VOL. 69, No.

5 SEPTEMBER 1962

PSYCHOLOGICAL REVIEW
COGNITIVE, SOCIAL, AND PHYSIOLOGICAL
DETERMINANTS OF EMOTIONAL STATE1
STANLEY SCHACHTER A FD JEROME E. SINGER
Columbia University Pennsylvania State University

The problem of which cues, internal mentally manipulated were character-

or external, permit a person to label ized by a general pattern of excitation
and identify his own emotional state of the sympathetic nervous system but
has been with us since the days that there appeared to be no clear-cut physi-
James (1890) first tendered his doc- ological discriminators of the various
trine that "the bodily changes follow emotions. This pattern of results was
directly the perception of the exciting so consistent from experiment to ex-
fact, and that our feeling of the same periment that Cannon (1929) offered,
changes as they occur is the emotion" as one of the crucial criticisms of the
(p. 449). Since we are aware of a James-Lange theory, the fact that "the
variety of feeling and emotion states, it same visceral changes occur in very dif-
should follow from James' proposition ferent emotional states and in non-emo-
that the various emotions will be ac- tional states" (p. 351).
companied by a variety of differentiable More recent work, however, has
bodily states. Following James' pro- given some indication that there may
nouncement, a formidable number of be differentiators. Ax (1953) and
studies were undertaken in search of Schachter (1957) studied fear and
the physiological differentiators of the anger. On a large number of indices
emotions. The results, in these early both of these states were characterized
days, were almost uniformly negative. by a similarly high level of autonomic
All of the emotional states experi- activation but on several indices they
1
did differ in the degree of activation.
This experiment is part of a program of Wolf and Wolff (1947) studied a sub-
research on cognitive and physiological de-
terminants of emotional state which is being ject with a gastric fistula and were able
conducted at the Department of Social Psy- to distinguish two patterns in the phys-
chology at Columbia University under PHS iological responses of the stomach wall.
Research Grant M-2S84 from the National It should be noted, though, that for
Institute of Mental Health, United States
Public Health Service. This experiment was many months they studied their subject
conducted at the Laboratory for Research in during and following a great variety of
Social Relations at the University of Minne- moods and emotions and were able to
sota. distinguish only two patterns.
The authors wish to thank Jean Carlin and Whether or not there are physiologi-
Ruth Hase, the physicians in the study, and
Bibb Latan6 and Leonard Weller who were cal distinctions among the various emo-
the paid participants. tional states must be considered an open
379
380 STANLEY SCHACHTER AND JEROME E. SINGER

question. Recent work might be taken factors are completely interrelated.

to indicate that such differences are at Imagine a man walking alone down a
best rather subtle and that the variety dark alley, a figure with a gun sud-
of emotion, mood, and feeling states are denly appears. The perception-cogni-
by no means matched by an equal vari- tion "figure with a gun" in some fash-
ety of visceral patterns. ion initiates a state of physiological
This rather ambiguous situation has arousal; this state of arousal is inter-
led Ruckmick (1936), Hunt, Cole, and preted in terms of knowledge about
Reis (1958), Schachter (1959) and dark alleys and guns and the state of
others to suggest that cognitive factors arousal is labeled "fear." Similarly a
may be major determinants of emo- student who unexpectedly learns that
tional states. Granted a general pat- he has made Phi Beta Kappa may ex-
tern of sympathetic excitation as char- perience a state of arousal which he will
acteristic of emotional states, granted label "joy."
that there may be some differences in Let us now consider circumstances in
pattern from state to state, it is sug- which these two elements, the physio-
gested that one labels, interprets, and logical and the cognitive, are, to some
identifies this stirred-up state in terms extent, independent. First, is the state
of the characteristics of the precipi- of physiological arousal alone sufficient
tating situation and one's apperceptive to induce an emotion? Best evidence
mass. This suggests, then, that an indicates that it is not. Maranon s
emotional state may be considered a (1924), in a fascinating study, (which
function of a state of physiological was replicated by Cantril & Hunt,
arousal 2 and of a cognition appropriate 1932, and Landis & Hunt, 1932) in-
to this state of arousal. The cognition, jected 210 of his patients with the sym-
in a sense, exerts a steering function. pathomimetic agent adrenalin and then
Cognitions arising from the immediate simply asked them to introspect. Sev-
situation as interpreted by past experi- enty-one percent of his subjects simply
ence provide the framework within reported their physical symptoms with
which one understands and labels his no emotional overtones; 29% of the
feelings. It is the cognition which de- subjects responded in an apparently
termines whether the state of physio- emotional fashion. Of these the great
logical arousal will be labeled as majority described their feelings in a
"anger," "joy," "fear," or whatever. fashion that Maranon labeled "cold" or
In order to examine the implications "as if" emotions, that is, they made
of this formulation let us consider the statements such as "I feel as if I were
fashion in which these two elements, afraid" or "as if I were awaiting a
a state of physiological arousal and cog- great happiness." This is a sort of
nitive factors, would interact in a vari- emotional "deja vu" experience; these
ety of situations. In most emotion in- subjects are neither happy nor afraid,
ducing situations, of course, the two they feel "as if" they were. Finally a
2
Though our experiments are concerned very few cases apparently reported a
exclusively with the physiological changes genuine emotional experience. How-
produced by the injection of adrenalin, which ever, in order to produce this reaction
appear to be primarily the result of sym-
pathetic excitation, the term physiological in most of these few cases, Maranon
arousal is used in preference to the more (1924) points out:
specific "excitation of the sympathetic nerv- 3
ous system" because there are indications, Translated copies of Marafion's (1924)
to be discussed later, that this formulation paper may be obtained by writing to the
is applicable to a variety of bodily states. senior author.
 DETERMINANTS OF EMOTIONAL STATE 381
One must suggest a memory with strong precisely such a state would lead to the
affective force but not so strong as to pro- arousal of "evaluative needs" (Festin-
duce an emotion in the normal state. For ex-
ample, in several cases we spoke to our pa- ger, 1954), that is, pressures would act
tients before the injection of their sick on an individual in such a state to un-
children or dead parents and they responded derstand and label his bodily feelings.
calmly to this topic. The same topic pre- His bodily state grossly resembles the
sented later, during the adrenal commotion,
was sufficient to trigger emotion. This ad-
condition in which it has been at times
renal commotion places the subject in a situ- of emotional excitement. How would
ation of 'affective imminence" (pp. 307-308). he label his present feelings ? It is sug-
gested, of course, that he will label his
Apparently, then, to produce a genu- feelings in terms of his knowledge of
inely emotional reaction to adrenalin, the immediate situation.4 Should he at
Maranon was forced to provide such the time be with a beautiful woman he
subjects with an appropriate cognition. might decide that he was wildly in love
Though Maranon (1924) is not ex- or sexually excited. Should he be at
plicit on his procedure, it is clear that a gay party, he might, by comparing
his subjects knew that they were re- himself to others, decide that he was
ceiving an injection and in all likeli- extremely happy and euphoric. Should
hood knew that they were receiving he be arguing with his wife, he might
adrenalin and probably had some order explode in fury and hatred. Or, should
of familiarity with its effects. In short, the situation be completely inappropri-
though they underwent the pattern of ate he could decide that he was excited
sympathetic discharge common to about something that had recently hap-
strong emotional states, at the same pened to him or, simply, that he was
time they had a completely appropriate sick. In any case, it is our basic as-
cognition or explanation as to why sumption that emotional states are a
they felt this way. This, we would function of the interaction of such cog-
suggest, is the reason so few of Mara- nitive factors with a state of physiologi-
fion's subjects reported any emotional cal arousal.
experience. This line of thought, then, leads to
Consider now a person in a state of the following propositions:
physiological arousal for which no im- 1. Given a state of physiological
mediately explanatory or appropriate arousal for which an individual has no
cognitions are available. Such a state immediate explanation, he will "label"
could result were one covertly to inject this state and describe his feelings in
a subject with adrenalin or, unknown terms of the cognitions available to him.
to him, feed the subject a sympathomi- To the extent that cognitive factors are
metic drug such as ephedrine. Under potent determiners of emotional states,
such conditions a subject would be it could be anticipated that precisely the
aware of palpitations, tremor, face same state of physiological arousal
flushing, and most of the battery of could be labeled "joy" or "fury" or
symptoms associated with a discharge "jealousy" or any of a great diversity
of the sympathetic nervous system. In
4
contrast to Maranon's (1924) subjects This suggestion is not new for several
psychologists have suggested that situational
he would, at the same time, be utterly factors should be considered the chief differ-
unaware of why he felt this way. What entiators of the emotions. Hunt, Cole, and
would be the consequence of such a Reis (1958) probably make this point most
explicitly in their study distinguishing among
state ? fear, anger, and sorrow in terms of situ-
Schachter (1959) has suggested that ational characteristics.
382 STANLEY SCHACHTER AND JEROME E. SINGER

of emotional labels depending on the in the framework of a study of the effects of

cognitive aspects of the situation. vitamin supplements on vision. As soon as a
subject arrived, he was taken to a private
2. Given a state of physiological room and told by the experimenter:
arousal for which an individual has a
completely appropriate explanation In this experiment we would like to
(e.g., "I feel this way because I have make various tests of your vision. We are
particularly interested in how certain vita-
just received an injection of adren- min compounds and vitamin supplements
alin") no evaluative needs will arise affect the visual skills. In particular, we
and the individual is unlikely to label want to find out how the vitamin compound
his feelings in terms of the alternative called 'Suproxin' affects your vision.
What we would like to do, then, if we
cognitions available. can get your permission, is to give you a
Finally, consider a condition in small injection of Suproxin. The injec-
which emotion inducing cognitions are tion itself is mild and harmless; however,
present but there is no state of physio- since some people do object to being in-
logical arousal. For example, an indi- jected we don't want to talk you into any-
thing. Would you mind receiving a
vidual might be completely aware that Suproxin injection?
he is in great danger but for some
reason (drug or surgical) remain in a If the subject agrees to the injection (and
all but 1 of 185 subjects did) the experi-
state of physiological quiescence. Does menter continues with instructions we shall
he experience the emotion "fear"? describe shortly, then leaves the room. In a
Our formulation of emotion as a joint few minutes a physician enters the room,
function of a state of physiological briefly repeats the experimenter's instruc-
arousal and an appropriate cognition, tions, takes the subject's pulse and then in-
jects him with Suproxin.
would, of course, suggest that he does Depending upon condition, the subject re-
not, which leads to our final proposi- ceives one of two forms of Suproxin—•
tion. epinephrine or a placebo.
3. Given the same cognitive circum- Epinephrine or adrenalin is a sympathomi-
stances, the individual will react emo- metic drug whose effects, with minor excep-
tions, are almost a perfect mimicry of a
tionally or describe his feelings as emo- discharge of the sympathetic nervous system.
tions only to the extent that he Shortly after injection systolic blood pressure
experiences a state of physiological increases markedly, heart rate increases
arousal.5 somewhat, cutaneous blood flow decreases,
while muscle and cerebral blood flow in-
PROCEDURE crease, blood sugar and lactic acid concentra-
The experimental test of these propositions tion increase, and respiration rate increases
requires (a) the experimental manipulation slightly. As far as the subject is concerned
of a state of physiological arousal, (6) the the major subjective symptoms are palpita-
manipulation of the extent to which the sub- tion, tremor, and sometimes a feeling of
ject has an appropriate or proper explanation flushing and accelerated breathing. With a
of his bodily state, and (c) the creation of subcutaneous injection (in the dosage ad-
situations from which explanatory cognitions ministered to our subjects), such effects usu-
may be derived. ally begin within 3-5 minutes of injection and
In order to satisfy the first two experi- last anywhere from 10 minutes to an hour.
mental requirements, the experiment was cast For most subjects these effects are dissipated
within 15-20 minutes after injection.
5
In his critique of the James-Lange theory Subjects receiving epinephrine received a
of emotion, Cannon (1929) also makes the subcutaneous injection of % cubic centimeter
point that sympathectomized animals and pa- of a 1 : 1000 solution of Winthrop Labora-
tients do seem to manifest emotional be- tory's Suprarenin, a saline solution of
havior. This criticism is, of course, as ap- epinephrine bitartrate.
plicable to the above proposition as it was to Subjects in the placebo condition receivpd a
the James-Lange formulation. We shall subcutaneous injection of % cubic centimeter
discuss the issues involved in later papers. of saline solution. This is, of course, com-
 DETERMINANTS OF EMOTIONAL STATE 383

pletely neutral material with no side effects These side effects are transitory, that is,
at all. they will only last for about 15 or 20
minutes. What will probably happen is
Manipulating an Appropriate that your feet will feel numb, you will have
an itching sensation over parts of your
Explanation body, and you may get a slight headache.
By "appropriate" we refer to the extent to Again these are side effects lasting IS or
which the subject has an authoritative, un- 20 minutes.
equivocal explanation of his bodily condi-
And again, the physician repeated these
tion. Thus, a subject who had been informed symptoms while injecting the subject.
by the physician that as a direct consequence None of these symptoms, of course, are
of the injection he would feel palpitations, consequences of an injection of epinephrine
tremor, etc. would be considered to have a and, in effect, these instructions provide the
completely appropriate explanation. A sub- subject with a completely inappropriate ex-
ject who had been informed only that the planation of his bodily feelings. This condi-
injection would have no side effects would tion was introduced as a control condition of
have no appropriate explanation of his state.
This dimension of appropriateness was ma- sorts. It seemed possible that the descrip-
tion of side effects in the Epi Inf condition
nipulated in three experimental conditions might turn the subject introspective, self-ex-
which shall be called: Epinephrine Informed amining, possibly slightly troubled. Differ-
(Epi Inf), Epinephrine Ignorant (Epi Ign), ences on the dependent variable between the
and Epinephrine Misinformed (Epi Mis).
Epi Inf and Epi Ign conditions might, then,
Immediately after the subject had agreed
be due to such factors rather than to dif-
to the injection and before the physician en-
ferences in appropriateness. The false symp-
tered the room, the experimenter's spiel in toms in the Epi Mis condition should simi-
each of these conditions went as follows:
larly turn the subject introspective, etc., but
Epinephrine Informed. I should also tell the instructions in this condition do not pro-
you that some of our subjects have ex- vide an appropriate explanation of the sub-
perienced side effects from the Suproxin. ject's state.
These side effects are transitory, that is, Subjects in all of the above conditions were
they will only last for about 15 or 20 min- injected with epinephrine. Finally, there was
utes. What will probably happen is that a placebo condition in which subjects, who
your hand will start to shake, your heart w^re injected with saline solution, were given
will start to pound, and your face may get precisely the same treatment as subjects in
warm and flushed. Again these are side the Epi Ign condition.
effects lasting about IS or 20 minutes.
Producing an Emotion Inducing
While the physician was giving the injec- tion
tion, she told the subject that the injection
was mild and harmless and repeated this de- Our initial hypothesis has suggested that
scription of the symptoms that the subject given a state of physiological arousal for
could expect as a consequence of the shot. In which the individual has no adequate ex-
this condition, then, subjects have a com- planation, cognitive factors can lead the in-
pletely appropriate explanation of their bodily dividual to describe his feelings with any of
state. They know precisely what they will a diversity of emotional labels. In order to
feel and why. test this hypothesis, it was decided to manipu-
Epinephrine Ignorant. In this condition, late emotional states which can be considered
when the subject agreed to the injection, the quite different—euphoria and anger.
experimenter said nothing more relevant to There are, of course, many ways to induce
side effects and simply left the room. While such states. In our own program of re-
the physician was giving the injection, she search, we have concentrated on social de-
told the subject that the injection was mild terminants of emotional states and have been
and harmless and would have no side effects. able to demonstrate in other studies that
In this condition, then, the subject has no ex- people do evaluate their own feelings by com-
perimentally provided explanation for his paring themselves with others around them
bodily state. (Schachter 1959; Wrightsman 1960). In
this experiment we have attempted again to
Epinephrine Misinformed. I should also manipulate emotional state by social means.
tell you that some of our subjects have ex- In one set of conditions, the subject is placed
perienced side effects from the Suproxin. together with a stooge who has been trained
384 STANLEY SCHACHTER AND JEROME E. SINGER
to act euphorically. In a second set of con- 2. "This scrap paper isn't even much
ditions the subject is with a stooge trained good for doodling" and crumples paper and
to act in an angry fashion. attempts to throw it into wastebasket in far
corner of the room. He misses but this
Euphoria leads him into a "basketball game." He
crumples up other sheets of paper, shoots a
Immediately9 after the subject had been few baskets, says "Two points" occasion-
injected, the physician left the room and the ally. He gets up and does a jump shot
experimenter returned with a stooge whom saying, "The old jump shot is really on
he introduced as another subject, then said: today."
Both of you have had the Suproxin shot 3. If the subject has not joined in, the
and you'll both be taking the same tests of stooge throws a paper basketball to the
vision. What I ask you to do now is just subject saying, "Here, you try it."
wait for 20 minutes. The reason for this 4. Stooge continues his game saying,
is simply that we have to allow 20 minutes "The trouble with paper basketballs is that
for the Suproxin to get from the injection you don't really have any control."
site into the bloodstream. At the end of 5. Stooge continues basketball, then
20 minutes when we are certain that most gives it up saying, "This is one of my good
of the Suproxin has been absorbed into days. I feel like a kid again. I think I'll
the bloodstream, we'll begin the tests of make a plane." He makes a paper airplane
vision. saying, "I guess I'll make one of the
longer ones."
The room in which this was said had been 6. Stooge flies plane. Gets up and re-
deliberately put into a state of mild disarray. trieves plane. Flies again, etc.
As he was leaving, the experimenter apolo- 7. Stooge throws plane at subject.
getically added: 8. Stooge, flying plane, says, "Even when
I was a kid, I was never much good at
The only other thing I should do is to this."
apologize for the condition of the room. 9. Stooge tears off part of plane saying,
I just didn't have time to clean it up. So, "Maybe this plane can't fly but at least
if you need any scratch paper or rubber it's good for something." He wads up
bands or pencils, help yourself. I'll be paper and making a slingshot of a rubber
back in 20 minutes to begin the vision tests. band begins to shoot the paper.
As soon as the experimenter had left, the 10. Shooting, the stooge says, "They
[paper ammunition] really go better if you
stooge introduced himself again, made a
series of standard icebreaker comments, and make them long. They don't work right
then launched his routine. For observation if you wad them up."
purposes, the stooge's act was broken into a 11. While shooting, stooge notices a
series of standard units, demarcated by a sloppy pile of manila folders on a table.
change in activity or a standard comment. He builds a tower of these folders, then
In sequence, the units of the stooge's routine goes to the opposite end of the room to
were the following: shoot at the tower.
12. He misses several times, then hits
1. Stooge reaches for a piece of paper and cheers as the tower falls. He goes
and starts doodling saying, "They said we over to pick up the folders.
could use this for scratch, didn't they?" 13. While picking up, he notices, behind
He doodles a fish for some 30 seconds, then a portable blackboard, a pair of hula hoops
says: which have been covered with black tape
with a few wires sticking out of the tape.
6
It was, of course, imperative that the se- He reaches for these, taking one for him-
quence with the stooge begin before the sub- self and putting the other aside but within
ject felt his first symptoms for otherwise the reaching distance of the subject. The
subject would be virtually forced to interpret stooge tries the hula hoop, saying, "This
his feelings in terms of events preceding the isn't as easy as it looks."
stooge's entrance. Pretests had indicated 14. Stooge twirls hoop wildly on arm,
that, for most subjects, epinphrine-caused saying, "Hey, look at this—this is great."
symptoms began within 3-5 minutes after in- 15. Stooge replaces the hula hoop and
jection. A deliberate attempt was made then sits down with his feet on the table.
to bring in the stooge within 1 minute after Shortly thereafter the experimenter re-
the subject's injection. turns to the room.
 DETERMINANTS OF EMOTIONAL STATE 385
This routine was completely standard, "Oh for Pete's sake, what did I have for
though its pace, of course, varied depending breakfast this morning?"
upon the subject's reaction, the extent to 3. Question 9 asks, "Do you ever hear
which he entered into this bedlam and the bells ? How often ? "
extent to which he initiated activities of his The stooge remarks, "Look at Question 9.
own. The only variations from this standard How ridiculous can you get? I hear bells
routine were those forced by the subject. every time' I change classes."
Should the subject originate some nonsense 4. Question 13 requests, "List the child-
of his own and request the stooge to join in, hood diseases you have had and the age
he would do so. And, he would, of course, at which you had them" to which the
respond to any comments initiated by the stooge remarks, "I get annoyed at this
subject. childhood disease question. I can't remem-
Subjects in each of the three "appropriate- ber what childhood diseases I had, and
ness" conditions and in the placebo condi- especially at what age. Can you?"
tion were submitted to this setup. The 5. Question 17 asks "What is your fath-
stooge, of course, never knew in which con- er's average annual income?" and the
dition any particular subject fell. stooge says, "This really irritates me. It's
none of their business what my father
Anger makes. I'm leaving that blank."
6. Question 25 presents a long series of
Immediately after the injection, the ex- items such as "Does not bathe or wash reg-
perimenter brought a stooge into the subject's ularly," "Seems to need psychiatric care,"
room, introduced the two and after explain- etc. and requests the respondent to write
ing, the necessity for a 20 minute delay for down for which member of his immediate
"the Suproxin to get from the injection site family each item seems most applicable.
into the bloodstream" he continued, "We The question specifically prohibits the an-
would like you to use these 20 minutes to swer "None" and each item must be an-
answer these questionnaires." Then handing swered. The stooge says, "I'll be damned if
out the questionnaires, he concludes with, I'll fill out Number 25. 'Does not bathe or
"I'll be back in 20 minutes to pick up the wash regularly'—that's a real insult." He
questionnaires and begin the tests of vision." then angrily crosses out the entire item.
Before looking at the questionnaire, the 7. Question 28 reads :
stooge says to the subject, "How many times each week do you
I really wanted to come for an experi- have sexual intercourse?" 0-1 2-3
ment today, but I think it's unfair for them 4-6 7 and over The
to give you shots. At least, they should stooge bites out, "The hell with it! I
have told us about the shots when they don't have to tell them all this."
called us; you hate to refuse, once you're 8. The stooge sits sullenly for a few mo-
here already. ments then he rips up his questionnaire,
crumples the pieces and hurls them to the
The questionnaires, five pages long, start floor, saying, "I'm not wasting any more
off innocently requesting face sheet informa- time. I'm getting my books and leaving"
tion and then grow increasingly personal and and he stamps out of the room.
insulting. The stooge, sitting directly op- 9. The questionnaire continues for eight
posite the subject, paces his own answers so more questions ending with: "With how
that at all times subject and stooge are work- many men (other than your father) has
ing on the same question. At regular points your mother had extramarital relation-
in the questionnaire, the stooge makes a ships?"
series of standardized comments about the 4 and under : 5-9 : 10 and
questions. His comments start off innocently over
enough, grow increasingly querulous, and
finally he ends up in a rage. In sequence, Subjects in the Epi Ign, Epi Inf and
he makes the following comments. Placebo conditions were run through this
"anger" inducing sequence. The stooge,
1. Before answering any items, he leafs again, did not know to which condition the
quickly through the questionnaire saying, subject had been assigned.
"Boy, this is a long one." In summary, this is a seven condition ex-
2. Question 7 on the questionnaire re- periment which, for two different emotional
quests, "List the foods that you would eat states, allows us (a) to evaluate the effects
in a typical day." The stooge comments, of "appropriateness" on emotional inducibility
386 STANLEY SCHACHTER AND JEROME E. SINGER

and (i>) to begin to evaluate the effects of stooge. Subjects who paid flatly no atten-
sympathetic activation on emotional induci- tion to the stooge or who, with or without
bility. In schematic form the conditions are comment, simply watched the stooge without
the following: joining in his activity were coded in these
categories.
EUPHORIA ANGER
For any particular unit of behavior, the
Epi Inf Epi Inf subject's behavior was coded in one or more
Epi Ign Epi Ign of these categories. To test reliability of
Epi Mis Placebo coding two observers independently coded
Placebo two experimental sessions. The observers
The Epi Mis condition was not run in the agreed completely on the coding of 88%
Anger sequence. This was originally con- of the units.
ceived as a control condition and it was felt Anger. For each of the units of stooge
that its inclusion in the Euphoria conditions behavior, an observer recorded the subject's
alone would suffice as a, means of evaluating responses and coded them according to the
the possible artifactual effect of the Epi Inf following category scheme:
instructions. Category 1: Agrees. In response to the
stooge the subject makes a comment indi-
Measurement cating that he agrees with the stooge's stand-
ardized comment or that he, too, is irked by
Two types of measures of emotional state a particular item on the questionnaire. For
were obtained. Standardized observation example, a subject who responded to the
through a one-way mirror was the technique stooge's comment on the "father's income"
used to assess the subject's behavior. To question by saying, "I don't like that kind
what extent did he act euphoric or angry? of personal question either" would be so
Such behavior can be considered in a way as coded (scored +2).
a "semiprivate" index of mood for as far as Category 2: Disagrees. In response to the
the subject was concerned, his emotional be- stooge's comment, the subject makes a com-
havior could be known only to the other per- ment which indicates that he disagrees with
son in the room—presumably another stu- the stooge's meaning or mood; e.g., in re-
dent. The second type of measure was self- sponse to the stooge's comment on the "fath-
report in which, on a variety of scales, the er's income" question, such a subject might
subject indicated his mood of the moment. say, "Take it easy, they probably have a
Such measures can be considered "public" good reason for wanting the information"
indices of mood for they would, of course, (scored -2).
be available to the experimenter and his Category 3: Neutral. A noncommittal or
associates. irrelevant response to the stooge's remark
(scored 0).
Observation Category 4: Initiates agreement or dis-
Euphoria. For each of the first 14 units agreement. With no instigation by the
of the stooge's standardized routine an ob- stooge, a subject, so coded, would have
server kept a running chronicle of what the volunteered a remark indicating that he felt
subject did and said. For each unit the ob- the same way or, alternatively, quite differ-
server coded the subject's behavior in one ently than the stooge. Examples would be
or more of the following categories: "Boy I hate this kind of thing^' or "I'm en-
Category 1: Joins in activity. If the sub- joying this" (scored +2 or —2).
ject entered into the stooge's activities, e.g., Category S: Watches. The subject makes
if he made or flew airplanes, threw paper no verbal response to the stooge's comment
basketballs, hula hooped, etc., his behavior but simply looks directly at him (scored 0).
was coded in this category. Category 6: Ignores. The subject makes
Category 2 : Initiates new activity. A sub- no verbal response to the stooge's comment
ject was so coded if he gave indications of nor does he look at him; the subject, paying
creative euphoria, that is, if, on his own, he no attention at all to the stooge, simply works
initiated behavior outside of the stooge's rou- at his own questionnaire (scored —1).
tine. Instances of such behavior would be A subject was scored in one or more of
the subject who threw open the window and, these categories for each unit of stooge be-
laughing, hurled paper basketballs at pass- havior. To test reliability, two observers in-
ersby; or, the subject who jumped on a table dependently coded three experimental ses-
and spun one hula hoop on his leg and the sions. In order to get a behavioral index
other on his neck. of anger, observation protocol was scored ac-
Categories 3 and 4: Ignores or watches cording to the values presented in parentheses
 DETERMINANTS OF EMOTIONAL STATE 387

after each of the above definitions of cate- To measure the physical effects of epi-
gories. In a unit-by-unit comparison, the two nephrine and determine whether or not the
observers agreed completely on the scoring injection had been successful in producing the
of 71% of the units jointly observed. The necessary bodily state, the following questions
scores of the two observers differed by a were asked:
value of 1 or less for 88% of the units coded 1. Have you experienced any palpitation
and in not a single case did the two observers (consciousness of your own heart beat) ?
differ in the direction of their scoring of
I
a unit.
Not at A slight A moderate An intense
all amount amount amount
Self Report of Mood and Physical (0) (0 (2) (3)
Condition 2. Did you feel any tremor (involuntary
When the subject's session with the stooge shaking of the hands, arms or legs) ?
was completed, the experimenter returned to
the room, took pulses and said: I l l I
Not at A slight A moderate An intense
Before we proceed with the vision tests, all amount amount amount
there is one other kind of information (0) (1) (2) (3)
which we must have. We have found, as To measure possible effects of the instruc-
you can probably imagine, that there are tions in the Epi Mis condition, the following
many things beside Suproxin that affect questions were asked:
how well you see in our tests. How
hungry you are, how tired you are, and 1. Did you feel any numbness in your feet?
2. Did you feel any itching sensation ?
even the mood you're in at the time—
whether you feel happy or irritated at the 3. Did you experience any feeling of head-
time of testing will affect how well you ache?
see. To understand the data we collect on To all three of these questions was at-
you, then, we must be able to figure out tached a four-point scale running from "Not
which effects are due to causes such as at all" to "An intense amount."
these and which are caused by Suproxin. In addition to these scales, the subjects
The only way we can get such informa- were asked to answer two open-end questions
tion about your physical and emotional on other physical or emotional sensations
state is to have you tell us. I'll hand out they may have experienced during the experi-
these questionnaires and ask you to an- mental session. A final measure of bodily
swer them as accurately as possible. Ob- state was pulse rate which was taken by the
viously, our data on the vision tests will physician or the experimenter at two times—•
only be as accurate as your description of immediately before the injection and immedi-
your mental and physical state. ately after the session with the stooge.
When the subjects had completed these
In keeping with this spiel, the questionnaire questionnaires, the experimenter announced
that the experimenter passed out contained a that the experiment was over, explained
number of mock questions about hunger, the deception and its necessity in detail, an-
fatigue, etc., as well as questions of more swered any questions, and swore the subjects
immediate relevance to the experiment. To to secrecy. Finally, the subjects answered a
measure mood or emotional state the follow- brief questionnaire about their experiences, if
ing two were the crucial questions: any, with adrenalin and their previous knowl-
edge or suspicion of the experimental setup.
1. How irritated, angry or annoyed would There was no indication that any of the sub-
you say you feel at present ?
jects had known about the experiment before-
! I ! hand but 11 subjects were so extremely sus-
I don't I feel I feel I feel I feel picious of some crucial feature of the
feel at all a little quite very extremely experiment that their data were automatically
irritated irritated irritated irritated irritated
or angry and angry and angry and angry and angry discarded.
(0) (1) (2) (3) (4)
2. How good or happy would you say you Subjects
feel at present? The subjects were all male, college stu-
dents taking classes in introductory psychol-
I don't I feel I feel I feel I feel ogy at the University of Minnesota. Some
feel at all a little quite very extremely 90% of the students in these classes volunteer
happy happy happy happy happy
or good and good and good and good and good for a subject pool for which they receive two
(0) (1) (2) (3) (4) extra points on their final exam for every
388 STANLEY SCHACHTER AND JEROME E. SINGER

hour that they serve as experimental sub- that the experimental situation effectively
jects. For this study the records of all po- produces sympathetic stimulation in placebo
tential subjects were cleared with the Student subjects, the proposition is difficult to test, for
Health Service in order to insure that no such a factor would attenuate differences be-
harmful effects would result from the tween epinephrine and placebo subjects.
injections. Both of these factors, then, can be expected
to interfere with the test of our several
Evaluation of the Experimental Design propositions. In presenting the results of
The ideal test of our propositions would this study, we shall first present condition by
require circumstances which our experiment condition results and then evaluate the effect
is far from realizing. First, the proposition of these two factors on experimental differ-
that: "A state of physiological arousal for ences.
which an individual has no immediate ex- RESULTS
planation will lead him to label this state in
terms of the cognitions available to him" Effects of the Injections on Bodily
obviously requires conditions under which State
the subject does not and cannot have a proper
explanation of his bodily state. Though we Let us examine first the success of
toyed with such fantasies as ventilating the the injections at producing the bodily
experimental room with vaporized adrenalin, state required to examine the proposi-
reality forced us to rely on the disguised in-
jection of Suproxin—a technique which was tions at test. Does the injection of
far from ideal for no matter what the ex- epinephrine produce symptoms of sym-
perimenter told them, some subjects would pathetic discharge as compared with
inevitably attribute their feelings to the in- the placebo injection? Relevant data
jection. To the extent that subjects did so,
are presented in Table 1 where it can
differences between the several appropriate-
ness conditions should be attenuated. be immediately seen that on all items
Second, the proposition that: "Given the subjects who were in epinephrine con-
same cognitive circumstances the individual ditions show considerably more evi-
will react emotionally only to the extent that dence of sympathetic activation than do
he experiences a state of physiological
arousal" requires for its ideal test the ma- subjects in placebo conditions. In all
nipulation of states of physiological arousal epinephrine conditions pulse rate in-
and of physiological quiescence. Though creases significantly when compared
there is no question that epinephrine effec- with the decrease characteristic of the
tively produces a state of arousal, there is
also no question that a placebo does not pre- placebo conditions. On the scales it is
vent physiological arousal. To the extent clear that epinephrine subjects experi-

TABLE 1
THE EFFECTS OF THE INJECTIONS ON BODILY STATE

Pulse Self-rating of
PnnHif ion N
Pre Post Palpitation Tremor Numbness Itching Headache

Euphoria
Epi Inf 27 85.7 88.6 1.20 1.43 0 0.16 0.32
Epi Ign 26 84.6 85.6 1.83 1.76 0.15 0 0.55
Epi Mis 26 82.9 86.0 1.27 2.00 0.06 0.08 0.23
Placebo 26 80.4 77.1 0.29 0.21 0.09 0 0.27
Anger
Epi Inf 23 85.9 92.4 1.26 1.41 0.17 0 0.11
Epi Ign 23 85.0 96.8 1.44 1.78 0 0.06 0.21
Placebo 23 84.5 79.6 0.59 0.24 0.14 0.06 0.06
 DETERMINANTS OF EMOTIONAL STATE 389

ence considerably more palpitation and of the point he checks on the irritation
tremor than do placebo subjects. In scale from the value of the point he
all possible comparisons on these symp- checks on the happiness scale. Thus,
toms, the mean scores of subjects in if a subject were to check the point "I
any of the epinephrine conditions are feel a little irritated and angry" on the
greater than the corresponding scores irritation scale and the point "I feel
in the placebo conditions at better than very happy and good" on the happiness
the .001 level of significance. Exami- scale, his score would be +2. The
nation of the absolute values of these higher the positive value, the happier
scores makes it quite clear that subjects and better the subject reports himself
in epinephrine conditions were, indeed, as feeling. Though we employ an in-
in a state of physiological arousal, while dex for expositional simplicity, it
most subjects in placebo conditions should be noted that the two com-
were in a relative state of physiological ponents of the index each yield results
quiescence. completely consistent with those ob-
The epinephrine injection, of course, tained by use of this index.
did not work with equal effectiveness Let us examine first the effects of the
for all subjects; indeed for a few sub- appropriateness instructions. Compari-
jects it did not work at all. Such sub- son of the scores for the Epi Mis and
jects reported almost no palpitation or Epi Inf conditions makes it immediately
tremor, showed no increase in pulse clear that the experimental differences
and described no other relevant phys- are not due to artifacts resulting from
ical symptoms. Since for such subjects the informed instructions. In both con-
the necessary experimental conditions ditions the subject was warned to ex-
were not established, they were auto- pect a variety of symptoms as a conse-
matically excluded from the data and quence of the injection. In the Epi
all further tabular presentations will not Mis condition, where the symptoms
include such subjects. Table 1, how- were inappropriate to the subject's
ever, does include the data of these sub- bodily state the self-report score is al-
jects. There were four such subjects most twice that in the Epi Inf condition
in euphoria conditions and one of them where the symptoms were completely
in anger conditions. appropriate to the subject's bodily state.
In order to evaluate further data on It is reasonable, then, to attribute dif-
Epi Mis subjects it is necessary to note ferences between informed subjects
the results of the "numbness," "itch- and those in other conditions to dif-
ing," and "headache" scales also pre- ferences in manipulated appropriate-
sented in Table 1. Clearly the subjects ness rather than to artifacts such as
in the Epi Mis condition do not differ introspectiveness or self-examination.
on these scales from subjects in any of It is clear that, consistent with expec-
the other experimental conditions. tations, subjects were more susceptible
to the stooge's mood and consequently
Effects of the Manipulations on more euphoric when they had no ex-
Emotional State planation of their own bodily states
Euphoria: Self-report. The effects than when they did. The means of
of the several manipulations on emo- both the Epi Ign and Epi Mis condi-
tional state in the euphoria conditions tions are considerably greater than the
are presented in Table 2. The scores mean of the Epi Inf condition.
recorded in this table are derived, for It is of interest to note that Epi Mis
each subject, by subtracting the value subjects are somewhat more euphoric
390 STANLEY SCHACHTER AND JEROME E. SINGER

TABLE 2 ateness, with the Epi Inf condition at

SELF-REPORT OF EMOTIONAL STATE IN one extreme and the Epi Mis condition
THE EUPHORIA CONDITIONS at the other.
Comparing the placebo to the epi-
Condi- Self- nephrine conditions, we note a pattern
tion N Report Comparison #"
scales which will repeat itself throughout the
Epi Inf 25 0.98
1.78
Epi Inf vs. Epi Mis
Epi Inf vs. Epi Ign
<.01
.02
data. Placebo subjects are less eu-
Epi Ign 25
Epi Mis 25 1.90 Placebo vs. Epi Mis, ns phoric than either Epi Mis or Epi Ign
Placebo 26 1.61 Ign, or Inf
subjects but somewhat more euphoric
All p values reported throughout paper are two- than Epi Inf subjects. These differ-
tailed. ences are not, however, statistically
significant. We shall consider the
than are Epi Ign subjects. This pat- epinephrine-placebo comparisons in de-
tern repeats itself in other data shortly tail in a later section of this paper fol-
to be presented. We would attribute lowing the presentation of additional
this difference to differences in the ap- relevant data. For the moment, it is
propriateness dimension. Though, as clear that, by self-report manipulating
in the Epi Ign condition, a subject is appropriateness has had a very strong
not provided with an explanation of his effect on euphoria.
bodily state, it is, of course, possible Behavior. Let us next examine the
that he will provide one for himself extent to which the subject's behavior
which is not derived from his interac- was affected by the experimental ma-
tion with the stooge. Most reasonably nipulations. To the extent that his
he could decide for himself that he mood has been affected, one should ex-
feels this way because of the injection. pect that the subject will join in the
To the extent that he does so he should stooge's whirl of manic activity and
be less susceptible to the stooge. It initiate similar activities of his own.
seems probable that he would be less The relevant data are presented in
likely to hit on such an explanation in Table 3. The column labeled "Activity
the Epi Mis condition than in the Epi
Ign condition for in the Epi Mis condi- TABLE 3
tion both the experimenter and the doc- BEHAVIORAL INDICATIONS OF EMOTIONAL
tor have told him that the effects of the STATE IN THE EUPHORIA CONDITIONS
injection would be quite different from
what he actually feels. The effect of Activity Mean num-
Condition N index ber of acts
such instructions is probably to make initiated
it more difficult for the subject himself Epi Inf 25 12.72 .20
to hit on the alternative explanation de- Epi Ign 25 18.28 .56
scribed above. There is some evidence Epi Mis 25 22.56 .84
Placebo 26 16.00 .54
to support this analysis. In open-end
questions in which subjects described P value
their own mood and state, 28% of the
subjects in the Epi Ign condition made Comparison Activity Initiates
index
some connection between the injection
Epi Inf vs. Epi Mis .05 .03
and their bodily state compared with Epi Inf vs. Epi Ign ns .08
the 16% of subjects in the Epi Mis Plac vs. Epi Mis, Ign, ns ns
condition who did so. It could be con- or Inf
sidered, then, that these three condi- a
Tested by Xz comparison of the proportion of sub-
tions fall along a dimension of appropri- jects in each condition initiating new acts.
 DETERMINANTS OF EMOTIONAL STATE 391

index" presents summary figures on the euphoric than placebo subjects but not
extent to which the subject joined in significantly so.
the stooge's activity. This is a Anger: Self-report. Before present-
weighted index which reflects both the ing data for the anger conditions, one
nature of the activities in which the point must be made about the anger ma-
subject engaged and the amount of time nipulation. In the situation devised,
he was active. The index was devised anger, if manifested, is most likely to be
by assigning the following weights to directed at the experimenter and his
the subject's activities: 5—hula hoop- annoyingly personal questionnaire. As
ing; 4—shooting with slingshot; 3— we subsequently discovered, this was
paper airplanes; 2—paper basketballs; rather unfortunate, for the subjects,
1—doodling; 0—does nothing. Pre- who had volunteered for the experiment
test scaling on 15 college students for extra points on their final exam,
ordered these activities with respect to simply refused to endanger these points
the degree of euphoria they repre- by publicly blowing up, admitting their
sented. Arbitrary weights were as- irritation to the experimenter's face or
signed so that the wilder the activity, spoiling the questionnaire. Though as
the heavier the weight. These weights the reader will see, the subjects were
are multiplied by an estimate of the quite willing to manifest anger when
amount of time the subject spent in they were alone with the stooge, they
each activity and the summed products hesitated to do so on material (self-
make up the activity index for each sub- ratings of mood and questionnaire)
ject. This index may be considered a that the experimenter might see and
measure of behavioral euphoria. It only after the purposes of the experi-
should be noted that the same between- ment had been revealed were many of
condition relationships hold for the two these subjects willing to admit to the
components of this index as for the experimenter that they had been irked
index itself. or irritated.
The column labeled "Mean number This experimentally unfortunate situ-
of acts initiated" presents the data on ation pretty much forces us to rely on
the extent to which the subject deviates the behavioral indices derived from ob-
from the stooge's routine and initiates servation of the subject's presumably
euphoric activities of his own. private interaction with the stooge.
On both behavioral indices, we find We do, however, present data on the
precisely the same pattern of relation- self-report scales in Table 4. These
ships as those obtained with self-re- figures are derived in the same way as
ports. Epi Mis subjects behave some- the figures presented in Table 2 for the
what more euphorically than do Epi euphoria conditions, that is, the value
Ign subjects who in turn behave more checked on the irritation scale is sub-
euphorically than do Epi Inf subjects. tracted from the value checked on the
On all measures, then, there is con- happiness scale. Though, for the rea-
sons stated above, the absolute magni-
sistent evidence that a subject will take
tude of these figures (all positive) is
over the stooge's euphoric mood to the
relatively meaningless, we can, of
extent that he has no other explana- course, compare condition means within
tion of his bodily state. the set of anger conditions. With the
Again it should be noted that on happiness-irritation index employed,
these behavioral indices, Epi Ign and we should, of course, anticipate pre-
Epi Mis subjects are somewhat more cisely the reverse results from those ob-
392 STANLEY SCHACHTER AND JEROME E. SINGER
TABLE 4 Table 5. For this analysis, the stooge's
SELF-REPORT OF EMOTIONAL STATE IN routine has been divided into two
THE ANGER CONDITIONS phases—the first two units of his be-
havior (the "long" questionnaire and
Self-
Condition N Report Comparison t "What did I have for breakfast?") are
scales considered essentially neutral revealing
Epi Inf 22 1.91 Epi Inf vs. Epi .08 nothing of the stooge's mood; all of the
Ign following units are considered "angry"
Epi Ign 23 1.39 Placebo vs. Epi ns
Ign or Inf units for they begin with an irritated re-
Placebo 23 1.63 mark about the "bells" question and
end with the stooge's fury as he rips
up his questionnaire and stomps out
tained in the euphoria conditions; that of the room. For the neutral units,
is, the Epi Inf subjects in the anger agreement or disagreement with the
conditions should again be less suscep- stooge's remarks is, of course, mean-
tible to the stooge's mood and should, ingless as an index of mood and we
therefore, describe themselves as in a should anticipate no difference between
somewhat happier frame of mind than conditions. As can be seen in Table
subjects in the Epi Ign condition. 5, this is the case.
This is the case; the Epi Inf subjects For the angry units, we must, of
average 1.91 on the self-report scales course, anticipate that subjects in the
while the Epi Ign subjects average Epi Ign condition will be angrier than
1.39. subjects in the Epi Inf condition. This
Evaluating the effects of the injec- is indeed the case. The Anger index
tions, we note again that, as antici- for the Epi Ign condition is positive
pated, Epi Ign subjects are somewhat and large, indicating that these subjects
less happy than Placebo subjects but, have become angry, while in the Epi
once more, this is not a significant Inf condition the Anger index is
difference. slightly negative in value indicating
Behavior. The subject's responses that these subjects have failed to catch
to the stooge, during the period when the stooge's mood at all. It seems clear
both were filling out their question- that providing the subject with an ap-
naires, were systematically coded to
provide a behavioral index of anger. TABLE 5
The coding scheme and the numerical
BEHAVIORAL INDICATIONS OF EMOTIONAL
values attached to each of the categories STATE IN THE ANGER CONDITIONS
have been described in the methodology
section. To arrive at an "Anger in- Condition Neutral Anger
N units units
dex" the numerical value assigned to
a subject's responses to the stooge is Epi Inf 22 +0.07 -0.18
summed together for the several units Epi Ign 23 +0.30 +2.28
Placebo 22» -0.09 +0.79
of stooge behavior. In the coding
scheme used, a positive value to this Comparison for anger units P
index indicates that the subject agrees
with the stooge's comment and is grow- Epi Inf vs. Epi Ign
Epi Ign vs. Placebo <.05
ing angry. A negative value indicates Placebo vs. Epi Inf ns
that the subject either disagrees with
the stooge or ignores him. « For one subject in this condition the sound system
went dead and the observer could not, of course, code
The relevant data are presented in his reactions.
 DETERMINANTS OF EMOTIONAL STATE 393

propriate explanation of his bodily tional level is consistently greater in the

state greatly reduces his tendency to Epi Mis and Epi Ign conditions than
interpret his state in terms of the cog- in the Placebo condition, this differ-
nitions provided by the stooge's angry ence is significant at acceptable prob-
behavior. ability levels only in the anger con-
Finally, on this behavioral index, it ditions.
can be seen that subjects in the Epi Ign In order to explore the problem fur-
condition are significantly angrier than ther, let us examine the experimental
subjects in the Placebo condition. Be- factors identified earlier, which might
haviorally, at least, the injection of epi- have acted to restrain the emotional
nephrine appears to have led subjects level in the Epi Ign and Epi Mis con-
to an angrier state than comparable ditions. As was pointed out earlier,
subjects who received placebo shots. the ideal test of our first two hypotheses
requires an experimental setup in which
Conformation of Data to Theoretical the subject has flatly no way of evalu-
Expectations ating his state of physiological arousal
Now that the basic data of this study other than by means of the experimen-
have been presented, let us examine tally provided cognitions. Had it been
closely the extent to which they con- possible to physiologically produce a
form to theoretical expectations. If our state of sympathetic activation by means
hypotheses are correct and if this ex- other than injection, one could have
perimental design provided a perfect approached this experimental ideal
test for these hypotheses, it should be more closely than in the present setup.
anticipated that in the euphoria condi- As it stands, however, there is always
tions the degree of experimentally pro- a reasonable alternative cognition avail-
duced euphoria should vary in the able to the aroused subject—he feels
following fashion: the way he does because of the injec-
tion. To the extent that the subject
Epi Mis ^ Epi Ign > Epi Inf = Placebo seizes on such an explanation of his
And in the anger conditions, anger bodily state, we should expect that he
should conform to the following pat- will be uninfluenced by the stooge.
tern : Evidence presented in Table 6 for the
anger condition and in Table 7 for the
Epi Ign > Epi Inf = Placebo
euphoria conditions indicates that this
In both sets of conditions, it is the is, indeed, the case.
case that emotional level in the Epi As mentioned earlier, some of the
Mis and Epi Ign conditions is con- Epi Ign and Epi Mis subjects in their
siderably greater than that achieved in answers to the open-end questions
the corresponding Epi Inf conditions. clearly attributed their physical state
The results for the Placebo condition, to the injection, e.g., "the shot gave
however, are ambiguous for consist- me the shivers." In Tables 6 and 7
ently the Placebo subjects fall between such subjects are labeled "Self-in-
the Epi Ign and the Epi Inf subjects. formed." In Table 6 it can be seen
This is a particularly troubling pattern that the self-informed subjects are con-
for it makes it impossible to evaluate siderably less angry than are the re-
unequivocally the effects of the state of maining subjects; indeed, they are not
physiological arousal and indeed raises angry at all. With these self-informed
serious questions about our entire theo- subjects eliminated the difference be-
retical structure. Though the emo- tween the Epi Ign and the Placebo
394 STANLEY SCHACHTER AND JEROME E. SINGER

TABLE 6 by assuming that emotional state is a

THE EFFECTS OF ATTRIBUTING BODILY STATE joint function of a state of physiologi-
TO THE INJECTION ON ANGER IN THE cal arousal and of the appropriateness
ANGER EPI IGN CONDITION of a cognition we are, in effect, as-
suming a multiplicative function, so
Anger that if either component is at zero, emo-
Condition N index P
tional level is at zero. As noted earlier
Self-informed subjects 3 -1.67 ns this expectation should hold if we can
Others 20 + 2.88 ns
Self-informed vs. .05 be sure that there is no sympathetic
Others activation in the Placebo conditions.
This assumption, of course, is com-
pletely unrealistic for the injection of
conditions is significant at the .01 level placebo does not prevent sympathetic
of significance. activation. The experimental situa-
Precisely the same pattern is evident tions were fairly dramatic and cer-
in Table 7 for the euphoria conditions. tainly some of the placebo subjects gave
In both the Epi Mis and the Epi Ign indications of physiological arousal. If
conditions, the self-informed subjects our general line of reasoning is correct,
have considerably lower activity indices it should be anticipated that the emo-
than do the remaining subjects. Elim- tional level of subjects who give indi-
inating self-informed subjects, compari- cations of sympathetic activity will be
son of both of these conditions with the greater than that of subjects who do
Placebo condition yields a difference not. The relevant evidence is pre-
significant at the .03 level of signifi- sented in Tables 8 and 9.
cance. It should be noted, too, that the As an index of sympathetic activa-
self-informed subjects have much the tion we shall use the most direct and
same score on the activity index as do unequivocal measure available—change
the experimental Epi Inf subjects
(Table 3). TABLE 7
It would appear, then, that the ex- THE EFFECTS OF ATTRIBUTING BODILY STATE
perimental procedure of injecting the TO THE INJECTION ON EUPHORIA IN
subjects, by providing an alternative THE EUPHORIA EPI IGN AND
cognition, has, to some extent, obscured EPI Mis CONDITIONS
the effects of epinephrine. When ac-
count is taken of this artifact, the evi- Epi Ign
dence is good that the state of physi-
Activity
ological arousal is a necessary com- Index
ponent of an emotional experience for Self-informed subjects 11.63 ns
8
when self-informed subjects are re- Others 17 21.14 ns
moved, epinephrine subjects give con- Self-informed vs. .05
Others
sistent indications of greater emotion-
ality than do placebo subjects.
Epi Mis
Let us examine next the fact that
consistently the emotional level, both Activity
N Index P
reported and behavioral, in Placebo
conditions is greater than that in the Self-informed subjects 5 12.40 ns
Epi Inf conditions. Theoretically, of Others 20 25.10 ns
Self-informed vs. .10
course, it should be expected that the Others
two conditions will be equally low, for
 DETERMINANTS OF EMOTIONAL STATE 395

TABLE 8 activation give indications of consider-

SYMPATHETIC ACTIVATION AND EUPHORIA ably more anger than do subjects who
IN THE EUPHORIA PLACEBO CONDITION show no such signs. Conforming to
expectations, sympathetic activation ac-
Subjects whose : Activity companies an increase in emotional
N index t
level.
Pulse decreased 14 10.67 ns
It should be noted, too, that the emo-
Pulse increased 12 23.17 ns
or remained same tional levels of subjects showing no
Pulse decrease vs. pulse .02 signs of sympathetic activity are quite
increase or same
comparable to the emotional level of
subjects in the parallel Epi Inf condi-
in pulse rate. It can be seen in Table tions (see Tables 3 and 5). The
1 that the predominant pattern in the similarity of these sets of scores and
Placebo condition is a decrease in pulse their uniformly low level of indicated
rate. We shall assume, therefore, that emotionality would certainly make it
those subjects whose pulse increases or appear that both factors are essential
remains the same give indications of to an emotional state. When either the
sympathetic activity while those sub- level of sympathetic arousal is low or
jects whose pulse decreases do not. In a completely appropriate cognition is
Table 8, for the euphoria condition, it available, the level of emotionality is
is immediately clear that subjects who low.
give indications of sympathetic activity DISCUSSION
are considerably more euphoric than Let us summarize the major findings
are subjects who show no sympathetic of this experiment and examine the ex-
activity. This relationship is, of course, tent to which they support the proposi-
confounded by the fact that euphoric tions offered in the introduction of this
subjects are considerably more active paper. It has been suggested, first, that
than noneuphoric subjects—a factor given a state of physiological arousal
which independent of mood could ele- for which an individual has no explana-
vate pulse rate. However, no such fac- tion, he will label this state in terms of
tor operates in the anger condition the cognitions available to him. This
where angry subjects are neither more implies, of course, that by manipulating
active nor talkative than calm subjects. the cognitions of an individual in such
It can be seen in Table 9 that Placebo a state we can manipulate his feelings
subjects who show signs of sympathetic in diverse directions. Experimental re-
sults support this proposition for fol-
TABLE 9 lowing the injection of epinephrine,
SYMPATHETIC ACTIVATION AND ANGER IN those subjects who had no explanation
ANGER PLACEBO CONDITION for the bodily state thus produced, gave
behavioral and self-report indications
/Va Anger
Subjects whose : index P that they had been readily manipulable
into the disparate feeling states of
Pulse decreased 13 +0.15 ns
Pulse increased 8 + 1.69 ns euphoria and anger.
or remained same From this first proposition, it must
Pusle decrease vs. pulse
increase or same .01 follow that given a state of physiologi-
cal arousal for which the individual has
*N reduced by two cases owing to failure of sound a completely satisfactory explanation,
system in one case and experimenter's failure to take
pulse in another. he will not label this state in terms of
396 STANLEY SCHACHTER AND JEROME E. SINGER

the alternative cognitions available. chlorpromazine. Laughter at a slap-

Experimental evidence strongly sup- stick movie was the dependent variable
ports this expectation. In those con- and the evidence is good that amuse-
ditions in which subjects were injected ment is a direct function of manipu-
with epinephrine and told precisely lated sympathetic activation.
what they would feel and why, they In order to make the epinephrine-
proved relatively immune to any effects placebo comparison under conditions
of the manipulated cognitions. In the which would rule out the operation of
anger condition, such subjects did not any self-informing tendency, two ex-
report or show anger; in the euphoria periments were conducted on rats. In
condition, such subjects reported them- one of these Singer (1961) demon-
selves as far less happy than subjects strated that under fear inducing condi-
with an identical bodily state but no tions, manipulated by the simultaneous
adequate knowledge of why they felt presentation of a loud bell, a buzzer,
the way they did. and a bright flashing light, rats in-
Finally, it has been suggested that jected with epinephrine were consider-
given constant cognitive circumstances, ably more frightened than rats injected
an individual will react emotionally with a placebo. Epinephrine-injected
only to the extent that he experiences rats defecated, urinated, and trembled
a state of physiological arousal. With- more than did placebo-injected rats. In
out taking account of experimental arti- nonfear control conditions, there were
facts, the evidence in support of this no differences between epinephrine and
proposition is consistent but tentative. placebo groups, neither group giving
When the effects of "self-informing" any indication of fear. In another
tendencies in epinephrine subjects and study, Latane and Schachter (1962)
of "self-arousing" tendencies in pla- demonstrated that rats injected with
cebo subjects are partialed out, the epinephrine were notably more capable
evidence strongly supports the proposi- of avoidance learning than were rats
tion. injected with a placebo. Using a modi-
The pattern of data, then, falls neatly fied Miller-Mowrer shuttlebox, these
in line with theoretical expectations. investigators found that during an ex-
However, the fact that we were forced, perimental period involving 200 massed
to some extent, to rely on internal trials, 15 rats injected with epinephrine
analyses in order to partial out the avoided shock an average of 101.2 trials
effects of experimental artifacts inevita- while 15 placebo-injected rats averaged
bly makes our conclusions somewhat only 37.3 avoidances.
tentative. In order to further test these Taken together, this body of studies
propositions on the interaction of cog- does give strong support to the proposi-
nitive and physiological determinants tions which generated these experi-
of emotional state, a series of additional mental tests. Given a state of sympa-
experiments, published elsewhere, was thetic activation, for which no imme-
designed to rule out or overcome the diately appropriate explanation is
operation of these artifacts. In the first available, human subjects can be readily
of these, Schachter and Wheeler manipulated into states of euphoria,
(1962) extended the range of manipu- anger, and amusement. Varying the
lated sympathetic activation by em- intensity of sympathetic activation
ploying three experimental groups— serves to vary the intensity of a vari-
epinephrine, placebo, and a group in- ety of emotional states in both rats and
jected with the sympatholytic agent, human subjects.
 DETERMINANTS OF EMOTIONAL STATE 397

Let us examine the implications of are specific to the state of sympathetic

these findings and of this line of activation or if they are generalizable
thought for problems in the general to other states of physiological arousal.
area of the physiology of the emotions. It is clear that from our experiments
We have noted in the introduction that proper, it is impossible to answer the
the numerous studies on physiological question for our studies have been con-
differentiators of emotional states have, cerned largely with the effects of an
viewed en masse, yielded quite incon- epinephrine created state of sympa-
clusive results. Most, though not all, thetic arousal. We would suggest,
of these studies have indicated no dif- however, that our conclusions are gen-
ferences among the various emotional eralizable to almost any pronounced in-
states. Since as human beings, rather ternal state for which no appropriate
than as scientists, we have no difficulty explanation is available. This sugges-
identifying, labeling, and distinguishing tion receives some support from the ex-
among our feelings, the results of these periences of Nowlis and Nowlis (1956)
studies have long seemed rather puz- in their program of research on the
zling and paradoxical. Perhaps because effects of drugs on mood. In their
of this, there has been a persistent work the Nowlises typically administer
tendency to discount such results as a drug to groups of four subjects who
due to ignorance or methodological in- are physically in one another's presence
adequacy and to pay far more atten- and free to interact. The Nowlises de-
tion to the very few studies which scribe some of their results with these
demonstrate some sort of physiological groups as follows:
differences among emotional states than At first we used the same drug for all
to the very many studies which indi- 4 men. In those sessions seconal, when com-
cate no differences at all. It is con- pared with placebo, increased the checking of
such words as expansive, forceful, coura-
ceivable, however, that these results geous, daring, elated, and impulsive. In our
should be taken at face value and that first statistical analysis we were confronted
emotional states may, indeed, be gener- with the stubborn fact that when the same
ally characterized by a high level of drug is given to all 4 men in a group, the N
sympathetic activation with few if any that has to be entered into the analysis is 1,
not 4. This increases the cost of an already
physiological distinguishes among the expensive experiment by a considerable fac-
many emotional states. If this is cor- tor, but it cannot be denied that the effects of
rect, the findings of the present study these drugs may be and often are quite con-
may help to resolve the problem. Ob- tagious. Our first attempted solution was to
run tests on groups in which each man had
viously this study does not rule out the a different drug during the same session, such
possibility of physiological differences as 1 on seconal, 1 on benzedrine, 1 on
among the emotional states. It is the dramamine, and 1 on placebo. What does
case, however, that given precisely the seconal do? Cooped up with, say, the ego-
tistical benzedrine partner, the withdrawn,
same state of epinephrine-induced sym- indifferent dramimine partner, and the
pathetic activation, we have, by means slightly bored lactose man, the seconal sub-
of cognitive manipulations, been able ject reports that he is distractible, dizzy,
to produce in our subjects the very drifting, glum, defiant, languid, sluggish, dis-
disparate states of euphoria and anger. couraged, dull, gloomy, lazy, and slow! This
is not the report of mood that we got when
It may indeed be the case that cognitive all 4 men were on seconal. It thus appears
factors are major determiners of the that the moods of the partners do definitely
emotional labels we apply to a common influence the effect of seconal (p. 350).
state of sympathetic arousal. It is not completely clear from this
Let us ask next whether our results description whether this "contagion"
398 STANLEY SCHACHTER AND JEROME E. SINGER

of mood is more marked in drug than of emotional labels depending on the

in placebo groups, but should this be cognitive aspects of the situation.
the case, these results would certainly 2. Given a state of physiological
support the suggestion that our findings arousal for which an individual has a
are generalizable to internal states completely appropriate explanation, no
other than that produced by an injec- evaluative needs will arise and the in-
tion of epinephrine. dividual is unlikely to label his feelings
Finally, let us consider the implica- in terms of the alternative cognitions
tions of our formulation and data for available.
alternative conceptualizations of emo- 3. Given the same cognitive circum-
tion. Perhaps the most popular cur- stances, the individual will react emo-
rent conception of emotion is in terms tionally or describe his feelings as emo-
of "activation theory" in the sense em- tions only to the extent that he experi-
ployed by Lindsley (1951) and Wood- ences a state of physiological arousal.
worth and Schlosberg (1958). As we An experiment is described which,
understand this theory, it suggests that together with the results of other
emotional states should be considered studies, supports these propositions.
as at one end of a continuum of activa-
tion which is denned in terms of degree REFERENCES
of autonomic arousal and of electro- Ax, A. F. Physiological differentiation of
encephalographic measures of activa- emotional states. Psychosom. Med., 1953,
15, 433-442.
tion. The results of the experiment de- CANNON, W. B. Bodily changes in pain,
scribed in this paper do, of course, hunger, fear and rage. (2nd ed.) New
suggest that such a formulation is not York: Appleton, 1929.
completely adequate. It is possible to CANTRIL, H., & HUNT, W. A. Emotional
have very high degrees of activation effects produced by the injection of adren-
alin. Amer. J. Psychol, 1932, 44, 300-307.
without a subject either appearing to be FESTINGER, L. A theory of social compari-
or describing himself as "emotional." son processes. Hum. Relat., 19S4, 7, 114-
Cognitive factors appear to be indis- 140.
pensable elements in any formulation HUNT, J. McV., COLE, M. W., & REIS, E. E.
of emotion. Situational cues distinguishing anger, fear,
and sorrow. Amer. J. Psychol., 1958, 71,
SUMMARY 136-151.
JAMES, W. The principles of psychology.
It is suggested that emotional states New York: Holt, 1890.
may be considered a function of a state LANDIS, C, & HUNT, W. A. Adrenalin and
of physiological arousal and of a cogni- emotion. Psychol. Rev., 1932, 39, 467-485.
tion appropriate to this state of arousal. LATANE, B., & SCHACHTER, S. Adrenalin
and avoidance learning. /. comp. physiol.
From this follows these propositions: Psychol., 1962, 65, 369-372.
1. Given a state of physiological LINDSLEY, D. B. Emotion. In S. S. Stevens
arousal for which an individual has no (Ed.), Handbook of experimental psychol-
immediate explanation, he will label ogy. New York: Wiley, 1951. Pp. 473-
516.
this state and describe his feelings in MARANON, G. Contribution a 1'etude de 1'ac-
terms of the cognitions available to him. tion emotive de 1'adrenaline. Rev. Fran-
To the extent that cognitive factors are cmse Endocrinol, 1924, 2, 301-325.
potent determiners of emotional states, NOWLIS, V., & NOWLIS, H. H. The descrip-
it should be anticipated that precisely tion and analysis of mood. Ann. N. Y.
Acad. Sci., 1956, 65, 345-355.
the same state of physiological arousal
RUCKMICK, C. A. The psychology of feeling
could be labeled "joy" or "fury" or and emotion. New York: McGraw-Hill,
"jealousy" or any of a great diversity 1936.
 DETERMINANTS OF EMOTIONAL STATE 399
SCHACHTER, J. Pain, fear, and anger in hy- non-stress conditions. Unpublished doc-
pertensives and normotensives: A psycho- toral dissertation, University of Minnesota,
physiologic study. Psychosom. Med., 1957, 1961.
19, 17-29. WOLF, S., & WOLFF, H. G. Human gastric
SCHACHTER, S. The psychology of affiliation, function. New York: Oxford Univer.
Stanford, Calif.: Stanford Univer. Press, Press, 1947.
1959. WOODWORTH, R. S., & SCHLOSBERG, H. Ex-
SCHACHTER, S., & WHEELER, L. Epi- perimental psychology. New York: Holt,
nephrine, chlorpromazine, and amusement. 19S8.
/. abnorm. soc. Psychol, 1962, 65, 121- WRIGHTSMAN, L. S. Effects of waiting with
128. others on changes in level of felt anxiety.
SINGER, J. E. The effects of epinephrine, J. abnorm. soc. Psychol., 1960, 61, 216-222.
chlorpromazine and dibenzyline upon the
fright responses of rats under stress and (Received February 17, 1961)