’s specific state largely from context (Bachorowski & Owren 2002).
Laughter also illustrates this conclusion. Laughs are produced not only by humor, but
also by anger and anxiety (Darwin 1872), attempted self-deprecation (Glenn 1991/1992),
attention (Martin & Gray 1996), appeasement or submission (Adams & Kirkevold 1978, Deacon
1997, Dovidio et al 1988, Grammer & Eibl-Eibesfeldt 1990) and sexual interest (Dunbar 1996,
Grammer 1990, Grammer & Eibl-Eibesfeldt 1990). From this variety, one might be tempted to
hypothesize that different types of laughs correspond to different states. We know of no
empirical support for this hypothesis. Although laugh acoustics are remarkably variable both
within and between laughers (Bachorowski et al 2001, Grammer & Eibl-Eibesfeldt 1990), they
have not been found to vary as a function of self-reported emotion. Instead, laughter varies with
social factors such as the sex of and familiarity with one’s social partner (J-A Bachorowski et al,
unpublished manuscript, Vanderbilt Univ., Devereux & Ginsburg 2001). Laughs also provide
Russell, Bachorowski, & Fernandez Dols 20
cues to individuality (Edmonson 1987) and elicit affective responses in listeners (Bachorowski &
Owren 2001). The emerging picture is one in which laughter serves to elicit cooperation and a
positive relationship with a specific receiver (Owren & Bachorowski 2001).
Vocal Expression in Speech. Many studies have examined the vocal characteristics of
speech in hopes of defining a vocal signature for each basic emotion. Leinonen et al (1997) and
Johnstone and Scherer (2000) recently provided detailed summaries. So far, the strongest single
association found for vocal acoustics has been with the sender’s general arousal level
(Bachorowski 1999, Cowie 2000, Kappas et al 1991). The still unanswered question is whether
reliable patterns beyond this simple relationship can be established.
A pattern of vocal cues unique to a valence (pleasure) dimension has proven elusive
(Bachorowski 1999, Leinonen et al 1997, Millot & Brand 2001, Paeschke & Sendlmeier 2000,
Pereira 2000, Protopapas & Lieberman 1997, Tolkmitt & Scherer 1986, Trouvain & Barry
2000). For example, anger and joy can both produce high F0 and high amplitude. These basic
acoustic effects have been shown for acted portrayals as well as naturally produced speech
(Johnstone & Scherer 2000, Scherer 1989) and suggest that the speech acoustics reflect what joy
and anger have in common (such as arousal). In a study of speech utterances produced
immediately after affectively charged success or failure feedback, changes in three F0-related
measures reflected increases in arousal (Bachorowski & Owren 1995). Effects associated with
valence were more ambiguous and depended on interactions with talker sex and trait differences
in emotional intensity.
Vocal differences due to arousal and valence are consistent with a dimensional account of
emotion, and therefore a test of predictions from Tomkins’ theory best focuses on negative
emotions. An important study by Banse and Scherer (1996) typifies this work. Twelve
Russell, Bachorowski, & Fernandez Dols 21
professional stage actors were asked to read two sentences for each of 28 scenarios (14 emotions
X 2 scenarios per emotion). A large number of acoustic features was examined. As expected, the
major differentiation was provided by mean F0, which most reliably indexes arousal. In addition,
statistically significant but modest differentiation occurred for separate emotions. Nevertheless,
two factors render this result questionable. First, acted portrayals do not necessarily correspond
to naturally produced vocal EEs; after all, the actors’ job was to convey emotion. Second, tests of
differentiated patterning were based not on all 1,344 vocal samples obtained but on a subset of
224 judged as the best acted.
Recent studies have attempted to induce emotion in the laboratory rather than to merely
simulate it using actors (Bachorowski & Owren 1995, Kappas 1997, Millot & Brand 2001, Sobin
& Alpert 1999) or to analyze recordings made outside the laboratory such as in radio and
television interviews (Gregory & Webster 1996) or horse-race commentaries (Trouvain & Barry
2001). These studies again confirm the link of vocal expression with sender’s general arousal,
and, importantly, sometimes show different patterns than those obtained with acted portrayals
(Streeter et al 1983). A vocal signature for each hypothesized basic emotion, however, remains
elusive.
Future Directions
Kappas (2002, p.10) summarized research on emotion and facial movement: “We might
be on safer ground than simply insisting, against our better knowledge, that there are fixed links
between facial expression and emotions.” The theory that a small number of discrete emotions
produce a corresponding set of facial signals has yielded at best weak results. And outside
Western societies, there is practically no relevant evidence. The evidence on vocal outbursts and
Russell, Bachorowski, & Fernandez Dols 22
on vocal characteristics of speech corroborates this trend. This same research, however, has
provided some hints of more promising directions.
First, research should not be limited to a small list of emotions or small set of signals. For
example, research in which the face or voice is the dependent variable has found many different
patterns. The traditional view of facial expressions focused on static visual configurations (the
sort of thing that can be captured in a painting, or later, in a photograph), one per emotion. Thus
much is neglected, such as blushing, paling, and blinking (see, for example, Leary et al 1992).
We suggest a much broader focus on the many possible dynamic patterns in nonverbal facets of
action. Although, in many respects, facial and vocal systems are separate, more study of their
joint occurrence (evident in laughter, sobbing, and yelling) is also needed.
In analyzing EEs, it is helpful to move beyond overly simple dichotomies. In response to
technical criticisms of research claimed to support Tomkins’ theory (Russell 1994) and to the
presentation of an alternative to that theory (Fridlund 1994), Rosenberg (1997, p. 88) stated,
“implicit in Russell’s argument and explicit in Fridlund’s is the notion that the face has nothing
to do with emotion.” Of course, there is some association between EEs and emotion, but the
question is the nature of that association.
Perhaps facial and vocal changes are more closely tied with what have been thought of as
components of the emotion and thus only indirectly with emotion per se. One possible such
component is the simple core affect of pleasure and activation (Bradley & Lang 2002b, Russell
2002). Another is the various cognitive steps involved in the processing of the emotion-eliciting
stimulus (Scherer 2001, Smith & Scott 1997). Still another is preparation for instrumental action
(Frijda & Tcherkassoff 1997). Because the components of emotion are at best loosely associated,
Russell, Bachorowski, & Fernandez Dols 23
one can find individual components in the absence of a full emotion and vice versa. The
interesting empirical question is what facial and vocal changes occur in these dissociated cases.
Evidence on the social nature of smiles suggests that more scrutiny of social norms and
context is warranted for EEs in general. This topic has typically been discussed through an idea
introduced by Klineberg (1940) and later named “display rules” (Ekman 1972): that persons
learn to voluntarily inhibit, produce, or alter their natural EEs. EEs are clearly influenced by
culture (Kupperbusch et al 1999), gender (LaFrance & Hess 1999), and group membership
(Kirouac & Hess 1999) – although whether “display rules” are the explanation remains largely
untested and perhaps untestable (Fridlund 1994).
Owren and Bachorowski’s (2001) account of smiling builds on modern evolutionary
theory. Two different but related systems underlie the smile (Rinn 1984). Smiles produced by
either system manipulate receiver affect. This account is thus consistent with evidence that
smiles are highly dependent on the presence of an audience (although that audience can be
psychologically rather than physically present, Fridlund 1994). A phylogenetically older,
simpler, reflex-like system produces “spontaneous” smiles as reliable signs of positive feelings
toward a specific receiver. Positive affect is therefore necessary but not sufficient for their
production. The second system is a more recently evolved version of the first in which
“volitional” smiles are produced in a controlled process. In contrast to spontaneous smiles,
volitional smiles are emancipated from affect in that they can occur during the experience of any
affective state. Sometimes thought of as being “deceptive” or “dishonest,” the power of
volitional smiles lies in their inherent unreliability as a cue to the sender’s state.
Fridlund’s (1994) evolutionary account places a similar emphasis on the smile being
directed at a receiver but substitutes “friendly intentions” for “positive feelings.” Although not
Russell, Bachorowski, & Fernandez Dols 24
denying that emotions and feelings exist or are correlated with EEs, Fridlund argued that the
most coherent causal story can be told in more behavior-relevant or functional terms. Fridlund
applied the same analysis to other EEs as well, centered on other social intentions, including
aggression, appeasement, and help-seeking. Perhaps because this account strays from the
traditional assumptions associated with EEs and maintained in previous accounts, it has been
frequently misunderstood. For example, “intentional” should not be taken to mean a conscious
state, but simply involving a behavioral disposition aimed at a specific receiver. Fridlund’s
account does not require a simple correlation between the amount of signaling and the degree of
sociality of the situation. Nor does Fridlund’s account deny that EEs can occur when the sender
is alone. Indeed, he offered evidence that EEs produced when alone are directed at an imaginary,
implicit, or animistic audience. Like Owren and Bachorowski’s, Fridlund’s account suggests the
power of modern evolutionary theory to overturn long held assumptions and open the door to
fresh perspectives on EEs.
Conclusion
The scientific study of emotional expressions has been pursued now for about two
centuries (e.g., in 1806, Bell published on the anatomical basis of facial expression). During
most of that time, the field was in the grip of an ancient set of assumptions, long ago
incorporated into common sense and embedded in our language: That “emotional expressions”
express emotions is a tautology, hardly something that seems to require empirical verification.
Even the best scientists, including Darwin, implicitly held these presuppositions. Of course,
science can progress even with dubious assumptions, but only so far. Emotional expressions may
not be expressions and may not be related to emotions in any simple way.
Russell, Bachorowski, & Fernandez Dols 25
Modern views of evolution are supplanting Darwin’s 130-year-old analysis. Rather than
broadcast for the benefit of anyone who happens to observe them, at least some EEs are directed
at a specific receiver. These serve to shape the affective and behavioral stance of the receiver and
likely evolved to do so. Other EEs are simply given off, by-products of actions performed
without communicative purpose. Still other EEs likely have other causal histories.
It is unlikely that the receiver simply decodes an emotional message in any simple,
reflex-like manner. There are quick, simple, and automatic responses to EEs, but these cannot be
assumed to reflect “decoding the emotion.” Receivers do sometimes attribute emotion to senders,
but doing so is not always quick or simple. Receivers make a variety of interpretations of an EE
besides emotional ones.
Of course, most of our conclusions here are tentative and await empirical test. Probably
the more important development in the study of EEs in the last decade is a shift in perspective.
Old assumptions need to be critically scrutinized, new ideas encouraged and pursued, rather than
vice versa. What is exciting is that the hold of the “vice versa” on the field is steadily yielding.
Russell, Bachorowski, & Fernandez Dols 26
Literature Cited
Adams RM, Kirkevold B. 1978. Looking, smiling, laughing, and moving in restaurants:
Sex and age differences. Envir. Psych. and Nonverb. Beh., 3:117-21
Bachorowski J-A. 1999. Vocal expression and perception of emotion. Cur. Dir. Psychol Sci.
8:53-57
Bachorowski J-A, & Owren MJ. 1995. Vocal expression of emotion: Acoustic properties of
speech are associated with emotional intensity and context. Psychol. Sci. 6:219-24
Bachorowski J-A, & Owren MJ. 2001. Not all laughs are alike: Voiced but not unvoiced laughter
readily elicits positive affect. Psychol. Sci. 12:252-57
Bachorowski J-A, & Owren MJ. 2002. The role of vocal acoustics in emotional intelligence. In
The Wisdom of Feelings: Processes Underlying Emotional Intelligence, eds. LF Barrett, P
Salovey. New York: Guilford. In press
Bachorowski J-A, Smoski MJ, & Owren MJ. 2001. The acoustic features of human laughter. J.
Acoust. Soc. Am. 110:1581-97
Banse R, & Scherer KR. 1996. Acoustic profiles in vocal emotion expression. J. Pers. Soc.
Psychol. 70:614-36
Barr RG, Hopkins B, Green JA, eds. 2000. Crying as a Sign, a Symptom, & a Signal. New York:
Cambridge Univ.
Bell C. 1806. Essays on the Anatomy of Expression in Painting. London: Longman, Hurst, Rees,
and Orme.
Berry KK. 1975. Developmental study of recognition of antecedents of infant vocalizations.
Percept. Mot. Skills 41:400-402
Bradley MM, Lang PJ. 2000a. Affective reactions to acoustic stimuli. Psychophysiology 37:204-
Russell, Bachorowski, & Fernandez Dols 27
15
Bradley MM, Lang PJ. 2000b. Measuring emotion: Behavior, feeling, and physiology. In
Cognitive Neuroscience of Emotion, eds. RD Lane, L Nadel, pp. 242-76. New York: Oxford
Univ.
Calder AJ, Young AW, Perrett DI, Etcoff NL, Roland D. 1996. Categorical perception
of morphed facial expressions. Vis. Cogn. 3:81-117
Camras LA. 1991. A dynamical systems perspective on expressive development. In
International Review of Studies on Emotion, ed. KT Strongman, pp. 16-28. New York: Wiley
Camras LA, Oster H, Campos JJ, Miyake K, Bradshaw D. 1992. Japanese and American infants’
responses to arm restraint. Dev. Psychol. 28:578-83
Carroll JM, Russell JA. 1996. Do facial expressions signal specific emotions? Judging the face in
context. J. Pers. Soc. Psychol. 70:205-18
Chong SCF, Werker JF, Russell JA, Carroll JM. 2002. Three facial expressions mothers direct to
their infants. Work. Pap., Dep. Psych. Univ. British Columbia
Cooper RP, Aslin RN. 1990. Preference for infant-directed speech in the first month after birth.
Child Dev. 61:1584-95
Cowie R. 2000. Describing the emotional states expressed in speech. See Cowie et al 2000, pp.
11-18
Cowie R, Douglas-Cowie E, Schröder M, eds. 2000. Proc. ISCA Workshop on Speech Emot.
Newcastle, Northern Ireland. Belfast: Textflow
Cummings KE, Clements MA. 1995. Analysis of the glottal excitation of emotionally styled and
stressed speech. J. Acoust. Soc. Am. 98:88-98
Darwin C. 1872. The Expression of the Emotions in Man and Animals. London: Murray.
Russell, Bachorowski, & Fernandez Dols 28
Davidson RJ. 2000. The functional neuroanatomy of affective style. In Cognitive Neuroscience
of Emotion, eds. RD Lane, L Nadel, pp. 371-88. New York: Oxford Univ.
Davidson RJ, Ekman P, Saron CD, Senulis JA, Friesen WV. 1990. Approach-withdrawal and
cerebral asymmetry: Emotional expression and brain physiology I. J. Pers. Soc. Psychol.
58:330-41
Dawkins R, Krebs JR. 1978. Animal signals: Information or manipulation? In Behavioral
Ecology: An Evolutionary Approach, eds. JR Krebs, NB Davies, pp. 282-309. London:
Blackwell Scientific
Deacon TW. 1997. The Symbolic Species. New York: Norton
Devereaux PG, Ginsburg GP. 2001. Sociality effects on the production of laughter. J. Gen.
Psychol. 128:227-40
Dimberg U, Öhman A. 1996. Behold the wrath: Psychophysiological responses to facial stimuli.
Motiv. Emot. 20:149-82
Dimberg U, Thunberg M, Elmehed K. 2000. Unconscious facial reactions to emotional facial
expressions. Psychol. Sci. 11:86-89
Dovidio JF, Brown CE, Heltman K, Ellyson SL, Keating, CF. 1988. Power displays between
women and men in discussions of gender-linked tasks: A multichannel study. J. Pers. Soc.
Psychol. 55:580-87
Dunbar R. 1996. Grooming, Gossip, and the Evolution of Language. Cambridge: Harvard Univ.
Edmonson, MS. 1987. “Notes on laughter,” Anthro. Ling. 29:23-33
Ekman P. 1975. Face muscles talk every language. Psychol. Today 9:35-39
Ekman P. 1994. Strong evidence for universals in facial expressions: A reply to Russell’s
mistaken critique. Psychol. Bull. 115:268-87
Russell, Bachorowski, & Fernandez Dols 29
Ekman P. 1997. Should we call it expression or communication? Innovation 10:333-44
Ekman P, Friesen WV, Ellsworth P. 1972. Emotion in the human face. Elmsford, NY: Pergamon
Ekman P, Davidson RJ, Friesen WV. 1990. The Duchenne smile: Emotional expression and
Brain Physiology II. J. Pers. Soc. Psychol. 58: 342-53
Ekman P, Friesen WV. 1978. Facial Action Coding System: A Technique for the Measurement of
Facial Movement. Palo Alto, CA: Consulting Psychologists
Ekman P, Friesen WV, Hager JC. 2002. New Version of the Facial Action Coding System.
http://dataface.nirc.com/Expression/FACS/New_Version/new_version.html
Ekman P, Oster H. 1979. Facial expressions of emotion. Ann. Rev. Psychol. 30:527-54
Elfenbein HA, Ambady N. 2002. On the universality and cultural specificity of emotion
recognition: A meta-analysis. Psychol. Bull. 128:203-35
Ellgring H. 1986. Nonverbal expression of psychological states in psychiatric patients. Eur.
Arch. Psychiatr. Neurol. Sci. 236:31-34
Fant G. 1960. Acoustic Theory of Speech Production. The Hague: Mouton
Fernald A. 1991. Prosody in speech to children: Prelinguistic and linguistic functions. In Annals
of Child Development, Vol. 8, ed. R Vasta, pp. 43-80. Philadelphia: Jessica Kingsley
Fernald A, Kuhl PK. 1987. Acoustic determinants of infant preference for motherese speech.
Infant Beh. Dev. 10:279-93
Fernández-Dols JM, Ruiz-Belda MA. 1995. Are smiles a sign of happiness? Gold medal winners
at the Olympic Games. J. Pers. Soc. Psychol. 69:1113-19
Fernández-Dols JM, Sanchez MA, Carrera P, Ruiz-Belda MA. 1997. Are spontaneous
expressions and emotions linked? An experimental test of coherence. J. Nonver. Behav.
21:163-77
Russell, Bachorowski, & Fernandez Dols 30
Frank MG, Stennett J. 2001. The forced-choice paradigm and the perception of facial expression
of emotion. J. Pers. Soc. Psychol. 80:75-85
Fridlund AJ. 1994. Human Facial Expression: An Evolutionary View. New York: Academic
Press
Frijda NA. 1969. Recognition of emotion. In Advances in Experimental Social Psychology, ed. L
Berkowitz, 4:167-223. New York: Academic
Frijda NH, Tcherkassof A. 1997. Facial expressions as modes of action readiness. See Russell &
Fernandez-Dols 1997, pp. 78-102
Glenn PJ. 1991/1992. Current speaker initiation of two-party shared laughter. Res. Lang. and
Social Inter. 25:139-62
Goffman E. 1959. The Presentation of Self in Everyday Life. Garden City, NY: Doubleday.
Grammer K. 1990. Strangers meet: Laughter and nonverbal signs of interest in opposite-sex
encounters. J. of Nonverb. Beh. 14:209-36
Grammer K, Eibl-Eibesfeldt I. 1990. The ritualization of laughter. In Naturlichkeit der Sprache
und der Kultur: acta colloquii, ed. W Koch, pp. 192-214. Bochum: Brockmeyer
Gregory SW, Webster S. 1996. A nonverbal signal in voices of interview partners effectively
predicts communication accomodation and social status perception. J. Pers. Soc. Psychol. 70:
1231-40
Hatfield E, Cacioppo JT, Rapson RL. 1992. Primitive emotional contagion. In Emotion and
Social Behavior, ed. MS Clark, pp. 151-177. London: Sage
Hebb DO. 1946. Emotion in man and animal: An analysis of the intuitive processes of
recognition. Psychol. Rev. 53:88-106
Hess U, Kirouac G. 2000. Emotion expression in groups. See Lewis & Haviland-Jones 2000, pp.
Russell, Bachorowski, & Fernandez Dols 31
368-81
Hess U, Scherer KR, Kappas A. 1988. Multichannel communication of emotion: Synthetic signal
production. In Facets of emotion: Recent research, ed. KR Scherer, pp. 161-249. Hillsdale,
NJ: Erlbaum
Hjortsjö, CH. 1969. Man’s Face and Mimic Language. Lund, Sweden: Studentlitteratur
Izard CE. 1971. The Face of Emotion. New York: Appleton Century Crofts
Izard CE. 1979. The Maximally Discriminative Facial Movement Coding System (MAX).
Newark, DE: Univ. of Delaware, Comp. and Network Serv., Univ. Media Serv.
Izard CE. 1994. Innate and universal facial expressions: Evidence from developmental and cross-
cultural research. Psychol. Bull. 115:288-99
Johnson K. 1997. Acoustic & Auditory Phonetics. Cambridge: Blackwell Publishers. 169 pp.
Johnson WF, Emde RN, Scherer KR, Klinnert MD. 1986. Recognition of emotion from vocal
cues. Arch. Gen. Psychiatry 43:280-83
Johnstone T, Scherer KR. 2000. Vocal communication of emotion. See Lewis & Haviland-Jones
2000, pp. 220-35
Kaplan PS, Bachorowski J-A, Smoski MJ, Hudenko, WJ. 2002. Infants of depressed mothers,
although competent learners, fail to learn in response to their own mother’s infant-directed
speech. Psychol. Sci. 13:268-71
Kaplan PS, Zarlengo-Strouse P, Kirk LS, Angel CL. 1997. Selective and nonselective
associations between speech segments and faces in human infants. Dev. Psychol. 33:990-99
Kappas A. 1997. His master’s voice: Acoustic analysis of sponataneous volcaizations in an
ongoing active coping task. Presented at Annu. Meet. Soc. Psychophys. Res. 37th, Cape Cod,
Mass.
Russell, Bachorowski, & Fernandez Dols 32
Kappas A. 2002. What facial activity can and cannot tell us about emotions. In The Human
Face: Measurement and Meaning, ed. M Katsikitis. Norwell, MA: Kluwer. In press
Kappas A, Hess U, Scherer KR. 1991. Voice and emotion. In Fundamentals of Nonverbal
Behavior, eds. B Rime, R Feldman, pp. 200-38. Cambridge: Cambridge Univ. Press
Katsikitis M, Pilowski I. 1988. A study of facial expression in Parkinson’s disease using a novel
microcomputer-based method. J. Neur. Neurosurg. Psychiatry 51:362-366
Katz GS, Cohen JF, Moore CA. 1996. A combination of vocal F0 dynamic and summary features
discriminates between three pragmatic categories of infant-directed speech. Child
Dev. 67:205-17
Keltner D. 1995. Signs of appeasement: Evidence for the distinct displays of embarrassment,
amusement, and shame. J. Pers. Soc. Psychol. 68:441-54
Keltner D, Bonanno G. 1997. A study of laughter and dissociation: Distinct correlates of laughter
and smiling during bereavement. J. Pers. Soc. Psychol. 73:687-702
Kraut RE, Johnston RE. 1979. Social and emotional messages of smiling: An ethological
approach. J. Pers. Soc. Psychol. 37:1539-53
LaFrance M, Hecht MA. 1995. Why smiles generate leniency. Pers. Soc. Psychol. Bull.
21:207-14
Lang PJ, Greenwald MK, Bradley MM, Hamm AO. 1993. Looking at pictures: Affective, facial,
visceral, and behavioral reactions. Psychophysiology 30:261-73
Leary MR, Britt TW, Cutlip WD, Templeton JL. 1992. Social blushing. Psychol. Bull. 11:446-
460
Leinonen L, Hiltunen T, Linnankoski I, Laakso, M-L 1997. Expression of emotional-
motivational connotations with a one-word utterance. J. Acoust. Soc. Am. 102:1853-63
Russell, Bachorowski, & Fernandez Dols 33
Levenson RW. 1996. Biological substrates of empathy and facial modulation of emotion: Two
facets of the scientific legacy of John Lanzetta. Motiv. Emot. 20:185-204
Lewis M, Haviland-Jones JM, eds. 2000. Handbook of Emotions. New York: Guilford. 2nd ed.
Lieberman P, Blumstein SE (1988). Speech Physiology, Speech Perception, and Acoustic
Phonetics. Cambridge: Cambridge Univ.
Marler P, Evans Ch. 1997. Animal sounds and human faces: Do they have anything in common?
See Russell & Fernandez-Dols 1997, pp. 133-157
Martin GN, Gray CD. 1996. The effect of audience laughter on men’s and women’s response to
humor. J. Soc. Psychol, 136:221-31
Millot J-L, Brand G. 2001. Effects of pleasant and unpleasant ambient odors on human voice
pitch. Neurosci. Let. 297:61-63
Montagu J. 1994. The Expression of the Passions: The Origin and Influence of Charles LeBrun’s
Conférence sur l’Expression Générale et Particulière. New Haven, CN: Yale Univ.
Motley MT, Camden CT. 1988. Facial expression of emotion: A comparison of posed
expressions versus spontaneous expressions in an interpersonal communication setting.
West. J. Speech Comm. 52:1-22
Murray L, Fiori-Cowley A, Hooper R, Cooper P. 1996. The impact of postnatal depression and
associated adversity on early mother-infant interactions and later infant outcomes. Child Dev.
67:2512-26
Nelson CA, de Haan M. 1997. A neurobehavioral approach to the recognition of facial
expressions in infancy. See Russell & Fernandez-Dols 1997, pp. 176-204
Neumann R, Strack F. 2000. “Mood contagion:” The automatic transfer of mood between
persons. J. Pers. Soc. Psychol. 79:211-23
Russell, Bachorowski, & Fernandez Dols 34
Niedenthal PM, Halberstadt JB, Margolin J, Innes-Ker AH. 2000. Emotional state and the
detection of change in facial expression of emotion. Eur. J. Soc. Psychol. 30:211-22
Owren MJ, Bachorowski J-A. (2001). The evolution of emotional expression: A “selfish-gene”
account of smiling and laughter in early hominids and humans. In Emotions: Current Issues
and Future Directions, eds. TJ Mayne, GA Bonanno, pp. 152-91. New York: Guilford
Owren MJ, Rendall D. 1997. An affect-conditioning model of nonhuman primate signaling. In
Perspectives in Ethology, Vol. 12: Communication, eds. DH Owings, MD Beecher, NS
Thompson, pp. 299-346. New York: Plenum
Owren MJ, Rendall D. 2001. Sound on the rebound: Bringing form and function back to the
forefront in understanding nonhuman primate vocal signaling. Evo. Anthro. 10:58-71
Owren MJ, Rendall D, Bachorowski J-A. 2002. Vocal communication: A non-representational
account of signaling in primates and humans. In Primate Psychology: Bridging the Gap
Between the Mind and Behavior of Human and Nonhuman Primates, ed. D Maestripieri,
Cambridge: Harvard Univ. In press
Pakosz M. 1983. Attitudinal judgments in intonation: Some evidence for a theory. J. Psycholing.
Res. 12:311-26
Paeschke A, Sendlmeier WF. 2000. Prosodic characteristics of emotional speech:
Measurements of fundamental frequency movements. See Cowie et al 2000, pp. 75-80
Pereira C. 2000. Dimensions of emotional meaning in speech. See Cowie et al 2000, pp. 25-28
Pittam J, Gallois C, Callan V. 1990. The long-term spectrum and perceived emotion. Speech
Comm. 9:177-87
Pittam J, Scherer KR. 1993. Vocal expression and communication of emotion. In Handbook of
Emotions, eds. M Lewis, JM Haviland, pp. 185-97. New York: Guilford. 1st ed.
Russell, Bachorowski, & Fernandez Dols 35
Protopapas A, Lieberman P. 1997. Fundamental frequency of phonation and perceived emotional
stress. J. Acoust. Soc. Am. 101:2267-77
Provine RR. 1997. Yawns, laughs, smiles, tickles, and talking: Naturalistic and laboratory studies
of facial action and social communication. See Russell & Fernandez-Dols 1997, pp. 158-75
Reisenzein R. 2000. Exploring the strength of association between the components of emotion
syndromes: The case of surprise. Cogn. Emot. 14:1-38
Rinn WE. 1984. The neuropsychology of facial expression. Psychol. Bull. 95:52-77
Rosenberg EL. 1997. Emotions as unified responses. In What the Face Reveals, eds. P Ekman,
EL Rosenberg, pp. 86-88. New York: Oxford
Rosenberg EL, Ekman P. 1994. Coherence between expressive and experiential systems in
emotion. Cogn. Emot. 8:201-29
Rosenberg EL, Ekman P, Jiang W, Babyak M, Coleman RE, et al 2001. Linkages between facial
expressions of anger and transient myocardial ischemia in men with coronary disease.
Emotion 1:107-15
Ruiz Belda MA, Fernandez-Dols JM, Carrera P, Barchard K. 2002. Spontaneous facial
expressions of happy bowlers and soccer fans. Cogn. Emot. In press
Russell JA. 1991. Culture and the categorization of emotion. Psychol. Bull. 110:426-50
Russell JA. 1994. Is there universal recognition of emotion from facial expressions? A review of
cross-cultural studies. Psychol. Bull. 115:102-41
Russell JA. 1995. Facial expressions of emotion: What lies beyond minimal universality?
Psychol. Bull. 118:379-99
Russell JA. 1997. Reading emotions from and into faces: Resurrecting a dimensional-
contextual perspective. See Russell & Fernandez Dols 1997, pp. 295-320
Russell, Bachorowski, & Fernandez Dols 36
Russell JA. 2002. Core affect and the psychological construction of emotion. Psychol. Rev. In
press
Russell JA, Bullock M. 1986. On the dimensions preschoolers use to interpret facial expressions
of emotion. Dev. Psychol. 22:97-102
Russell JA, & Feldman Barrett L. 1999. Core affect, prototypical emotional episodes, and other
things called emotion: Dissecting the elephant. J. Pers. Soc. Psychol. 76:805-19
Russell JA, Fernández-Dols JM, eds. 1997. The Psychology of Facial Expression. New York:
Cambridge Univ.
Russell JA, Fernández-Dols JM. 1997. What does a facial expression mean? See Russell &
Fernandez-Dols 1997, pp. 3-30
Scherer KR. 1989. Vocal measurement of emotion. In Emotion: Theory, Research, and
Experience Vol. 4. The Measurement of Emotions, eds. R Plutchik, H Kellerman, pp. 233-59.
New York: Academic
Scherer KR. 2001. Appraisal considered as a process of multilevel sequential checking. In
Appraisal Processes in Emotion, eds. KR Scherer, A Schorr, T Johnstone, pp. 92-120. New
York: Oxford Univ.
Schneider K, Josephs I. 1991. The expressive and communicative functions of preschool
children’s smiles in an achievement-situation. J. Nonverb. Behav. 15:185-98
Schneider K, Unzner L. 1992. Preschoolers’ attention and emotion in an achievement and an
effect game: A longitudinal study. Cogn. Emot. 6:37-63
Schröder M. 2000. Experimental study of affect bursts. See Cowie et al 2000, pp. 132-35
Smith CA, Kirby LD. 2001. Toward delivering on the promise of appraisal theory. In
Appraisal Processes in Emotion, eds. KR Scherer, A Schorr, T Johnstone, pp. 121-38. New
Russell, Bachorowski, & Fernandez Dols 37
York: Oxford Univ.
Smith CA, Scott HS. 1997. A componential approach to the meaning of facial expressions. See
Russell & Fernandez-Dols 1997, pp. 229-54
Sobin C, Alpert M. 1999. Emotion in speech: The acoustic attributs of fear, anger, sadness, and
joy. J. Psycholing. Res. 28:347-65
Soussignan R, Schaal B. 1996. Forms and social signal value of smiles associated with pleasant
and unpleasant sensory experience. Ethology 102:1020-41
Stevens KN, House AS. 1955. Development of a quantitative description of vowel articulation. J.
Acoust. Soc. Am. 27:484-493
Streeter LA, Macdonald NH, Apple W, Krauss RM, Galotti KM. 1983. Acoustic and perceptual
indicators of emotional stress. J. Acoust. Soc. Am. 73:1354-60
Tassinary LG, Cacioppo JT. 1992. Unobservable facial actions and emotion. Psychol. Sci. 3:28-
33
Tiedens L. 2001. Anger and advancement versus sadness and subjugation: The effect of
negative emotion expressions on social status conferral. J. Pers. Soc. Psychol. 80:86-
94
Titze IR. 1994. Principles of Voice Production. Englewood Cliffs, NJ: Prentice-Hall
Tolkmitt FJ, Scherer KR. 1986. Effects of experimentally induced stress on vocal parameters. J.
Exp. Psychol: Hum. Percep. Perf. 12:302-13
Tomkins SS. 1962. Affect, Imagery, Consciousness, Vol. 1. New York: Springer.
Trouvain J, Barry WJ. 2000. The prosody of excitement in horse race commentaries. See Cowie
et al 2000, pp. 86-91
Turner TJ, Ortony A. 1992. Basic emotions: Can conflicting criteria converge? Psycholo. Rev
Russell, Bachorowski, & Fernandez Dols 38
99:566-571
Wallbott HG, Scherer KR. 1986. Cues and channels in emotion recognition. J. Pers. Soc.
Psychol. 51:690-99
Widen SC, Russell JA. 2002. Gender and preschoolers perception of emotion. Merrill-Palmer
Quart. In press
Wild B, Erb M, Bartels M. 2001. Are emotions contagious? Evoked emotions while viewing
emotionally expressive faces: Quality, quantity, time course and gender differences.
Psychiatry Res. 102:109-24
Winkielman P, Cacioppo JT. 2001. Mind at ease puts a smile on the face: Psychophysiological
evidence that processing facilitation elicits positive affect. J. Pers. Soc. Psychol. 81:989-1000
Yik MSM, Meng Z, Russell JA. 1998. Adults’ freely produced emotion labels for babies’
spontaneous facial expressions. Cogn. Emot. 12:723-30
Yik MSM, Russell JA. Interpretation of faces: A cross-cultural study of a prediction from
Fridlund’s theory. Cogn. Emot. 13:93-104
Zivin G. 1977. Facial gestures predict preschoolers’ encounter outcomes. Soc. Sci. Inf. 16:715-30
