Content uploaded by James Douglas Laird
Author content
All content in this area was uploaded by James Douglas Laird on Dec 26, 2013
Content may be subject to copyright.
joumii
of
ftnoBihty
sod Socnl Psychology
l»4,V«.47.No.4,«»-9nCooyndu 1984 bribe
Amcncu PiycbologKil Aaooalioa. Inc.
The Real Role
of
Facial Response
in the
Experience
of
Emotion:
A Reply
to
Tourangeau
and
Ellsworth,
and
Others
James D. Laird
Clark University
The facial feedback hypothesis holds that emotional experiences are derived from
facial expressions.
Ten
published studies indicating that manipulated facial
expressions
do
produce corresponding emotional experience are contrasted with
Tourangeau
and
Ellsworth's sole published failure
to
demonstrate this relation.
Six other studies using
a
different
but
theoretically consistent paradigm also
demonstrate facial feedback effects. Related results
in
many
of
these
16
studies
effectively rule
out
experimental demand
as an
explanation
and
instead suggest
similarities
in
process between facial feedback and hunger, attitude change
and
self-evaluation.
Recently Tourangeau and Ellsworth (1979)
failed
in an
attempt
to
show that subjects
induced
to
adopt facial expressions
of
emo-
tion,
would then report feeling those
emo-
tions.
The
authors believed that they were
testing
a
hypothesis related
to,
though
not
directly contained
in the
emotion theories
of
Tomkins (1962)
and
Izard (1977), namely,
that
the
experience
of
emotion
is
derived
from feedback from facial expressions. Tour-
angeau and Ellsworth (1979) concluded from
their study that the "facial feedback" hypoth-
esis
was not
tenable
and
that this cast some
doubt
on the
larger theories
as
well. Subse-
quently, Tomkins (1981)
and
Izard (1981)
both wrote replies
to
that article,
in
large
measure disowning this empirical form
of
the facial feedback hypothesis
on
grounds
that feeling required more than such simple
facial movements. Hager and Ekman (1981),
in
the
same issue, argue
on
methodological
grounds that
the
experimental hypothesis
would
not be
expected
to be
confirmed.
Ellsworth and Tourangeau (1981) then found
themselves defending their attack
on
"this
authorless hypothesis"
(p. 363) and
their
"failure
to
disconfirm what nobody ever said"
(p.
363).
The reader of the Tourangeau and Ellsworth
article
and the
subsequent series
of
replies
Requests
for
reprints should
be
sent
to
James
D.
Laird, Department
of
Psychology, Clark University,
Worcester, Massachusetts 01610.
might easily infer that the weight of the
evidence is against the facial feedback hy-
pothesis and that no one expected otherwise.
Both inferences are false.
Evidence for Facial Feedback
The studies relevant to the facial feedback
hypothesis have all used one of two general
methods. In one type, of which Tourangeau
and Ellsworth's study is a good example,
facial expressions are manipulated more or
less muscle by muscle. For example, a subject
might be asked to contract muscles between
the eyebrows and at the corners of the jaw,
to produce an angry expression. Tourangeau
and Ellsworth's is the only published study
which failed to show an effect of the muscle
by muscle procedure on emotional experi-
ence.
Izard (1981) also mentions an unsuc-
cessful unpublished study by Kotsch, Izard,
and Walker. Ranged against these two are
quite a number of successful published dem-
onstrations of the effect in our lab (Duncan
& Laird, 1977, 1980; Edelman, 1984; Kell-
erman & Laird, 1982; Laird, 1974; Laird &
Crosby, 1974; Laird, Wagener, Halal, &
Szegda, 1982) and by others (Kleinke &
Walton, 1982; MacArthur, Solomon, & Jaffee,
1980;
Rhodewalt & Comer, 1979). In this
type of study, the box score favors the facial
feedback hypothesis, 10 to 1 on published
articles, and probably by the same margin
among unpublished articles.
909
r.t
910JAMES D. LAIRD
The second type of study gives subjects an
occasion for an emotional reaction, such as
an electric shock or viewing a film. The
subjects are asked to exaggerate or minimize
their expressive reactions, usually to deceive
a purported observer. These studies charac-
teristically have demonstrated effects of vary-
ing the magnitude of expressive behavior on
both self-reports of emotional experience and
on various measures of physiological arousal
such as heart rate and skin conductance
(Colby et al, 1977; Kleck, et al. 1976; Kopel
& Arkowitz, 1974; Kraut, 1982; Lanzetta,
Cartwright-Smith, & Kleck, 1976; Zucker-
man, Klorman, Larrance, & Spiegel, 1981).
Only one study in the exaggerate/minimize
paradigm reported negative results (McCaul,
Holmes, & Solomon, 1982).
Overall, then, the box scores favor the
facial feedback hypothesis by 10 to 1 in the
muscle by muscle paradigm and 6 to 1 in
the exaggerate/minimize paradigm. Further-
more, successful demonstrations of an effect
must weigh more heavily in such an argument,
because there are only a few ways in which
one can succeed but many in which to fail.
This box score approach does not get us very
far, however. The important question is, what
differentiates the studies that support the
facial feedback hypothesis from those that do
not? To answer this, we need to consider each
experimental paradigm separately.
In the muscle by muscle paradigm the
basic technique for manipulating expression
is always similar. Subjects are given some
plausible excuse, most often electromyo-
graphic recording and then are asked to
contract and relax different muscles that in
fact are chosen to create the expression of an
emotion.
The most obvious difference between
Tourangeau and Ellsworth's procedures and
those of all of the other studies is whether
the comparisons were within or between sub-
jects.
Tourangeau and Ellsworth deliberately
chose to run subjects in only one expression
condition in order to minimize the likelihood
that subjects would become aware of the
experimenter's hypotheses. Obviously if sub-
jects are first asked to adopt a smile and then
a frown, they are more likely to guess the
real intent of the study. In experiments with
a within-subjects design, positive results could
be created by experimenter demand (see also
Buck, 1980, for a discussion of this problem).
In short, a possible explanation of the dis-
crepancy between Tourangeau and Ellsworth
and the other studies is that the positive
results were all due to artifact.
This issue has been a concern of previous
experimenters, and at least three general fea-
tures of these studies were designed to cope
with this possibility. First, all employ elaborate
deceptions to misdirect subjects about the
experimenter's intent. A second response to
this problem (Laird, 1974) was to run a
parallel "observer" subject who sat beside the
manipulated subject and could see and hear
everything the experimenter said and did
during the experiment. These observer sub-
jects showed no variations in emotional ex-
perience between smile and frown conditions,
suggesting that -the effects were not due to
the experimenter's behavior. However, these
observer subjects did not have available to
them the information that the manipulated
subjects might have had from their own faces,
that the facial manipulations produced smile
and frown expressions. Thus, the possibility
that the manipulated subjects were simply
complying with perceived expectations is re-
duced but not eliminated. The third proce-
dure used in this and all subsequent studies
was an extensive postexperimental question-
naire. These questionnaires asked a series of
questions about the procedures, the subjects'
understanding of the purpose of the experi-
ment or any other purpose they could imag-
ine,
what they thought the experimenter ex-
pected them to do, and what factors in the
experiment may have affected their mood.
Any subjects who revealed awareness of the
hypotheses were excluded from subsequent
analyses.
These three procedures for dealing with
the awareness problem have made crude
awareness and experimental demand unlikely,
but do not rule out more subtle effects. Of
course, at this level of subtlety the same kind
of demand effects might also have occurred
in the Tourangeau and Ellsworth study (cf.
Hager & Ekman, 1981). Running subjects in
only one condition reduces, but does not
entirely remove the possibility of awareness
and/or experimenter bias.
Fortunately, a number of the more recent
FACIAL RESPONSE
AND
EMOTION911
studies provide
an
indirect
but
much more
compelling argument against
the
possibility
of subject awareness
and
experimenter
de-
mand. These studies developed from
an em-
pirical observation
and
some theoretical
as-
sumptions which will
be
discussed
in
greater
detail later.
The
empirical observation
was
that although
the
overall effects
of
expressions
were significant,
not all
subjects responded
in
an
equally strong manner (Laird, 1974).
Following
up on
this observation
we ran a
study
in
which subjects were
run
through
the
expression manipulation procedure twice
at
an interval
of a few
days.
We
found that
people consistently
did or did not
respond
to
the expression manipulation
on
both occa-
sions (Laird
&
Crosby, 1974). These differ-
ences
in
expression response were also related
in theoretically consistent ways
to how sub-
jects described their emotional experience
in
everyday life.
In
short,
it
appeared that
the
facial feedback effect
was not
equally strong
in
all
people.
The theory that guided this work was
self-
perception theory (Bern,
1967;
1972).
In
this
view,
we are in the
same position
as any
observer
of
us,
who
must infer
our
psychic
states from observation
of our
actions
and
the circumstances
in
which
we act.
Like
the
observer, we
can
only know something about
ourselves
by
observing what
we do and say.
For example Bern (1967) argued that we infer
our attitudes from
our
speeches
and self-
descriptions of attitudes. From the perspective
of self-perception theory
the
relation between
facial expressions
and
emotional experience
is
a
particular case
of the
general relation
between behaviors
and
psychic states.
Putting together
the
observation that some
people responded
to
their expressions
and
some
did
not, with
the
theoretical assumption
that this was
an
instance of a general process,
we were lead
to
propose that there were
general differences between people
in the
kinds of information they used
in
identifying
their own attributes (Laird
&
Berglas, 1975).
There seem
to
be two such varieties
of
infor-
mation.
One
kind, which
we
labeled
self-
produced cues, arise from
our
actions, such
as expressive behaviors, bodily activities
in
arousal,
and
instrumental action.
The
other
kind
of
cues, situational cues, consist
of
normative information from
the
situation,
about what anyone
in the
situation should
or
probably would feel.
In his work
on
the eating
of
normal weight
and obese people, Schachter (1971) has drawn
a similar distinction, between external
and
internal cues. Among external cues
for
eating
he would include
the
time
of day, the ap-
pearance,
and the
taste
of
food, whereas
internal cues would consist
of
stomach
con-
tractions. Schachter's labels seem quite
straightforward
and
useful
in the
context
of
the eating research,
but
less
so in
regard
to
other self-perceptions.
The
problem
is
that
the external/internal labels imply that
the distinction
is
essentially geographical—
whether
the
cues arise inside
or
outside
of
the individual's skin.
In the
case
of a
smile,
for example,
it is
difficult
to
say whether
it is
inside
or
outside
the
skin,
and the
attempt
to
make such
a
distinction seems even more
inappropriate
in
regard to
a
counterattitudinal
speech.
The
point
is
that
the
distinction
is
not really between inside
and
outside. Instead,
one kind of cue arises from
a
person's partic-
ular activities, whereas
the
other consists
of
general expectations about
how
most people
would behave
in a
situation, which require
no attention
to
the individual. This latter way
of characterizing the distinction seemed better
captured
by the
labels
of
self-produced
or
situational.
The
issue however,
is not the
nature of the distinction,
but
simply how that
distinction is
to
be described. What Schachter
would call external cues,
we
would call situ-
ational,
and his
internal would
be our self-
produced.
If indeed
we
were talking about
the
same
kinds of
cues,
then clearly response
to
expres-
sion manipulations should
be
related
to
body
weight. This is indeed the
case.
In
four studies,
it
has
been found that normal weight people
who presumedly
eat in
response
to
self-pro-
duced cues also feel
the
emotions they
are
induced
to
express.
In
contrast overweight
people who are unresponsive to self-produced
cues
for
eating
are
similarly unresponsive
to
self-produced cues from their expressions
(Comer,
1975;
Comer
&
Rhodewalt,
1979;
Edelman, 1984; MacArthur,
et al.
1980).
A number
of
other studies have examined
the relation between
the
expression manipu-
lation procedure
and
other self-perception
procedures.
For
example, Bern (1967) origi-
912JAMES D. LAIRD
nally proposed self-perception theory as an
explanation of the effect of counter-attitudinal
behavior and our analysis suggested that there
should be individual differences in these ef-
fects (Laird & Berglas, 1975). Such differences
in the self-perception of attitudes should be
related to the effects of expression on feeling,
and they are. People whose emotional expe-
rience was affected by the expression manip-
ulations also changed their attitudes in the
induced-compliance paradigm, whereas sub-
jects who did not respond to the expression
manipulation also did not respond to the
induced-compliance procedure (Duncan &
Laird, 1977; Rhodewalt & Comer, 1979). In
another study a reverse of
this
effect occurred.
Subjects who do not respond to the
self-
produced cues of the expression manipulation
do respond to the situational cues of a con-
formity manipulation (Comer, 1975). Simi-
larly, subjects who respond to a placebo in
the standard, positive way, do not feel the
emotions they are induced to express. Subjects
who are more responsive to self-produced
cues instead show a reverse placebo effect in
which they feel the opposite of the placebo
message (Duncan & Laird, 1980). Another
study demonstrated that mood produced by
expression manipulations affected recall of
emotional material, but only among subjects
more responsive to self-produced cues (Laird,
Wagener, Halal, & Szegda, 1982). Appearance
changes also affected the self-perceptions of
subjects who responded to the expression
manipulation, but did not affect the others
(Kellerman & Laird, 1982). Finally, subjects
who show the facial feedback effect also tend
to be field independent, as measured by the
Rod and Frame Test (Rhodewalt & Comer,
1979;
Edelman, 1984).
All of these relations between facial feed-
back and other tasks or measures cannot be
explained by experimenter bias. In all of
these studies the experimenter was completely
unaware of the subjects' performance on the
other tasks or tests, so he or she could not
have known which subjects to bias in the
appropriate way. Similarly, in most cases the
subjects could not have identified their own
performance on one task or test, so they
could not have known whether or not they
should respond on another. Thus, the exper-
imenter bias/cooperative subjects explanations
will not suffice.
This pattern of relations also renders un-
likely the possibility that subjects who respond
to the expression manipulation procedure are
simply the more suggestible or complaint.
For example, much evidence indicates that
field dependent people are more susceptible
to suggestion (e.g., Witkin, et al. 1954), but
field dependent subjects do not respond to
the expression manipulation procedure. In-
stead it is the less suggestible field independent
subjects who do (Rhodewalt & Comer, 1979;
Edelman, 1984). Similarly, obese individuals
have been found to be more conforming
(Rodin & Slochower, 1974) but they do not
respond to the expression manipulation pro-
cedure, instead their normal weight cohorts
do.
In addition, subjects who respond to the
expression manipulations do not accept the
suggestion of a placebo message (Duncan &
Laird, 1980). Finally, in a recent study sub-
jects who were unaffected by the expression
manipulation accepted a direct suggestion
about how they should judge themselves,
whereas subjects who did respond to the
expression manipulation responded opposite
to the suggestion (Kellerman & Laird, 1982).
In sum, there is a clear, consistent pattern in
which one group of subjects are responsive
to self-produced cues, including those arising
from their manipulated facial expressions.
These subjects do not respond to suggestion,
conformity and other situational definitions
of how they should feel. Another group do
not respond to the expression manipulations,
and these subjects are very responsive to
suggestion.
At this point it seems reasonable to rule
out methodological explanations of the studies
that show that facial feedback affects emo-
tional experience. Instead, it seems necessary
to find a methodological explanation of the
one study which did not find such an effect.
However, before we go on, one qualification
that has slipped rather casually into this
discussion should be highlighted. The evi-
dence seems consistent and strong that the
facial feedback effect occurs. It is equally
clear that it occurs only among some people,
not all. Tourangeau and Ellsworth may have
been half wrong, but they were also half
right.
If the facial feedback effect occurs, and is
not due to experimental demand, then why
did Tourangeau and Ellsworth fail to detect
FACIAL RESPONSE AND EMOTION913
it? A strong possibility is that individual
differences in the effects of expressions on
emotional experience may be the cause. If
Tourangeau and Ellsworth's sample contained
a relatively high proportion of subjects who
were more responsive to situational cues,
then the effects among subjects more respon-
sive to self-produced cues might have been
obscured in the aggregate. Of
course,
there is
no way to know if this did occur, but this
possibility seems more plausible in the light
of a study showing that some kinds of subject-
recruitment procedures do differentially at-
tract subjects responsive to self-produced or
situational cues (Wagener & Laird, 1980).
An equally likely possibility arises from
Tourangeau and Ellsworth's choice of a be-
tween-subjects design. They chose this design
in the reasonable attempt to reduce subject
awareness,
but they also accepted some serious
limitations with its advantages. In particular,
we
have consistently found that subjects arrive
at experimental sessions in widely different
moods and/or use the mood scales in different
ways.
These differences tend to endure
through the session for many subjects, espe-
cially those more responsive to situational
cues,
and expression manipulations often
modify existing moods without entirely over-
riding them. All the differences between how
subjects feel when they arrive and how they
react to the experimental situation will appear
as error variance in a between-subject design.
In addition to the possible effects of subject
differences and the error variance introduced
with the berween-subjects design, a number
of possible problems with Tourangeau and
Ellsworth's method were suggested by Izard,
Tomkins, Hager, and Ekman. However, most
of these features of their method were shared
with some of the successful studies, so it
seems unlikely that they were at fault For
example, Hager and Ekman were concerned
that there was not very tight control of the
facial muscle manipulations, but their pro-
cedure was at least as good as most of the
other studies. Similarly, Hager and Ekman
noted that Tourangeau and Ellsworth asked
their subjects to maintain the expressions for
2 min, whereas in most of the other studies
10 to 15 s was the norm. However, in three
recent studies we have asked subjects to
maintain expressions for longer lengths of
time,
up to about 3 minutes, and have still
found the usual effect (Laird, et al. 1982;
Wagener & Laird, 1980; Wixon & Laird,
1981).
Thus, within this range the amount
of time does not seem to be a critical variable
either.
Another difference between the Tourangeau
and Ellsworth study and many others in
which expressions were directly created by
the experimenter is in the kinds of expressions
used. Most of the published positive studies
have used expressions of happiness and anger,
whereas Tourangeau and Ellsworth used fear
and sadness. However, MacArthur, Solomon
and Jaffee (1980) used a sad expression suc-
cessfully. In addition, in two recent studies
(Laird, et al. 1982; Wagener & Laird, 1980)
we have used fearful and sad expressions
successfully in conceptually related experi-
ments. Thus, the nature of the expressions
used does not seem likely to be the origin of
the different results.
In sum, then, it is apparent that there is a
great deal of evidence from the muscle by
muscle paradigm in support of the facial
feedback hypothesis. This evidence cannot
readily be explained except by the assumption
that, at least for some people, expressions
of emotion lead to feelings of that emo-
tion. Tourangeau and Ellsworth's results are
anomalies in this pattern, and seem attrib-
utable to purely methodological differences.
In addition to these studies, which shape
expressions muscle by muscle, there is the
body of work by Lanzetta, Kleck, and their
colleagues, and Zuckerman et al. (1981).
These studies employed a different method-
ology that is not susceptable to the problems
of
how
well or how long subjects' expressions
are formed. In these studies, subjects are lead
to believe that there are people observing
them as they receive electric shocks or watch
emotionally evocative films, and the subjects'
task is to mislead the observers about the
intensity of the stimulus. Thus, in these stud-
ies
the subjects are minimizing or exaggerating
their own expressive behavior in whatever
ways they find natural. These variations in
expressed feeling are then reflected in subjects'
reports of their real pain, disgust, or amuse-
ment. Tourangeau and Ellsworth were aware
of some of these studies and therefore reason-
ably restricted their negative conclusions:
"Lanzetta, Cartwright-Smith, and Kleck's
(1976)
finding
of an effect of facial expression
914JAMES D. LAIRD
cm feelings of pain does not seem to extend
to feelings of fear or sadness" (Tourangeau &
Ellsworth, p. 528). They do not offer any
explanation for this discrepancy.
In the only failure to find facial feedback
in this paradigm, McCaul et al. (1982) con-
cluded that mimicking emotional expressions
could produce arousal variations without any
effects on experienced affect. They report two
studies, but one of these apparently did not
employ any measure of emotional experience,
and so is irrelevant to this conclusion. Their
other study did have a measure of experience,
but this measure was deployed in a consid-
erably less powerful way than their measures
of physiological response. The physiological
measures were compared between conditions
in which subjects posed feeling "afraid, calm,
or normal," so that all comparisons were
within subjects. The measures of experience
on the other hand were based only on the
first trial for each subject, with the comparison
being between subjects who were assigned
different orders of expression conditions.
Thus,
the negative results appeared in only
one of their two studies, and only in a
between-subjects comparison that was less
powerful than any of the others. In the more
powerful comparisons within subjects, their
results replicate the observation that mimick-
ing facial expressions can produce changes in
heart rate. However, their article provided
only weak evidence that expressions do not
affect experience. In any case, the bulk of the
data from this paradigm still clearly favors
the facial feedback hypothesis.
Conclusions
In both of the two paradigms used to study
facial feedback the situation is the same at
the moment. In both there are a number of
studies demonstrating that independently
manipulated facial expressions do affect emo-
tional experience, whereas the contrary evi-
dence is meager and probably attributable to
differences in method. The facial feedback
effect has been demonstrated with a wide
variety of emotions, including anger, happi-
ness,
sadness, fear, pain, and humor, with
various cover stories and procedures for pro-
ducing expressions, and with various dura-
tions of expression. In sum, contrary to Tour-
angeau and Ellsworth and others, the facial
feedback effect has been demonstrated fre-
quently and consistently.
The evidence is equally strong that this
effect is not due to experimental demand or
the compliance of subjects. Those subjects
who are most affected by manipulations of
their facial expressions are those who seem
least susceptible to demand and suggestion,
because they are field independent, normal
weight, do not accept placebo instructions,
nor do they respond to direct experimenter
suggestions about how they should feel. Ex-
perimenters could not be producing differ-
ential effects because in all of these studies
they are blind to the subjects' performances
on both the expression manipulations and
other tasks. In short, at least some people are
happy because they smile, are angry because
they scowl and are sad because they pout.
Just as some people are affected by facial
feedback, others are not. Unaffected subjects
appear, instead, to define their emotional
states in terms of situational expectations.
This interpretation is less certain than the
first two however, because much of the evi-
dence on the role of situational cues requires
an additional step of inference. Most of the
studies that directly demonstrate effects of
situational cues have involved phenomena
other than emotion. For example, we know
with considerable confidence that people who
don't respond to facial feedback tend to eat
in response to external, or situational cues.
They also accept direct suggestions about
their self-concepts and abilities. These studies
show a greater response to situational cues,
but not for emotion. Only one study (Duncan
& Laird, 1980) has actually demonstrated
directly that people who don't show facial
feedback do accept the emotional state im-
plied by their circumstances. Thus, the evi-
dence is less extensive and more inferential,
but the most reasonable conclusion is that
people who do not show the facial feedback
effect instead base their emotional experience
on situational cues.
It is at least possible that the facial feedback
effect is real but is no more than a curiosity
of the experimental lab, like some perceptual
illusion that can only occur in the contrived
circumstances of the laboratory. Naturally it
is difficult for an experiment to bear directly
on the issues of external validity. However, a
FACIAL RESPONSE AND EMOTION915
Dumber of the relations discussed in this
article do connect facial feedback to the real
world. All of these connections relate to the
individual differences in response to
self-
produced and situational cues. First of all,
people who respond to self-produced cues
describe their everyday emotional experiences
differently from those who do not (Laird &
Crosby, 1974). In addition, all of the evidence
connecting cue response to body weight takes
us outside of the laboratory, to the places
and times when people eat. The studies con-
necting cue response to field dependence
themselves remain within the laboratory, but
because
field
dependence is related to a variety
of real-life events, these studies add at least a
small increment of confidence to this con-
nection. Overall however, the evidence for
this connection to the real world is not as
strong as for the earlier propositions. None-
theless, it seems reasonable at this point to
assume that we are indeed studying the pro-
cesses by which people feel emotions in their
everyday lives.
How large is the role of facial feedback in
emotional experience? The studies reviewed
here do not provide a very clear answer.
Within the confines of the experiments them-
selves, the effects of expressions on mood can
be
estimated by omega square (Hays
&
Wink-
ler, 1971, p. 728), and range from .12 to .17
(Laird, 1974) to .44 (Duncan & Laird, 1980).
This last value is quite impressive, because it
indicates that 44% of the variance has been
accounted for by the expression manipula-
tions.
However, this is variance in experimen-
tal contexts carefully designed to minimize
any other sources of variance. Thus, these
values derived from controlled experiments
tell us nothing about the relative importance
of facial expressions in everyday life.
In fact, self-perception theory suggests that
although facial expressions might be among
the most important determinants of emo-
tional experience, there are certainly others.
For example, the importance of situational
cues has already been discussed. In addition,
any other behaviors that permit an observer
to infer an individual's mood might also
serve the individual as well. Thus, it seems
likely that patterns of movement, posture,
expressive language and overt actions would
also contribute to the individual's experience.
Nonetheless, considering the importance of
facial expressions in judgments of other peo-
ple's moods, it certainly seems likely that
facial expression is also one of the most
important determinants of
one's
own feelings.
There is one body of research that suggests
how important expressive behavior may be.
This is the research demonstrating the effec-
tiveness of psychotherapeutic techniques such
as systematic desensitization (e.g., Kazdin &
Wilson, 1978). These therapies seek to inhibit
the experience of negative emotions by con-
trolling their
expression.
Specifically, the ther-
apist teaches the client to relax all of their
muscles and to keep them relaxed during the
remainder of the therapeutic procedures. The
relaxed muscles then prevent the individual
from feeling anxiety, guilt, and so forth. No
doubt a number of mechanisms could be
proposed for this effect, but surely the
most straightforward mechanism is discussed
here—that in order to feel an emotion at
least some people must first enact its expres-
sion.
Throughout this article I have used the
language of self-perception to discuss the
facial feedback effect. However, the facial
feedback hypothesis more
frequently
has been
derived from the emotion theories of Izard
and Tomkins (as for example by Tourangeau
& Ellsworth) and clearly the results might
have been described in those terms as well.
In fact, both differential emotions theory
(Izard,
1977;
Tomkins, 1962) and self-percep-
tion theory can account equally well for the
basic facial feedback effect. Although both
Izard and Tomkins were dubious about the
kinds of manipulations that produce facial
feedback, their doubts were essentially meth-
odological, and their theories are consistent
with these results. Indeed, these results sup-
port an important part of their theories that
had not previously received direct empirical
confirmation.
Izard and Tomkins do not explicitly discuss
individual differences of the kind described
here,
but their theories contain elements
which could account effortlessly for individual
differences too. For example, both have dis-
cussed the obvious fact that facial expressions
can occur without appropriate feeling, and
the explanations they propose might be ex-
panded to represent stable differences between
916JAMES
D.
LAIRD
people,
as
well
as
between occasions.
I
noted
earlier that
the
evidence
is
persuasive, though
not quite
as
strong, that subjects
who do not
show the facial feedback effect instead respond
to situational cues. This finding
is
less
ob-
viously compatible with
the
theories
of
Izard
and Tomkins,
but I am
confident
it can be
incorporated as well.
The
final
group of results
that must
be
accommodated
by
differential
emotions theory demonstrates connections
between
how
people feel emotions
and how
they feel hunger, attitudes, self-evaluations,
and abilities. This
is the
point
at
which
self-
perception theory
and the
theories
of
Izard
and Tomkins most differ.
I
don't think,
how-
ever, that this
is a
competitive difference,
where
one
must
be
right and the other wrong.
Instead,
the two
theoretical positions seem
to
deal primarily with quite different phenom-
ena,
and
overlap only
in
regard
to the
origins
of the subjective experience of emotion, where
they
are in
essential agreement.
For example, relatively little
is
known about
the details
of
the facial feedback process,
but
most
of
what
is
known
or
assumed
is the
same
for
both self-perception theory
and
differential emotions theory. Both assume
that
the
process partly depends
on
sensory
information from
the
facial muscles
or
skin.
Izard
and
Tomkins have been more elaborate
in their speculations about
the
neural path-
ways
and the
central nervous system struc-
tures involved,
but
none of these speculations
is incompatible with
a
self-perception
per-
spective. Both perspectives also assume that
the information from facial feedback must
be integrated
and
combined,
so
that
the
actual emotional experience represents
a pat-
tern of facial muscle activity
and
other inputs,
including such cognitive factors
as the
occa-
sion
for the
expressive activity.
(The
recog-
nition of this integrative aspect of the process
is
the
reason
for the
"perception" part
of
the
self-perception label.) Both perspectives also
assume that this integrative process
is
auto-
matic, very rapid, and that people are unaware
of either
the
process
or the
constituent
ele-
ments
of the
final integrated experience.
1
have
in the
past (Laird,
1974;
Kellerman
&
Laird, 1982) suggested that this process seems
to
be
very similar
to the
process
of
depth
perception
in its
automatic, rapid, integrative,
and inaccessible qualities,
and
this parallel
seems consistent with Izard's
and
Tomkin's
views
as
well.
The two theories differ
in
how they develop
from this
set of
shared assumptions. Izard,
for example, explicitly notes that
the
facial
feedback issue concerns
how
emotions
are
initially activated,
and
distinguishes
between
(the
facial feedback) activation
of
emotion
and
other aspects of emotional processes. Propositions relating
to
the
latter
and to
emotion-consciousness-cognition-
action relationships
are the
real heart
of
differential
emotions theory.
The
latter propositions
are the
essential
proving grounds
for
differential emotions theory
. . "
(Izard,
1981, p. 351).
The issue
of
facial feedback
and
subjective
experience
is not the
central concern
of his
theory.
He is
primarily interested
in the
effects
of emotional processes
on
later thoughts
and
actions.
On the
other hand, subjective expe-
rience is
the
central concern
of
self-perception
theory, which
in
fact deals with nothing else
but experience. However, self-perception
the-
ory
is
concerned about emotional experience
as
one
among
a
variety
of
kinds
of self-
experience
and
self-knowledge.
In sum,
then,
differential emotions theory
and
self-percep-
tion theory essentially agree about
the
basic
nature
of
facial feedback,
and
only differ
in
the kinds
of
phenomena they attempt
to
relate
to
facial feedback.
Whatever minor theoretical differences may
remain
or
emerge
in
future research,
self-
perception
and
differential emotions theory
agree
on the
major point
of
this review:
Contrary
to
Tourangeau
and
Ellsworth, facial
feedback does occur,
and in
fact
is a
major
component
of
normal emotional processes.
References
Bern,
D. J.
(1967). Self-perception:
An
alternative inter-
pretation
of
cognitive dissonance phenomena. Psycho-
logical
Review.
74,
183-200.
Bern,
D. J.
(1972). Self-perception theory.
In L.
Bcrkowitz
(Ed.)
Advances
in
experimental social psychology (Vol.
6,
pp.). New
York: Academic Press.
Buck,
R.
(1980). Nonverbal behavior
and the
theory
of
emotion:
The
facial feedback hypothesis. Journal
of
Personality
and
Social
Psychology,
38,
811-824.
Colby,
C. Z.,
Lanzetta,
J. T., &
Kleck,
R. E.
(1977)
Effects
of
expressions
of
pain
on
autonomic
and
pain
tolerance responses
to
subject-controlled pain.
Psycho-
physiology,
14,
537-540.
Comer,
R.
(1975). Individual differences
in
self-attribution
behaviors: Dimensions
and
child-rearing correlates.
Ph.D.
Dissertation, Clark University.
FACIAL RESPONSE AND EMOTION917
Comer, R., & Rhodewait, F. (1979). Cue utilization in
the self-attribution of emotions and attitudes. Person-
ality & Social Psychology Bulletin. 5, 320-324.
Duncan, J. W., & Laird, J. D. (1977). Cross-modality
consistencies in individual differences in self-attribution.
Journal
of Personality. 45, 191-206.
Duncan, J. W., & Laird, J. D. (1980). Positive and
reverse placebo effects as a function of differences in
cues used in self-perception. Journal of Personality
and Social
Psychology.
39, 1024-1036.
Edelman, B. (1984). A multiple-factor of body weight
control. Journal of General
Psychology,
110, 99-114.
Ellsworth, P. C, & Tourangeau, R. (1981). On our failure
to disconfirm what nobody ever said. Journal of Per-
sonality and Social
Psychology,
40, 363-369.
Hager, J. C, & Ekman, P. (1981). Methodological prob-
lems in Tourangeau and Ellsworth's study of facial
expression and experience of emotion. Journal of
Personality and Social
Psychology.
40, 358-362.
Hays,
W. L., & Winkler, R. L. (1971). Statistics. Proba-
bility,
inference
and
decision.
New York; Holt, Rinehart
& Winston.
Izard, C. E. (1977). Human emotions. New York &
London: Plenum Press.
Izard, C. E. (1981). Differential emotions theory and the
facial feedback hypothesis of emotion activation: Com-
ments on Tourangeau and Ellsworth's "The cole of
facial response in the experience of
emotion."
Journal
of Personality and Social
Psychology,
40, 2, 350-354.
Kazdm, A. E., & Wilson, G. T. (1978). Evaluation of
behavior therapy: Issues, evidence and research strate-
gies Cambridge, MA: Ballinger.
Kellerman, J., & Laird, J. D. (1982). The effect of
appearance on self-perception. Journal of Personality,
50, 296-315.
Kleck, R. E., Vaughan, R. C, Cartwright-Smith, J.,
Vaughan, K_ B., Colby, C. Z., & Lanzetta, J. T. (1976).
Effects of being observed on expressive, subjective, and
physiological responses to painful stimuli. Journal of
Personality and Social
Psychology,
34, 1211-1218.
Kleinke, C. L., & Walton, J. H. (1982). Influence of
reinforced smiling on affective responses in an interview.
Journal of Personality and Social Psychology, 42, 3,
557-565.
Kopel, S., & Arkowitz, H. S. (1974). Role playing as a
source of self-observation and behavior change. Journal
of Personality and Social
Psychology.
29, 677-686.
Kraut, R. E. (1982). Social pressure, facial feedback, and
emotion. Journal of Personality and Social
Psychology,
42.
853-863.
Laird, J. D. (1974). Self-attribution of emotion: The
effects of expressive behavior on the quality of emotional
experience. Journal of Personality and Social Psychol-
ogy,
475-486.
Laird, J. D., & Berglas, S. (1975). Individual differences
in the effects of engaging in counter-attitudinal behavior.
Journal of Personality, 43, 286-304.
Laird, J. D., & Crosby, M. (1974). Individual differences
in the self-attribution of emotion. In H. London & R.
Nisbett (Eds.) Thinking and feeling. The cognitive
alteration of feeling states (pp.). Chicago: Aldine.
Laird, J. D., Wagenet, J. J., Halal, M., & Szegda, M.
(1982).
Remembering what you feel: The effects of
emotion on memory. Journal of Personality and Social
Psychology, 42. 4, 646-657.
Lanzetta, J. T, Cartwright-Smith, J., & Kleck, R. E.
(1976).
Effects of nonverbal dissimilation on emotional
experience and autonomic arousal. Journal of Person-
ality and Social
Psychology.
33. 354-370.
McArthur, L. A., Solomon, M. R., & Jaffee, R. H.
(1980).
Weight and sex differences in emotional re-
sponsiveness to proprioceptive and pictoral stimuli.
Journal of Personality and Social
Psychology,
39. 308-
319.
McCaul, K. D., Holmes, D. S., & Solomon, S. (1982).
Voluntary expressive changes and emotion. Journal of
Personality and Social
Psychology,
42, 145-152.
Rhodewait, F. & Comet; R. (1979). Induced-compliance
attitude change: Once more with feeling. Journal of
Experimental Social
Psychology,
15. 35-47.
Rodin, J., & Slochower, J. (1974). Fat chance for a favor
Obese-normal differences in compliance and incidental
learning. Journal of Personality and Social
Psychology,
29. 557-565.
Schachter, S. (1971). Some extraordinary facts about
obese humans and rats. American Psychologist, 26,
129-144.
Tomkins, S. S. (1962). Affect, imagery, consciousness.
New York: Springer.
Tomkins, S. S. (1981). The role of facial response in the
experience of emotion: A reply to Tourangeau and
Ellsworth. Journal of Personality and Social
Psychology,
40. 355-357.
Tourangeau, R., & Ellsworth, P. C. (1979). The role of
facial response in the experience of emotion. Journal
of Personality and Social
Psychology.
37, 1519-1531.
Wagener, J. J., & Laird, J. D. (1980). A "new look" at
judgments about emotion in others: The effects of the
emotional experience of the judge. Paper presented at
EPA. Eastern Psychological Association.
Wagener, J. J. & Laird, J. D. (1980). The experimenters
foot-in-the-door Self-perception, body weight and vol-
unteering. Personality and Social Psychology Bulletin,
6. 441-446.
Witkin, H. A., Lewis, H. D., Herteman, M., Machover,
K., Meissna; P. B., & Wapner, S. (1954). Personality
through perception New York: Harper.
Wixon, D. R., & Laird, J. D. (1981). Effects of mimicry
on the judgment of facial expressions in others. Paper
presented at the Annual Meeting of the Eastern Psy-
chological Association.
Zuckerman, M., Kiorman, R., Larrance, D. T. & Spiegel,
N.
H. (1981). Facial, autonomic, and subjective com-
ponents of emotion: The facial feedback hypothesis
versus the extemalizer-intemalizer distinction. Journal
of Personality and Social
Psychology,
41. 929-944.
Received February 22, 1983
Revision received June 10, 1983
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Personality and Social Psychology
This content is subject to copyright. Terms and conditions apply.
