Journal of Personality and Social Psychology

1975, Vol. 31, No. 2, 311-327

Generality of Learned Helplessness in Man

Donald S. Hiroto and Martin E. P. Seligman
University of Pennsylvania
Learned helplessness, the interference with instrumental responding following
inescapable aversive events, has been found in animals and man. This study
tested for the generality of the debilitation produced by uncontrollable events
across tasks and motivational systems. Four experiments with college students
were simultaneously conducted: (a) pretreatment with inescapable, escapable,
or control aversive tone followed by shuttlebox escape testing; (b) pretreat-
ment with insoluble, soluble, or control discrimination problems followed by
anagram solution testing; (c) pretreatments with inescapable, escapable, or
control aversive tone followed by anagram solution testing; (d) pretreatments
with insoluble, soluble, or control discrimination problems followed by
shuttlebox escape testing. Learned helplessness was found with all four ex-
periments: Both insolubility and inescapability produced failure to escape and
failure to solve anagrams. We suggest that inescapability and insolubility both
engendered expectancies that responding is independent of reinforcement.
The generality of this process suggests that learned helplessness may be an
induced "trait."

Inescapable aversive events presented to specific environmental cues of training medi-

animals or to men result in profound inter- ate the interference. In contrast, Hiroto
ference with later instrumental learning (e.g., (1974) and Miller and Seligman (1973)
Hiroto, 1974; Overmier & Seligman, 1967; hypothesized that expectancy of indepen-
Seligman & Maier, 1967; Thornton & Jacobs, dence is an internal state of the organism
1971). If a subject can escape the aversive that is broadly transferred.
event, later instrumental behavior remains If interference with learning occurs fol-
normal. This phenomenon has been inter- lowing uncontrollable events from very dif-
preted as learned helplessness (Maier, Selig- ferent response and stimulus modalities,
man, & Solomon, 1969; Seligman, Maier, & learned helplessness would seem to be a
Solomon, 1971). This interpretation claims stable and pervasive process. To put it loosely,
that organisms learn that responding and is learned helplessness a state or trait? This
reinforcement (e.g., shock termination) are question is of particular interest since learned
independent when shock is inescapable. Such helplessness has been postulated as under-
learning undermines the motivation for initi- lying human depression (Seligman, 1973, in
ating instrumental responses. press a; Seligman, Klein, & Miller, in press).
Is learned helplessness a specific state Hiroto (1974) reported results which dem-
which only impairs performance in situations onstrated parallel behaviors between animals
similar to original training, or does it impair and man in a learned helplessness paradigm.
a broad range of behavior? It is possible that One group received aversive loud noise which
it could escape by button pressing. A second
The research reported in this article was con- group received inescapable noise, and a third
ducted while the first author was a Postdoctoral group received no pretreatment. All groups
Fellow at the University of Pennsylvania. This re-
search was supported, in part, by National Institute
then received controllable noise in a two-way
of Mental Health Grant MHS3982. shuttlebox. As with animals, the inescapable
Requests for reprints should be sent to Donald S. group tended to sit and take the noise with-
Hiroto, University of California Medical Center, out responding, while the escape and no-
Ambulatory Psychiatric Services, Room A-830, 400 pretreatment groups escaped readily. A per-
Parnassus Avenue, San Francisco, California 94143.
This study is part of a program of research at
sonality measure, external control of rein-
the University of Pennsylvania supported, in part, forcement (Rotter, 1966), as well as the in-
by National Institute of Mantal Health Grant structions of chance both produced passivity
MH19604 to Martin E. P. Seligman. similar to the effects of inescapability. Hiroto
311
312 DONALD S. HIROTO AND MARTIN E. P. SELIGMAN
INS1 I N S I COG INS! INST-COG COG-COG
button Pressing Levine Discrimination Button Pressing Levine Discrimination
Pretreatment- Pretreatment Pretreatment- Pretreatment-
Shuttlebox Testing Shuttlebox Testing Anagram Testing Anagram Testing

Escapable Tone Soluble Problem

E S Escapable Tone Soluble Problem
Pretreatmenl E S
Contingency
Inescapable Tone Insoluble Problem
E S
Inescapably Tone Insoluble Problem
E S
Control Control
C C

Control Control
FIGURE 1. Designs of experiments. (Inst. refers to
instrumental; Cog. refers to cognitive.) C C

FIGURE 2. Designs of experiments. (Inst. refers to

instrumental; Cog. refers to cognitive.)
20 •

18 •

16 •

Z
0 14 •
5
in
au 12 •

u
a 10 •
<
u
in
m 8 -
0

in 6 •
4
a
4 •

4
ill
2 •

S.O. 2.61 S.D. 8.38 S.D. 5.58

Esc APE INESCAPE CONTRO

— P<.01 ' 1 P<.05

FIGURE 3. Group means, standard deviations, and significance levels of three

dependent variables for the instrumental pretreatment-instrumental test for
helplessness experiment: mean trials to criterion for escape acquisition.
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 313

20

12

n 10
u

8
E
J

I
U.
0
0

B.D. 2.45 8.D. 6.80 B.D. 5.87

ESCAPE INHCAPE

.05-

-P> .05
FIGURE 4. Group means, standard deviations, and significance levels of three
dependent variables for the instrumental pretreatment-instrumental test for
helplessness experiment: mean number of failures to escape.

concluded that all three variables—inescapa- of responding. So we determine whether in-

bility, externality, and chance instructions— strumental helplessness or mastery transfer
engender a similar state, that is, the expec- to cognitive task and whether cognitive help-
tancy that reinforcement is independent of lessness or mastery transfer to instrumental
responding. tasks.
The Hiroto study, like most others in the METHOD
helplessness literature, tested its subjects in a Overview
situation similar, but not identical, to the one
The study consisted of four independent, but si-
in which helplessness is trained: Both train- multaneously conducted, experiments. Two of the
ing and testing used aversive loud noise and four experiments used an instrumental (Inst.) or
an instrumental problem. The study we report cognitive (Cog.) pretreatment followed by an
trains its subjects in either instrumental help- instrumental test for helplessness (Inst.-Inst., Cog.-
Inst.). The other two experiments used an instru-
lessness with loud noise or insoluble cognitive
mental or cognitive pretreatment followed by a cog-
problems and then tests them in both new nitive test for helplessness (Inst.-Cog., Cog.-Cog.).
instrumental or cognitive problems which are Each experiment consisted of three groups which
soluble. It should be pointed out that insolu- received escapable (soluble), inescapable (insoluble),
bility in a cognitive task is formally anal- or a control pretreatment prior to the test for help-
lessness. Figure 1 presents the design for the Inst.-
ogous to inescapability, since in both the Inst. and Cog.-Inst. experiments, and Figure 2 pre-
probability of reinforcement (correct or in- sents the design for the Inst.-Cog. and Cog.-Cog.
correct, or shock or no shock) is independent experiments.
314 DONALD S. HIEOTO AND MARTIN E. P. SELIGMAN

Subjects Cognitive pretreatment. A series of four-dimen-

sional stimulus patterns used in previous discrimi-
Ninety-six undergraduate students, consisting of nation learning studies (Levine, 1966, 1971) was
SI men and 45 women, at the University of Penn- the cognitive task. Each of the four dimensions had
sylvania participated in the study. They responded two associated values: (a) letter (A or T ) ; (b)
to advertisements for a study in noise pollution and letter color (black or white); (c) letter size (large
were advised prior to their participation that they or small) ; (d) border surrounding letter (circle or
might be exposed to a "slightly unpleasant" tone. square). In the soluble condition, one value of one
Two women refused to participate and were re- of the dimensions, for example, square border, was
placed. None of the subjects had previously par- always correct. In the insoluble condition, no value
ticipated in an experiment using an aversive stimu- was consistently correct. Levine (1971) provides a
lus. Each subject was paid $2.00 at the end of the detailed description of the patterns. Each pattern
experiment and debriefed. was on a 4 X 6-in. (10 X 16 cm.) "wire-index" card.
Soluble instrumental test task. The apparatus was
Apparatus a modified Turner and Solomon (1962) human
shuttlebox. Moving a knob from one side to the
Instrumental pretreatment. The apparatus was a other escaped and avoided the tone. Specifications of
spring-loaded button in the center of a 12-in. (30 the manipulandum have been described elsewhere
cm.) circular base. Symmetrically on each side of the (Hiroto, 1974).
button were two 24-V dc lights. In the escapable The aversive stimulus in both instrumental tasks
condition, pressing the button four times terminated was a 3,000 hertz tone emanating from an Eico
the tone. In the inescapable condition the button audio generator (Model 377). The tone was pre-
had no effect on the tone. sented to the subject at 90 decibels through cali-

4
IBB^^

7 ^

ui 6 i
in

•
5
\
h
2 4 i

| •3

2 •

1 •

S.D. 1.96 S.D. 1.91 S.D. 1.43

ESCAPE 1NEBCAPE CONTP.O

P<.02 ' ' P>.05

O~> OS —^^— ———

FIGURE S. Group means, standard deviations, and significance levels of three

dependent variables for the instrumental pretreatment-instrumental test for
helplessness experiment: mean response latency.
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 315

20r

* 18 .
g
2 16 -
E
u 14 m

Ul

\ 12 •
u
in
u 10 m

0
h
8 m
in
1 6•
h
Z 4 -
Ul
S
2 '
S.D. 6.71 so. 6 4 7

SOL.UBLK CONTHOI-

-P<.01 p<.02

• P> 05

FIGURE 6. Group means, standard deviations, and significance levels of three

dependent variables for the cognitive pretreatment-instrumental test for help-
lessness experiment: mean trials to criterion for escape acquisition.

brated Grason-Stadler earphones (Model TDH-39), Procedure

and as we see the tone was generally judged "mod-
erately aversive." The subjects were assigned into 1 of the 12
Soluble cognitive test task, A series of 20 ana- possible groups generated by the four experiments
each with the three pretreatment contingencies. The
grams taken from a list of five-letter anagrams
first study used an instrumental pretreatment (but-
(Tresselt & Mayzner, 1966) was used as the cogni- ton press) followed by a second instrumental task
tive test task. The anagrams were placed indi- (shuttlebox) to test for helplessness. The second
vidually on 4 X 6-in. (10 X 16 cm.) wire-index study used a cognitive task (Levine discrimination
cards and composed of J in. (.64 cm.) letters spaced problems) for the pretreatment followed by the
•fa in. (.48 cm.) apart. Examples of the anagrams instrumental task. The third and fourth studies used
were: (a) I A R D T; (b) B I A T H; (c) an instrumental and cognitive pretreatment task,
U L A T F; (d) E R L K C. The letter order respectively, and followed by the cognitive test task.
for all anagrams was 3-4-2-5-1. This enabled us to There were 24 subjects in each experiment.
look at both the time to solution of each anagram The three groups pretreated with the instrumental
and trials to catching on to the pattern. task were: (a) escapable (E) subjects who received
Communication between subject and experimenter 4$ trials of unsignaled escapable noise; (b) ines-
was conducted with a two-way intercom. Control- capable yoked (E) subjects who received the iden-
ling circuitry was located in an adjacent room sepa- tical 45 trials of unsignaled inescapable loud noise;
rated by a one-way mirror. All response variables and (c) control (C) subjects who passively listened
were measured by (1/100-sec.) times for the instru- to the identical number and duration of tones as E
mental tasks and a (1/100-min.) stopwatch for the and fi subjects. The C group differed from £ group
cognitive tasks. in that C was instructed merely to sit and listen to
316 DONALD S. HIROTO AND MARTIN E. P. SELIGMAN
the tones, while E was instructed to try to terminate to brief samples of the tone. The control groups
the tones (see "Instructions," below). Both C and were told only "From time to time a loud tone
E received identical patterns of tones, since we will come on for awhile. Please sit and listen to it."
deemed it important that the inescapable group The E and E groups were told "From time to time
differ from the control group not in habituation to a loud tone will come on for awhile. When that
tone, but only in perceived inescapability (Geer, tone comes on there is something you can do to
Davison, & Gatchel, 1970; Glass, Singer, & Fried- stop it." The function of the two lights on the
man, 1969). The groups pretreated with the cogni- apparatus was next described, but only to E and E
tive task were: (a) soluble (S) which received three groups.
soluble concept identification problems; (b) insolu-
ble (S) which received three identical concept prob- There are two lights located on this base. The
lems but with no consistent correct values; (c) lights will tell you how the noise on each trial
control (C) which inspected the three problems but was controlled. If you find the way to stop the
like the instrumental control group received no noise then the green light marked 'S-out' will
instructions to solve the problems and no feedback. momentarily flash on after each time you stop
Instructions for instrumental pretreatment. The the loud tone. If you don't stop the tone then
button press-shuttlebox and button press-anagram the red light marked 'time out' will flash when
groups were pretreated with the same instrumental the tone stops. Remember, when the green light
button pressing task. Each subject was initially flashes on this means that you have stopped the
informed that the study involved listening to noise tone. But if the red light flashes this means you
and was given the option of leaving after listening did not stop the tone but that it stopped auto-

20

14
ui
a
< 12
u
in
ui
10
in
u
a
a
I
u.
0
b

4
ui

S.D. 5.79 SO 5.66

BOLUBLB INBOLUBLE

I P<.01 • -P< .02

•P> .05

FIGURE 7. Group means, standard deviations, and significance levels of three

dependent variables for the cognitive pretreatment-instrumental test for help-
lessness experiment: mean number of failures to escape.
 GENERALITY OP LEARNED HELPLESSNESS IN MAN 317

8 -

•
7 -

6 •
0
111
§ 5 •

4 •
<
1U

3 •

2 •

1
-

S.D. 2.19 S.D. 1.49 S.D. 1.09

SOLUBLE INSOLUBLE CoryTHa

-P>.05-

P> .05

FIGURE 8. Group means, standard deviations, and significance levels of three

dependent variables for the cognitive pretreatment-instrumental test for help-
lessness experiment: mean response latency.

matically. Taking the earphones off or disman- terns on it. The sample patterns are composed of
tling the apparatus is not the way to stop the five different dimensions and two values associ-
noise. ated with each dimension. The dimensions and
their values are [experimenter described each
The pretreatments consisted of 45 unsignaled dimension and value]. Each stimulus pattern has
trials with the 90-decibel tone. If a subject in the E one value from each of the five dimensions.
group failed to terminate the tone it lasted for 5
sees. The onset of the red (failure) light was corre- Here the C instructions ceased and the rest of the
lated with the end of the S-sec. interval. The onset instructions were given to S and S groups.
of the green (success) light was correlated with
subject's termination of the tone. The intertrial I have arbitrarily chosen one of the ten values as
interval (ITI) ranged from 10 to 25 sees, with a being correct. For each card I want you to choose
14-sec. mean ITI. Triads of subjects in all groups which side contains this value, and I will then
received identical, yoked durations of tone. At the tell you if your choice was correct or incorrect.
conclusion of the pretreatments subjects rated the In a few trials you can learn what the correct
aversiveness of the tone. value is by this feedback. The object for you is
Instructions for cognitive pretreatment. The Le- to figure out what the answer is so you can
vine discrimination-anagram and Levine discrimina- choose correctly as often as possible.
tion-shuttlebox groups received the following in-
structions to introduce the cognitive task: Five sample trials of a single five-dimension prob-
lem was first presented. This clarified the task of
In this experiment you will be looking at cards finding the "correct" value. The C group was merely
like this one. Each card has two stimulus pat- shown the sample trials without any clarifying in-
318 DONALD S. HIEOTO AND MARTIN E. P. SELIGMAN

structions. All subjects were asked if there were any Instructions to the S and S groups were "We are
questions. No subject asked if the experimental now starting a new problem. You do not know at
problems were soluble or insoluble. this point if I have chosen a different value for this
The experimental stimulus patterns were composed problem. I will continue telling you if you are
of four dimensions. Three different problems were correct or incorrect."
presented in blocks of 10 trials each. At the end of The C group was instructed to "Please continue
each 10-trial problem S and S groups were asked for studying each stimulus pattern carefully and turn
the correct answer. The criterion for acquisition was each card when I ask you to."
subject identifying the correct value after each 10- The S and § groups were allowed 10 sees, to make
trial block. Six subjects were discarded for not a decision before the experimenter warned them that
meeting criterion on all of the three problems and a decision must be made within 5 sees. In other
were subsequently replaced. The S group received words, a trial could never be longer than 15 sees., but
a predetermined schedule of "correct" and "incor- no subject took longer than 10 sees, on any trial.
rect" regardless of what value was guessed. In this Instructions for instrumental test trials. The in-
manner reinforcements were independent and not strumental test trials were conducted at a different
contingent on the S groups responding. The schedule location but within the same experimental room as
of reinforcements were: (a) C-I-I-C-C-I-I-C-C-I the pretreatments. The manipulandum was covered
for the first problem; (b) I-C-I-C-C-I-C-I-C-I for until subject received the following instructions:
the second; and (c) I-C-I-C-I-C-C-I-C-I for the
last problem. In addition, the S group was told You will be given some trials in which a rela-
"that's the wrong answer" when subject tried to tively loud tone will be presented to you. When-
guess the correct value after each problem. ever you hear the tone come on there is some-

16
-

^^•MH

2 14 •
0
a
u 12
H •

II
0
0 10 •

h
I^^HH
U)
< 8 •

a
h
6 1

4
-

2 •

S.D. 8.88 S.O. 14.38 5.D. 9.75

ESCAPE INEBCAPE CONTP.O

P>.05 ' ' P>.05 —-J

-P> .05

FIGURE 9. Group means, standard deviations, and significance levels of three

dependent variables for the instrumental pretreatment-cognitive test for help-
lessness experiment: mean trials to criterion for anagram solution.
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 319

20

8 -

7 •

6 -

5 •

__

0
Z 3 - ^••^^•H

Z
<l
111 2 -

1•

S0 2.45 S0 3.96 S D 3.53

ESCAPE INESCAPE CONTHC

P< .02 • P < 05 •

-P> .05

FIGURE 10. Group means, standard deviations, and significance levels of three
dependent variables for the instrumental pretreatment-cognitive test for help-
lessness experiment: mean number of failures to solve.

thing you can do to stop it. Taking the earphones side before the noise began. A response latency
off or dismantling the apparatus is not the way to under 5 sees, terminated the warning light and
stop the noise. I'll answer all questions and pay avoided the tone. If the subject d'd not terminate
you for your time at the completion of the study. the light (latency less than S sees.) or escape the
Uncover the apparatus and we'll begin. tone (between 5 to 9.99 sees.) a latency of 10 sees,
was given for that trial. At the completion of the
The sliding knob was always located at the mid- test phase subjects rated the unpleasantness of the
point of the manipulandum such that the subject 3,000 hertz tone and completed qusstionnaires relat-
could slide the knob with equal ease to either the ing to the pretreatment and test trials.
left or right end of the box. A 24-V dc warning light Instructions jor cognitive test trials. The cognitive
at the midpoint of the manipulandum cover was on test was located in the same location as the instru-
for S sees, before and terminated when the S-sec. mental test task. The following instructions intro-
tone began. The test phase consisted of 20 signaled duced the trials:
10-sec. trials with the IT1 ranging from 10 to 45 You will be asked to solve some anagrams. As
sees, and a mean ITI of 21 sees. you know anagrams are words with the letters
The appropriate response was moving the knob scrambled. The problem for you is to unscramble
to one side of the manipulation to throw the micro- the letters so they form a word. When you've
switch controlling the stimulus light or noise. On found the word tell me what it is over the inter-
the next trial, moving to the opposite side escaped com system. Now [subject's name], there could be
or avoided the noise. The instructions specified a pattern or principle by which to solve the ana-
escape contingencies only, but an avoidance response grams. But that's up to you to figure out. I
was possible by sliding the knob to the appropriate can't answer any questions now. After the cxperi-
320 DONALD S. HIEOTO AND MARTIN E. P. SELIGMAN

ment is over I'll answer all questions and pay mental measures were: (a) trials to criterion for
you for your time. escape acquisition, defined as subject completing
All 20 anagrams were soluble and had the same three consecutive escape responses; (b) number of
letter sequence. The anagrams could be solved indi- failures to escape, defined as the number of trials
vidually; but the easiest method was to learn to use with latencies of 10 sees.; and (c) the mean latency
the letter sequence. The anagrams were selected such for the 20 trials. Three analogous measures were
that only one word could be arranged with each analyzed for the cognitive test task: (a) trials to
anagram. But there were two instances in which criterion for anagram solution was defined as sub-
subjects found additional words. The anagram ject solving three consecutive anagrams in less than
B L O E N ("noble") was rearranged as "Nobel" 15 sees. each. (Reaching this criterion meant that
by three subjects, but experimenter disqualified this subject recognized the principle of fixed-letter se-
answer since it was a proper name. On these occa- quence. The definition, although arbitrary, was
sions subject was asked to try again, and each sub- highly reliable. When subject "caught on" to the
ject eventually found the acceptable word. The anagram construction, latencies dropped dramatically
anagram U N A T J ("jaunt") was seen as "junta" from an average of 45 sees, to well below 15 sees.) ;
by three different subjects and scored as acceptable (b) number of failures to solve, defined as the
answers (experimenter was tempted to award these number of trials with latencies of 100 sees., the point
subjects a medal). In cases in which subject gave a at which the trial ended; (c) mean response latency
nonsense word experimenter replied, "that's not a for the 20 anagrams. The last two measures for the
word, please try again." instrumental and cognitive tests parallel the indices
Three dependent variables were analyzed on the reported in the human and animal learned helpless-
instrumental and cognitive test tasks. The instru- ness literature.

IOC

50

40
U
III
Ifl

U
30
III

J
20
III

10

BO 15.31 S.D. 21.S3 S.D. 20.64

ESCAPE INEBCAPE CONTROL

|—— P< .02- -P<.05

•P> .05
FIGURE 11. Group means, standard deviations, and significance levels of three
dependent variables for the instrumental prelreatment-cognitive test for help-
lessness experiment: mean response latency.
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 321

20

18

16

14
g
\5 12
u
o 10
n
5

Ul

P> .05

FIGURE 12. Group means, standard deviations, and significance levels of three
dependent variables for the cognitive pretreatment-cognitive test for helplessness
experiment: mean trials to criterion for anagram solution.

RESULTS differences appeared in solution of the Levine

Overall, interference was produced by in- discrimination pretreatment between the solu-
escapability and insolubility in three of the ble group followed by the cognitive test and
four experiments: Inst.-Inst.; Inst.-Cog., the soluble group followed by the instru-
Cog.-Inst. The Cog.-Cog. experiment did not mental test (all ps > .OS). All subjects in the
show significant effects although the results escapable and soluble pretreatments reached
were in the predicted direction. the criteria of learning in pretreatment.
A check on the pretreatment manipulation
Pretreatment Trials was conducted to assess the effectiveness of
All subjects who received the escapable or the inescapability and insolubility manipula-
soluble pretreatments performed appropri- tions. The subjects in the helplessness-induc-
ately to the relevant contingencies during tion groups (i.e., E and S) believed they had
pretreatment. No significant differences in the no control over solution. Two items from a
number of trials to escape the tone by button questionnaire relating to controllability in the
pressing were found between the button-press pretreatment provided the relevant informa-
escapable group followed by the instrumental tion. The items were worded to determine if
task and the button-press escapable group subject attributed uncontrollability to himself
followed by the cognitive test. Likewise, no or to the task. Question 1 asked if subject
322 DONALD S. HIROTO AND MARTIN E. P. SELIGMAN

believed "you couldn't solve the problem" tion and S groups rated the question at 1.56.
while the second asked if subject believed the The difference between the pooled E and S
"problem was unsoluble—that it couldn't be versus pooled E and S groups was significant,
solved." Answers were recorded on a 7-point F ( 1 , 6 2 ) = 5.13, p< .001.
scale with higher scores denoting greater in-
solubility. All E and S groups believed they Helplessness Test Trials 1
lacked the ability to solve their respective Inst.-Inst. The group pretreated with ines-
tasks relative to E and S groups. Subjects in capable tone in button pressing escaped sig-
the E groups had a mean rating of 4.4 on nificantly more poorly in the shuttlebox than
Question 1 and the S groups a rating of 4.9. the escapable and control pretreated groups
This compares with a rating of 2.25 for the on all measures. Figures 3, 4, and 5 present
E groups. The difference between the pooled these results and significance levels for trials
E and S groups versus pooled E and S groups 1
was significant, F ( 1 , 6 2 ) = 43.54, p < .001. Since it was predicted that groups pretreated
with inescapability (insolubility) would demon-
In addition, the E and S groups believed the strate an interference to learning relative to groups
problem was unsolvable. The mean rating for without such pretreatments, the following statistical
tests were one-tailed: E versus E, E versus C, S
E groups on Question 2 was 4.63 while the versus S, and S versus C; tests between escapable
mean rating for S groups was 4.31; the E (soluble) and control pretreated groups were two-
groups had a rating of 2.37 on the same ques- tailed: E versus C and S versus C.

201

7
1E «
Ul *
•

•^^'

5 -

! •

i ».
f 2

S.D. 3.54 B.D. 5.67 8.D. 3.11

SOLUBLK INBQLUBLK CONTHOI

•P> .05

FIGURE 13. Group means, standard deviations, and significance levels of three
dependent variables for the cognitive pretreatment-cognitive test for helplessness
experiment: mean number of failures to solve.
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 323

100

40 mm^mm

u —

HI
> 30 •

S ^^^*m

5 20
\
Ul
•

10

8.D. 20.56 8.D. 28.82 8.D. 18.77

SOLUBLE INSOLUBLE CONTROL

P>.05 1 1 P>.05

FIGURE 14. Group means, standard deviations, and significance levels of three
dependent variables for the cognitive pretreatment-cognitive test for helplessness
experiment: mean response latency.

to escape criterion, number of failures to es- soluble, soluble, or control discrimination

cape, and mean latency. problems. Figures 12, 13, and 14 present the
Cog.-Inst. The group pretreated with in- trials to criterion for anagram solution, num-
soluble discrimination problems did signifi- ber of failures to solve, mean latency, and
cantly worse on shuttlebox escape than the significance levels for anagram solution.
soluble and control pretreated groups in the Questionnaire scores. The subjects pre-
shuttlebox. Figures 6, 7, and 8 present these treated with inescapable tone were yoked to
results and significance levels for trials to subjects pretreated with escapable tone such
escape criterion, number of failures to escape, that density, frequency, and duration of tone
and mean latency.
were equated. It is possible that differential
Inst.-Cog. The group pretreated with ines-
subjective aversiveness of tone have produced
capable tone in button pressing generally was
worse at solving anagrams than the soluble the subsequent retarded performance. How-
or control pretreated groups. Figures 9, 10, ever, all subjects rated the pretreatment tone
and 11 present these results and significance as moderately unpleasant on a 7-point scale.
levels for trials to criterion for anagram solu- The pretreated E groups rated the tone at
tion, number of failures to solve, and mean 3.91, while E groups rated the stimulus at
latency. 3.70. The C groups who passively listened to
Cog.-Cog. No significant differences were the 3,000 hertz tone rated the same stimulus
found on anagram performance following in- somewhat lower at 3.14. These differences
324 DONALD S. HIROTO AND MARTIN E. P. SELIGMAN

were not significant, F (2,45) = 1.02, p> problems rather than three. The subjects, ap-
.05. paratus, and procedures were identical to the
Additional post hoc comparisons on the Cog.-Cog. experiment reported above, except
questionnaires were conducted. There were no that the S group (n = 8) received four 10-
differences in "amount of trying" between trial block insoluble problems, the S group
pretreated groups in either the pretreatment (n = 8) received four soluble problems, and
or test trials (p > .10), nor did they change the C group (n = 8) looked passively at four
in amount of trying from pretreatment to 10-trial block problems without attempting
test trials (p > .10). All subjects were asked solution.
if they felt "frustrated" during any part of When four insoluble problems were used
the experiment. The E and S groups rated the interference with anagram solution was found.
pretreatment as more frustrating than the Figures 15, 16, and 17 present the trials to
pooled E and S groups (4.5 vs. 2.2, respec- solution criterion, number of failures to solve,
tively), F (1,62) = 42.3, p < .001. Within and mean latency to solution as well as sig-
the helplessness-induction groups two S pre- nificance levels for anagram performance of
treated groups indicated greater frustration the three pretreated groups.
than the two E pretreated groups (5.3 vs. 3.6,
respectively), F (1, 62) = 6.96, p < .018. DISCUSSION
Since only the cognitive pretreated group There were three main findings of the
tested with anagrams failed to show helpless- experiment: (a) a group pretreated with an
ness, we repeated the Cog.-Cog. procedure inescapable aversive tone showed greatly de-
with four insoluble Levine discrimination bilitated tone-escape performance in a shut-

20 •

18 • MBIBM

16 -
0
5 14
u •

a.
u l*i
12 •

0
10 •
^H^H

n
ia, 8 •

6 •

Ul
4 •

2 •

S.D. 6.3 S.D. 7.2 s.D. 6. 5

SOLIJBLE INSOLUBLE COIMT POU

FIGURE IS. Group means, standard deviations, and significance levels of three
dependent variables for the revised cognitive pretreatment-cognitive test for
helplessness experiment: mean trials to criterion for anagram solution.
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 325

20
^
/

§
0
0)
10

§ 6
u.
0

-P> .05

FIGURE 16. Group means, standard deviations, and significance levels of three
dependent variables for the revised cognitive pretreatment-cognitive test for
helplessness experiment: mean number of failures to solve.

tlebox; a control group and an escapable pre- capability. In addition, a group pretreated
treated group performed well in shuttlebox with inescapable tone was debilitated at ana-
escape. This replicates the findings of Hiroto gram solution to the same extent as the group
(1974) and again demonstrates a learned pretreated with four insoluble discrimination
helplessness phenomenon in man, directly problems.
parallel to learned helplessness in dogs, cats, The finding of cross-modal helplessness is
and rats (cf. Seligman, in press b, for a of considerable theoretical interest. Initially,
cross-species review), (b) A group pretreated critics of the animal helplessness findings
with four insoluble discrimination problems argued that failure to escape in a shuttlebox
was debilitated at solving later anagrams following inescapable shock might have re-
relative to a control and soluble pretreated sulted from a competing motor response,
groups. This demonstrates that learned help- rather than a more general "organismic" de-
lessness can be produced within cognitive bilitation (e.g., Miller & Weiss, 1969). Maier
tasks, without aversive unconditioned stimuli (1970) and Seligman and Maier (1967) dis-
or instrumental components, (c) Cross-modal confirmed this peripheral interpretation by
helplessness was also found. A group pre- using very different training and testing
treated with insoluble cognitive problems was situations. These situations, however, were
debilitated at instrumental escape. Interest- similar at least in the fact that pretreatment
ingly, this group was just as debilitated as and testing both involved the same uncondi-
the group pretreated with instrumental ines- tioned stimulus-shock-and instrumental re-
326 DONALD S. HIKOTO AND MARTIN E. P. SELIGMAN

100

40
0
I
30

20

10

-P> .05

FIGURE 17. Group means, standard deviations, and significance levels of three
dependent variables for the revised cognitive pretreatment-cognitive test for
helplessness experiment: mean response latency.

spending. Our present findings provide the differentiate from learned helplessness, which
coup de grace against any peripheralist in- is not post hoc (e.g., Schmeck & Clements,
terpretation. No competing motor response 1971; Schmeck & Ribich, 1969). Since the
could generate cognitive interference from groups all reported similar levels of aversive-
instrumental inescapability or instrumental ness to the tone, this variable does not ex-
interference from cognitive insolubility. plain the findings. Finally, during our de-
Alternatively, neither frustration, differen- briefing no subject gave evidence of having
tial aversion of tone, nor demand character- detected the purpose of the study or what
istics (Orne, 1962) seem able to account for was predicted of him.
these differences. Frustration was experienced One limitation on the generality of these
to a greater extent in the inescapable and effects should be mentioned. The subjects
insoluble pretreatments than in the soluble, clearly perceived both tasks, as different as
escapable, or control pretreatment groups. they are, as part of the same experiment. We
Frustration is usually interpreted as energiz- do not know whether any learned helpless-
ing, rather than deenergizing, motivation ness was carried out of the laboratory. Fu-
(e.g., Amsel & Roussel, 1952). One can ture research with unobtrusive tests should
choose to interpret frustration post hoc as determine the extraexperimental generality of
rigidifying rather than energizing responses, our effects.
but then the explanation becomes difficult to What is the process of learned helplessness,
 GENERALITY OF LEARNED HELPLESSNESS IN MAN 327

then? If the interference produced was both Maier, S. F., Seligman, M. E. P., & Solomon, R. L.
short-lived and highly specific to the condi- Pavlovian fear conditioning and learned helpless-
ness. In B. A. Campbell & R. M. Church (Eds.),
tions of training, we might call it a transient Punishment. New York: Appleton-Century-Crofts,
and peripheral "state." Since it is general 1969.
across motivations and tasks, however, we Miller, N. E., & Weiss, J. M. Effects of the somatic
suggest that the process induced by uncon- or visceral responses to punishment. In B. A.
trollability may be the rudiment of a "trait." Campbell & R. M. Church (Eds.), Punishment.
New York: Appleton-Century-Crofts, 1969.
Further research on the long-livedness of this Miller, W. R., & Seligman, M. E. P. Depression and
process in humans may bear this out. At any the perception of reinforcement. Journal of Ab-
rate the expectancy that responding and rein- normal Psychology, 1973, 82, 62-73.
forcement are independent generalizes widely. Orne, M. T. On the social psychology of the psycho-
Seligman (1973, in press a) and Seligman logical experiment: With particular references to
demand characteristics and their implications.
et al. (in press) proposed that learned help- American Psychologist, 1962, 17, 776-783.
lessness is a model for depression in man. Overmier, J. B., & Seligman, M. E. P. Effects of
This speculation entails the consequence that inescapable shock upon subsequent escape and
our procedure for producing debilitation avoidance responding. Journal of Comparative
should also produce those symptoms associ- and Physiological Psychology, 1967, 63, 28-33.
ated with mild depression: self devaluation, Rotter, J. B. Generalized expectancies for internal
versus external control of reinforcement. Psycho-
psychomotor retardation, feelings of sadness, logical Monographs, 1966, 50(1, Whole No. 609).
etc. Future research should answer these Schmeck, R. R., & Clements, P. Resistance to set
questions. breaking as a function of frustrative error making.
In conclusion, we have produced learned Psychonomic Science, 1971, 24, 297-298.
helplessness in humans. The process engen- Schmeck, R. R., & Ribich, F. Rigidity as a func-
tion of task complexity. Psychonomic Science,
dered debilitates performance well beyond 1969, 17, 323.
the condition under which helplessness is first Seligman, M. E. P. Fall into helplessness. Psychol-
trained. We suggest therefore that learned ogy Today, 1973, 7, 43-48.
helplessness may involve a trait-like system Seligman, M. E. P. Depression and learned help-
of expectancies that responding is futile. lessness. In R. J. Friedman & M. M. Katz (Eds.),
The psychology of depression: Contemporary
REFERENCES theory and research, in press, (a)
Seligman, M. E. P. Helplessness. San Francisco:
Amsel, A., & Roussel, J. Motivational properties of Freeman, in press, (b)
frustration: I. Effect on a running response of the
Seligman, M. E. P., Klein, D., & Miller, W. R.
addition of frustration to the motivational com- Depression. In H. Leitenberg (Ed.), Handbook
plex. Journal of Experimental Psychology, 1952,
of behavior therapy. New York: Appleton-Cen-
43, 363-368. tury-Crofts, in press.
Geer, J. H., Davison, G. C., & Gatchel, R. L. Re-
duction of stress in humans through nonveridical Seligman, M. E. P., & Maier, S. F. Failure to es-
perceived control of aversive stimulation. Journal cape traumatic shock. Journal of Experimental
of Personality and Social Psychology, 1970, 16, Psychology, 1967, 74, 1-9.
731-738. Seligman, M. E. P., Maier, S. F., & Solomon, R. L.
Glass, D. C., Singer, J, E., & Friedman, L. C. Psy- Unpredictable and uncontrollable aversive events.
chic cost of adaptation to an environmental stres- In F. R. Brush (Ed.), Aversive conditioning and
sor. Journal of Personality and Social Psychology, learning. New York: Academic Press, 1971.
1969, 12, 200-210. Thornton, J. W., & Jacobs, P. D. Learned help-
Hiroto, D. S. Locus of control and learned helpless- lessness in human subjects. Journal of Experi-
ness. Journal of Experimental Psychology, 1974, mental Psychology, 1971, 87, 367-372.
102, 187-193. Tresselt, M. E., & Mayzner, M. S. Normative solu-
Levine, M. Hypothesis behavior by humans during tion times for a sample of 134 solution words and
discrimination learning. Journal of Experimental 378 associated anagrams. Psychonomic Monograph
Psychology, 1966, 71, 331-338. Supplements, 1966, 1, 293-298.
Levine, M. Hypothesis theory and nonlearning de- Turner, L., & Solomon, R. L. Human traumatic
spite ideal S-R reinforcement contingencies. avoidance learning: Theory and experiments on
Psychological Review, 1971, 78, 130-140. the operant-respondent distinction and failures to
Maier, S. F. Failure to escape traumatic shock: learn. Psychological Monographs, 1962, 76(1,
Incompatible skeletal motor responses or learned Whole No. 559).
helplessness? Learning and Motivation, 1970, 1,
157-170. (Received July 31, 1973)