Journal of Abnormal Psychology 1978, Vol. 87, No. 1, 49-74

Learned Helplessness in Humans: Critique and Reformulation

Lyn Y. Abramson and Martin E. P. Seligman University of Pennsylvania John D. Teasdale Oxford University, England

The learned helplessness hypothesis is criticized and reformulated. The old hypothesis, when applied to learned helplessness in humans, has two major problems: (a) It does not distinguish between cases in which outcomes are uncontrollable for all people and cases in which they are uncontrollable only - for some people (univervsal vs. personal helplessness), and (b) it does not explain when helplessness is general and when specific, or when chronic and when acute. A reformulation based on a revision of attribution theory is pro- posed to resolve these inadequacies. According to the reformulation, once people perceive noncontingency, they attribute their helplessness to a cause. This cause can be stable or unstable, global or specific, and internal or external. The attribution chosen influences whether expectation of future helplessness will be chronic or acute, broad or narrow, and whether helplessness will lower self-esteem or not. The implications of this reformulation of human helplessness for the learned helplessness model of depression are outlined.

Over the past 10 years a large number of 1967). Paralleling the experimental findings experiments have shown that a variety of orga- with dogs, the debilitating consequences of nisms exposed to uncontrollable events often uncontrollable events have been demonstrated exhibit subsequent disruption of behavior (see in cats (Masserman, 1971; Seward & Hum- Maier & Seligman, 1976, for a review of the phrey, 1967; Thomas & Dewald, 1977), in infrahuman literature). For example, whereas fish (Frumkin & Brookshire, 1969; Padilla, naive dogs efficiently learn to escape shock by 1973; Padilla, Padilla, Ketterer, & Giacolone, jumping over a barrier in a shuttle box, dogs 1970), and in rats (Maier, Albin, & Testa, that first received shocks they could neither 1973; Maier & Testa, 197S; Seligman & avoid nor escape show marked deficits in ac- Beagley, 1975; Seligman, Rosellini, & Kozak, quisition of a shuttle escape response (Over- 197S). Finally, the effects of uncontrollable mier & Seligman, 1967; Seligman & Maier, events have been examined in humans (Fosco & Geer, 1971; Gatchel & Proctor, 1976; Glass & Singer, 1972; Hiroto, 1974; Hiroto & Selig- This work was supported by U.S. Public Health man, 1975; Klein, Fencil-Morse, & Seligman, Service Grant MH-19604, National Science Founda- 1976; Klein & Seligman, 1976; Krantz, Glass, tion Grant SOC-74 12063, and a Guggenheim fel- & Snyder, 1974; Miller & Seligman, 1975; lowship to Martin Seligman. We thank Lauren Al- Racinskas, 1971; Rodin, 1976; Roth, 1973; loy, Judy Garber, Suzanne Miller, Frank Irwin, S. J. Rachman, and Paul Eelen for their critical com- Roth & Bootzin, 1974; Roth & Kubal, 1975; ments on earlier drafts of this paper. Thornton & Jacobs, 1971; among others). Ivan Miller (Note 1) has proposed an almost Hiroto's experiment (1974) is representative identical reformulation. We believe this work to have been done independently of ours, and it should and provides a human analogue to the animal be so treated. studies. College student volunteers were as- Requests for reprints should be sent to Martin E. signed to one of three groups. In the con- P. Seligman, Department of Psychology, University of Pennsylvania, 3815 Walnut Street, Philadelphia, trollable noise group, subjects received loud Pennsylvania 19174. noise that they could terminate by pushing

Copyright 1978 by the American Psychological Association, Inc. All rights of reproduction in any form reserved.

49 50 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE a button four times. Subjects assigned to the Historically, the learned helplessness hy- uncontrollable noise group received noise that pothesis was formulated before helplessness terminated independently of subjects' respond- experiments were performed with human sub- ing. Finally, a third group received no noise. jects. In the main, early studies of human In the second phase of the experiment all helplessness attempted to reproduce the ani- groups were tested on a hand shuttle box. In mal findings in humans and were rather less the shuttle box, noise termination was con- concerned with theory building. Recently, trollable for all subjects; to turn off the noise, however, investigators of human helplessness subjects merely had to move a lever from one (e.g., Blaney, 1977; Golin & Terrell, 1977; side of the box to the other. The results of the Wortman & Brehm, 1975; Roth & Kilpatrick- test phase were strikingly similar to those ob- Tabak, Note 2) have become increasingly dis- tained with animals. The group receiving prior enchanted with the adequacy of theoretical controllable noise as well as the group receiv- constructs originating in animal helplessness ing no noise readily learned to shuttle, but the for understanding helplessness in humans. typical subject in the group receiving prior And so have we. We now present an attribu- uncontrollable noise failed to escape and tional framework that resolves several the- listened passively to the noise. oretical controversies about the effects of un- Although a number of alternative hypothe- controllability in humans. We do not know ses (see Maier & Seligman, 1976, for a re- whether these considerations apply to infra- view) have been proposed to account for the humans. In brief, we argue that when a per- debilitating effects of experience with uncon- son finds that he is helpless, he asks why he is trollability, only the learned helplessness hy- helpless. The causal attribution he makes then pothesis (Maier & Seligman, 1976; Maier, determines the generality and chronicity of Seligman, & Solomon, 1969; Seligman, 1975; his helplessness deficits as well as his later Seligman et al., 1971) provides a unified the- self-esteem. In developing the attributional oretical framework integrating the animal and framework, we find it necessary to refine at- human data. The cornerstone of the hypothe- tribution theory (cf. Heider, 1958; Weiner, sis is that learning that outcomes are uncon- 1972, 1974). Finally, we discuss the implica- trollable results in three deficits: motivational, tions of the reformulation for the helplessness cognitive and emotional. The hypothesis is model of depression (Seligman, 1972, 1975; "cognitive" in that it postulates that mere ex- Seligman, Klein, & Miller, 1976). posure to uncontrollability is not sufficient to render an organism helpless; rather, the orga- Personal Helplessness Versus Universal nism must come to expect that outcomes are Helplessness uncontrollable in order to exhibit helplessness. Inadequacy 1 of the Old Theory In brief, the motivational deficit consists of retarded initiation of voluntary responses and Several examples highlight a conceptual is seen as a consequence of the expectation problem encountered by the existing learned that outcomes are uncontrollable. If the orga- helplessness hypothesis when applied to hu- nism expects that its responses will not affect man helplessness. Consider a subject in Hi- some outcome, then the likelihood of emitting roto's experiment (1974) who is assigned to such responses decreases. Second, the learned the group that received uncontrollable noise. helplessness hypothesis argues that learning The experimenter tells the subject there is something he can do to turn off the noise. that an outcome is uncontrollable results in a Since the noise is actually uncontrollable, the cognitive deficit since such learning makes it subject is unable to find a way to turn off the difficult to later learn that responses produce noise. After repeated unsuccessful attempts, that outcome. Finally, the learned helpless- the subject may come to believe the problem ness hypothesis claims that depressed affect is is unsolvable; that is, neither he nor any other a consequence of learning that outcomes are subject can control noise termination. Alterna- uncontrollable. tively, the subject may believe that the prob- CRITIQUE AND REFORMULATION 51 lem is solvable but that he lacks the ability jects perceived tasks of skill as if they were r to solve it; that is, although he can't control tasks of chance. A puzzling finding, however, noise termination, other subjects could suc- was consistently obtained in these studies. On cessfully control the noise. The old helpless- postexperimental questionnaires, helpless and ness hypothesis does not distinguish these two nonhelpless subjects rated skill as playing the states, either of which could be engendered by same large role in a person's performance on the procedure of presenting uncontrollable the skill task. Both helpless and nonhelpless outcomes. subjects said they viewed the skill task as a In a recent publication, Bandura (1977) skill task. Thus, the relation between the con- discussed a similar distinction: cepts of external control and uncontrollability may be more complex than implied by the old Theorizing and experimentation on learned helpless- hypothesis. ness might well consider the conceptual distinction between efficacy and outcome expectations. People Taken together, these examples point to can give up trying because they lack a sense of ef- one conceptual problem concerning the no- ficacy in achieving the required behavior, or they tions of uncontrollability and helplessness. Re- may be assured of their capabilities but give up try- call the distinction made by the old helpless- ing because they expect their behavior to have no ness hypothesis between controllable and effect on an unresponsive environment or to be con- sistently punished. These two separable expectancy uncontrollable outcomes. An outcome is said sources of futility have quite different antecedents to be uncontrollable for an individual when and remedial implications. To alter efficacy-based the occurrence of the outcome is not related futility requires development of competencies and to his responding. That is, if the probability expectations of personal effectiveness. By contrast, to change outcome-based futility necessitates changes in of an outcome is the same whether or not a prevailing environmental contingencies that restore given response occurs, then the outcome is in- the instrumental value of the competencies that peo- dependent of that response. When this is ple already possess, (pp. 204-205) true of all voluntary responses, the outcome is said to be uncontrollable for the individual A final way of illustrating this inadequacy (Seligman, 1975; Seligman, Maier, & Solo- concerns the relation between helplessness and mon, 1971). Conversely, if the probability of external locus of control. Early perspectives of the outcome when some response is made is learned helplessness (Hiroto, 1974; Miller & different from the probability of the outcome Seligman, 1973; Seligman, Maier, & Geer, when the response is not made, then the out- 1968) emphasized an apparent similarity be- come is dependent on that response: The out- tween the helplessness concept of learning come is controllable. The early definition, that outcomes are uncontrollable and Rotter's then, makes no distinction between cases in (1966) concept of external control. Rotter ar- which an individual lacks requisite controlling gued that people's beliefs about causality can responses that are available to other people be arrayed along the dimension of locus of and cases in which the individual as well as control, with "internals" tending to believe all other individuals do not possess controlling outcomes are caused by their own responding responses. These three examples all illustrate and "externals" tending to believe outcomes the same inadequacy. In the next section we are not caused by their own responding but outline a framework that resolves this inade- by luck, chance, or fate. Support for this pro- quacy, and we discuss the implications of this posed conceptual similarity of externals and framework. helpless individuals was provided by studies of verbalized expectancies for success in tasks Resolution of Inadequacy 1 of skill (Klein & Seligman, 1976; Miller & Seligman, 1975). Helpless subjects gave smallj Suppose a child contracts leukemia and the expectancy changes, which suggests a belief1 father bends all his resources to save the in external control, whereas subjects not child's life. Nothing he does, however, im- made helpless gave large expectancy changes, proves the child's health. Eventually he comes which suggests a belief in internal control. to believe there is nothing he can do. Nor is These findings indicated that helpless sub- there anything anyone else can do since leu- 52 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE

future responses would be futile in obtaining Objective noncontingency —» Perception of the outcome. The usual sequence of events present and past noncontingency —> Attribu- leading from objective noncontingency to the tion for present or past noncontingency ~* Expectation of future noncontingency —> helplessness is diagrammed in Figure 1. Symptoms of helplessness. Both the old and reformulated hypotheses hold the expectation of noncontingency to be the crucial determinant of the symptoms of Figure 1. Flow of events leading to symptoms of learned helplessness. Objective noncontingency helplessness. is predicted to lead to symptoms of helpless- kemia is incurable. He subsequently gives up ness only if the expectation of noncontingency trying to save the child's life and exhibits is present (Seligman, 1975, pp. 47-48). The signs of behavioral helplessness as well as de- old model, however, was vague in specifying pressed affect. This example fits the specifica- the conditions under which a perception that tions of the old learned helplessness hypothe- events are noncontingent (past or present ori- sis. The parent believed the course of the ented) was transformed into an expectation child's disease was independent of all of his that events will be noncontingent (future responses as well as the responses of other oriented). Our reformulation regards the at- people. We term this situation universal help- tribution the individual makes for noncontin- lessness. gency between his acts and outcomes in the Suppose a person tries very hard in school. here and now as a determinant of his subse- He studies endlessly, takes remedial courses, quent expectations for future noncontingency. hires tutors. But he fails anyway. The person These expectations, in turn, determine the comes to believe he is stupid and gives up generality, chronicity, and type of his help- trying to pass. This is not a clear case of un- lessness symptoms. In the context of this gen- controllability according to the old model, eral account of the role of attribution in the since the person believed there existed re- production of symptoms, the distinction be- sponses that would contingently produce pass- tween universal and personal helplessness can ing grades although he did not possess them. now be clarified. Regardless of any voluntary response the per- Table 1 explicates the distinction between son made, however, the probability of his ob- universal helplessness and personal helpless- taining good grades was not altered. We term ness and ultimately serves to define our usage this situation personal helplessness. of the attributional dimension of internality. Before discussing the distinction between We take the self-other dichotomy as the cri- terion of internality. When people believe universal and personal helplessness, it is use- ful to spell out the flow of events leading to that outcomes are more likely or less likely symptoms of helplessness in both examples. to happen to themselves than to relevant First, the person perceived that all of his acts others, they attribute these outcomes to in- were noncontingently related to the desired ternal factors. Alternatively, persons make ex- outcome; regardless of what the father did, ternal attribution for outcomes that they be- the child's illness did not improve, and the lieve are as likely to happen to themselves as to relevant others. student continued to do poorly no matter how hard he tried. The person then made an at- In the table, the x-axis represents the per- tribution for the perceived noncontingency son's expectations about the relation between the desired outcome and the responses in his between his acts and the outcome; the father 1 came to believe leukemia was incurable and repertoire. The person expects the outcome the student came to believe he was stupid. In each case, the attribution led to an expec- 1 For the purpose of exposition, dichotomies tancy of noncontingency between future acts rather than continua are used. The person expects of the individual and the outcome. Finally, that the controlling response is or is not available to him and that the controlling response is or is not the symptoms of helplessness were a conse- available to others. These two dichotomies allow for quence of the person's expectancy that his four possible belief states. Strictly speaking, how- CRITIQUE AND REFORMULATION S3

Table 1 Personal Helplessness and Universal Helplessness

Self

The person expects the The person expects the outcome is contingent outcome is not contingent on a response in his on any response in his Other repertoire. repertoire.

The person expects the outcome is contingent personal helplessness on a response in the 3 repertoire of a relevant (internal attribution) other. The person expects the outcome is not contingent universal helplessness on a response in the 4 repertoire of any relevant (external attribution) other. either to be contingent on some response in ample makes it clear why we use a "relevant his repertoire or not to be contingent on any other" rather than a "random other" or "any response in his repertoire. The y-axis repre- other": It is of no solace to a floundering sents his expectations about the relation be- graduate student in mathematics that "ran- tween the desired outcome and the responses dom others" are unable to do topological in the repertoires of relevant others. The per- transformations. Crucial to the student's self- son expects the outcome to be either con- evaluation is his belief that his peers, "rele- tingent on at least one response in at least one vant others," have a high probability of relevant other's repertoire or not contingent on being able to do topological transformations. any response in any relevant other's reper- Nor is it self-esteem damaging for a grade toire. Cell 4 represents the universal helpless- school student to fail to solve mathematical ness case and includes the leukemia example, problems that only professional mathemati- and Cell 3 represents the personal helplessness cians can solve, although he may have low case and includes the school failure example. self-esteem if his peers can solve them. There- Because the person does not believe he is fore, the y-axis is best viewed as representing helpless in Cells 1 and 2, these cells are not the person's expectations about the relation relevant here and are not discussed. It should between the desired outcome and the re- be pointed out, however, that a person in Cell sponses in the repertoires of relevant others.2 2 would be more likely to make an internal attribution for his perceived control than would a person in Cell 1. 2 Our formulation of "internal" and "external" In Table 1, the y-axis represents the per- attributions resembles other attributional frame- works. Heider (19S8), who is generally considered son's expectations about whether someone the founder of attribution theory, made a basic dis- else, a relevant other, had the controlling re- tinction between "factors within the person" and sponse in his repertoire. The following ex- "factors within the environment" as perceived de- terminants of outcomes. Similarly, in the locus of control literature, Rotter (1966) distinguished be- ever, the x-axis is a continuum. At the far right, the tween outcomes that subjects perceive as causally person expects there is a zero probability that the related to their own responses and personal char- desired outcome is contingent on any response in acteristics and outcomes that subjects perceive as his repertoire. Conversely, on the far left he expects caused by external forces such as fate. Unlike these there is a probability of one that the desired out- previous formulations that ask whether a factor come is contingent on a response in his repertoire. resides "within the skin" or "outside the skin" to Similar considerations apply to the y-axis as a con- determine whether it is internal or external, we de- tinuum. fine the self-other dichotomy as the criterion of 54 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE

Implications tasks as skill tasks, not as chance, is no longer puzzling. The task is one of skill (relevant The distinction between universal and per- others can solve it), but they do not have the sonal helplessness resolves the set of inade- relevant skill. These subjects view them- quacies with which we began the article, Sit- selves as personally rather than universally uations in which subjects believe they cannot helpless. solve solvable problems are instances of per- The distinction between universal helpless- sonal helplessness according to the reformu- ness and personal helplessness also clarifies lated hypothesis. Alternatively, situations in the relation of uncontrollability to failure. In which subjects believe that neither they nor the literature these two terms have often been relevant others can solve the problem are in- used synonymously. Tennen (Note 3), argu- stances of universal helplessness. Similarly, ing from an attributional stance, suggested Bandura's (1977) conceptual distinction be- that the terms are redundant and that we tween efficacy and outcome expectancies re- abandon the concept of uncontrollability for lates to the reformulation in the following the simpler concept of failure. We believe way: Personal helplessness entails a low ef- this suggestion is misguided both from the ficacy expectation coupled with a high out- point of view of attribution theory and from come expectation (the response producing the common usage of the term failure. outcome is unavailable to the person), whereas In current attribution theories (e.g., Weiner, universal helplessness entails a low outcome 1972) success and failure refer to outcomes. expectation (no response produces the out- Success refers to obtaining a desired outcome come). Finally, the reformulation regards "ex- and failure to not obtaining a desired outcome. ternal locus of control" and "helplessness" as According to this framework, then, the term orthogonal. One can be either internally or failure does not embrace all cases of uncon- externally helpless. Universally helpless in- trollability. Thus, from a strict attributional dividuals make external attributions for fail- point of view, failure and uncontrollability ures, whereas personally helpless individuals are not synonymous: Failure is a subset make internal attributions. The experimental of uncontrollability involving bad outcomes. finding that helpless individuals view skill Early theoretical accounts of helplessness sug- gested that good things received independently internality. Although these two formulations may of responding should lead to helplessness defi- appear to be at odds, analysis reveals strong simi- cits. Recent evidence bears this out: Uncon- larities. For example, Heider (1958) argued that in trollable positive events produce the motiva- making a causal attribution, individuals hold a con- dition responsible for an outcome if that condition is tional and cognitive deficits in animals (Good- present when the outcome is present and absent when kin, 1976; Welker, 1976) and in humans the outcome is absent. Likewise, Kelley (1967) sug- (Griffiths, 1977; Eisenberger, Mauriello, Carl- gested that the procedure individuals use in deter- son, Frank, & Park, Note 4; Hirsch, Note 5; mining the cause of events is similar to an analysis Nugent, Note 6; but see Benson & Kennelly, of variance procedure employed by scientists. The factor that consistently covaries with an outcome is 1976, for contrary evidence) but probably do considered to be its cause. not produce sad affect. Similarly Cohen, Roth- Let us examine the leukemic child and the school bart, and Phillips (1976) produced helpless- failure examples from the perspective of Kelley and ness effects in the absence of perceived failure. Heider. The responses of no person are consistently associated with improvement of the leukemic child's In the future, such studies should measure disease. If the father performed Kelley or Heider's perception of noncontingency as well as per- causal analysis, he would conclude that his failure formance, since Alloy and Abramson (Note was due to some external factor (e.g., leukemia is 7) found that noncontingency is more dif- incurable). Alternatively, in the school example, failing is consistently associated with the student ficult to perceive when one is winning than and not associated with his peers. Here, the student when one is losing. So the notion of uncon- would conclude that some internal factor (e.g., stu- trollability means more than just failure, and pidity) was the cause of his failure. Thus, Heider and Kelley also rely on social comparison as a major it makes predictions concerning both failure determinant of internality. and noncontingent success. CRITIQUE AND REFORMULATION 55

In ordinary language, failure means more We now turn to the second set of inade- than merely the occurrence of a bad outcome. quacies. The old hypothesis was vague about People say they have failed when they have when helplessness would be general across tried unsuccessfully to reach a goal and at- situations and when specific, and when it tribute this to some internal factor. Obtaining would be chronic and when acute. We now poor grades in school is considered failure, formulate this inadequacy and develop an but being caught in a flash flood is generally attributional framework that resolves it. considered misfortune. The concepts of trying and personal helplessness are both necessary Generality and Chronicity of Helplessness to analyze failure in the ordinary language sense. According to the reformulated model, Inadequacy 2 of the Old Theory then, failure, seen from the individual's point A second set of examples point to the other of view, means the subset of personal help- inadequacy of the old helplessness hypothesis. lessness involving unsuccessful trying. Consider debriefing in a typical human help- The final ramification of the distinction be- lessness study: The subject is presented with tween universal and personal helplessness is an unsolvable problem, tested on a second that it deduces a fourth deficit of human help- solvable task, and finally debriefed. The sub- lessness—low self-esteem. A major determi- ject is told that the first problem was ac- nant of attitudes toward the self is com- tually unsolvable and therefore no one could parison with others (Clark & Clark, 1939; have solved it. Experimenters in human help- Festinger, 1954; Morse & Gergen, 1970; lessness studies seem to believe that telling Rosenberg, 1965). Our analysis suggests that a subject that no one could solve the problem individuals who believe that desired outcomes will cause helplessness deficits to go away. are not contingent on acts in their repertoires The prior discussion suggests that convincing but are contingent on acts in the repertoires a subject that his helplessness is universal of relevant others, will show lower self-esteem rather than personal will remove self-esteem than individuals who believe that desired out- deficits suffered in the experiment. Neither comes are neither contingent on acts in their the old nor the new hypothesis, however, pre- repertoires nor contingent on acts in the rep- dicts that such debriefing will remove the ertoires of relevant others. That is, an unin- cognitive and motivational deficits. What does telligent student who fails an exam his peers debriefing undo and why? pass will have lower self-esteem than a stu- A second way of illustrating this inade- dent who fails an exam that all of his peers quacy is the following: A number of investi- fail as well. gators (Hanusa & Schulz, 1977; Tennen & The dichotomy between universal and per- Eller, 1977; Wortman & Brehm, 1975) have sonal helplessness determines cases of help- emphasized those cases of learned helplessness lessness (and depression, see below) with in which a person inappropriately generalizes and without low self-esteem. But it is neutral the expectation of noncontingency to a new, with regard to the cognitive and motivational controllable situation. It is important to point deficits in helplessness. It is important to em- out that the old hypothesis does not require phasize that the cognitive and motivational an inappropriate generalization for helpless- deficits occur in both personal and universal ness. Helplessness exists when a person shows helplessness. Abramson (1977) has demon- motivational and cognitive deficits as a con- sequence of an expectation of uncontrollabil- strated this empirically while showing that ity. The veridicality of the belief and the lowered self-esteem occurs only in personal range of situations over which it occurs are helplessness. According to both the old and irrelevant to demonstrating helplessness. But the new hypotheses, the expectation that out- the old hypothesis does not specify where and comes are noncontingently related to one's when a person who expects outcomes to be own responses is a sufficient condition for uncontrollable will show deficits. In keeping motivational and cognitive deficits. with the resolution of the first inadequacy, an 56 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE attributional framework is now presented to Again, the explanation was largely post hoc. resolve the second inadequacy by explaining Helplessness that dissipated rapidly was as- the generality and chronicity of deficits as- sumed to have strong proactive or retroactive sociated with helplessness. interference; that which persisted was not. The reformulated hypothesis makes a ma- A Resolution: The Attributional Dimensions jor new set of predictions about this topic: of Stability and Generality The helpless individual first finds out that cer- Helplessness deficits are sometimes highly tain outcomes and responses are independent, general and sometimes quite specific. An ac- then he makes an attribution about the cause. countant, fired from his job, may function This attribution affects his expectations about poorly in a broad range of situations: he can- future response-outcome relations and thereby not get started on his income tax, he fails to determines, as we shall see, the chronicity, look for a new job, he becomes impotent, he generality, and to some degree, the intensity neglects his children, and he avoids social of the deficits. Some attributions have global, gatherings. In contrast, his helplessness may others only specific, implications. Some at- be situation specific: He does not do his in- tributions have chronic, others transient, im- come tax and fails to look for a new job, but plications. Consider an example: You submit he remains an adequate lover, father, and a paper to a journal and it is scathingly re- party-goer, When helplessness deficits occur jected by a consulting editor. Consider two in a broad range of situations, we call them possible attributions you might make: "I am global; when the deficits occur in a narrow stupid" and "The consulting editor is stupid." range of situations, we call them specific. The first, "I am stupid," has much more dis- The time course of helplessness (and de- astrous implications for your future paper- pression, see below) also varies from individ- submitting than the second. If "I am stupid" ual to individual. Some helplessness deficits is true, future papers are likely to be rejected may last only minutes and others may last as well. If "The editor is stupid" is true, fu- years. Helplessness is called chronic when it is ture papers stand a better chance of being either long-lived or recurrent and transient accepted as long as you do not happen on the when short-lived and nonrecurrent. same consulting editor. Since "I" is something The old hypothesis was vague about gen- I have to carry around with me, attributing erality and chronicity. The helpless person had the cause of helplessness internally often but learned in a particular situation that certain not always (see below) implies a grimmer responses and outcomes were independent. future than attributing the cause externally, The deficits resulting could crop up in new since external circumstances are usually but situations if either the responses called for or not always in greater flux than internal fac- the outcomes desired were similar to the re- tors. sponses and outcomes about which original Recent attribution theorists have refined learning had occurred. Helplessness was global the possible attribution for outcomes by sug- when it depressed responses highly dissimilar gesting that the dimension "stable-unstable" to those about which original learning had oc- is orthogonal to "internal-external" (Weiner, curred or when it extended to stimuli highly 1974; Weiner, Frieze Kukla, Reed, Rest, & dissimilar to those about which original learn- Rosenbaum, 1971). Stable factors are thought ing had occurred. No account was given about of as long-lived or recurrent, whereas un- why helplessness was sometimes specific and stable factors are short-lived or intermittent. sometimes global. When a bad outcome occurs, an individual When helplessness dissipated in time, for- can attribute it to (a) lack of ability (an in- getting produced by interference from prior ternal-stable factor), (b) lack of effort (an or later learning was invoked (e.g., Seligman, internal-unstable factor), (c) the task's being 1975, pp. 67-68). Forgetting of helplessness too difficult (an external-stable factor), or could be caused either by earlier mastery (d) lack of luck (an external-unstable fac- learning or by subsequent mastery learning. tor). CRITIQUE AND REFORMULATION 57

Table 2 Formal Characteristics of Attribution and Some Examples

Internal External Dimension Stable Unstable Stable Unstable

Global Failing student Lack of intelligence Exhaustion ETS gives unfair tests. Today is Friday the 13th. (Laziness) (Having a cold, which (People are usually (ETS gave experimental tests makes me stupid) unlucky on the ORE.) this time which were too hard for everyone.) Rejected woman I'm unattractive to My conversation some- Men are overly compet- Men get into rejecting men, times bores men. itive with intelligent moods. women. Specific Failing student Lack of mathematical Fed up with math ETS gives unfair math The math test was from ability problems tests. No. 13. (Math always bores (Having a cold, which (People are usually (Everyone's copy of the me.) ruins my arithmetic) unlucky on math tests.) math test was blurred.) Rejected woman I'm unattractive to My conversation bores He's overly competi- He was in a rejecting mood. him. him. tive with women.

Note. ETS = Educational Testing Service, the maker of graduate record examinations (GRE).

While we applaud this refinement, we be- ation). Table 2 describes the formal character- lieve that further refinement is necessary to istics of the attributions and exemplifies them. specify the attributions that are made when Table 1 relates to Table 2 in the following an individual finds himself helpless. In par- way: Table 2 uses the attributional dimen- ticular, we suggest that there is a third di- sions of stability and generality to further mension—"global-specific"—orthogonal to in- subdivide the cases of personal helplessness ternality and stability, that characterizes the (Cell 3—internal attribution) and universal attributions of people. Global factors affect helplessness (Cell 4—external attribution) in a wide variety of outcomes, but specific fac- Table 1. tors do not.3 A global attribution implies that According to the reformulated hypothesis, helplessness will occur across situations, if the individual makes any of the four global whereas a specific attribution implies helpless- attributions for a low math score, the deficits ness only in the original situation. This di- observed will be far-reaching: Global attribu- mension (like those of stability and internal- tions imply to the individual that when he ity) is a continuum, not a dichotomy; for the confronts new situations the outcome will sake of simplicity, however, we treat it here again be independent of his responses. So, if as a dichotomy. he decides that his poor score was caused by Consider a student taking graduate record his lack of intelligence (internal, stable, examinations (GREs) measuring mathemati- global) or his exhausted condition (internal, cal and verbal skills. He just took the math test and believes he did very poorly. Within the three dimensions, there are eight kinds of 3 In principle, there are a large number of di- attribution he can make about the cause of mensions on which attributions can be specified. his low score (Internal-External X Stable-Un- Weiner (Note 8) suggested that the criterion for a stable X Global-Specific). These attributions dimension, as opposed to a mere property, of attri- bution be that we can sensibly ask, Does it apply to have strikingly different implications for how all the causes that we assign to behavior? So stable- he believes he will perform in the next hour on unstable is a dimension because we can sensibly ask, the verbal test (generality of the helplessness Is ability a factor that persists stably over time? Is deficit across situations) and for how he be- patience a factor that persists stably?, and so on. Similarly, global-specific qualifies as a dimension lieves he will do on future math tests when he since we can ask sensibly, Is ability a factor that retakes the GRE some months hence (chro- affects many situations or only few? Is patience a nicity of the deficit over time in the same situ- factor that affects many situations?, and so on. 58 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE unstable, global) or that the Educational men (internal, unstable, global); men are Testing Service (ETS; the creator of GREs) overly competitive with intelligent women gives unfair tests (external, stable, global) or (external, stable, global); men get into re- that it is an unlucky day (external, unstable, jecting moods (external, unstable, global). All global), when he confronts the verbal test in four of these attributions will produce help- a few minutes, he will expect that here, as lessness deficits in new situations with most well, outcomes will be independent of his re- other men. The four specific attributions will sponse, and the helplessness deficits will ensue. produce deficits only with this particular man: If the individual makes any of the four spe- I'm unattractive to him (internal, stable, cific attributions for a low math score, help- specific); my conversation somtimes bores lessness deficits will not necessarily appear him (internal, unstable, specific); he is overly during the verbal test: i.e., lack of mathe- competitive with intelligent women (external, matical ability (internal, stable, specific) or stable, specific); he was in a rejecting mood being fed up with math problems (internal, (external, unstable, specific). Any of the four unstable, specific) or that ETS asks unfair stable attributions will produce chronic defi- math questions (external, stable, specific) or cits either with that man (if specific) or with being unlucky on that particular math test most men (if global); the four unstable at- (external, unstable, specific). tributions will produce transient deficits. The In a parallel manner, chronicity of the four internal attributions will produce self- deficits follows from the stability dimension. esteem deficits; the four external attributions Chronic deficits (he will be helpless on the will not. next math GRE when he retakes it at a later Having stated what we believe are the de- time) will ensue if the attribution is to stable terminants of the chronicity and generality of factors: lack of intelligence, lack of mathe- helplessness deficits, a word about intensity or matical ability, ETS gives unfair tests, ETS severity is in order. Severity is logically inde- gives unfair math tests. Attribution to stable factors leads to chronic deficits because they imply to the individual that he will lack the 4 A critical remark is in order on the adequacy of controlling response in the future as well as ability, effort, task difficulty, and luck as embody- ing, respectively, internal-stable, internal-unstable, now. If the attribution is to an unstable factor external-stable, external-unstable attributions (Wei- —exhaustion, fed up with the math problems, ner et al., 1971). While we find the orthogonality of unlucky day, or unlucky on the math tests— internality and stability dimensions useful and im- he will not necessarily be helpless on the portant, we do not believe that the ability-effort/ next math GRE. If he makes any of the in- task difficulty-luck distinctions map into these di- mensions. Table 2 presents (in parentheses) attri- ternal attributions—lack of intelligence, lack butions that systematically violate the mapping. An of math ability, exhaustion, or fed up with internal-stable attribution for helplessness need not math problems—the self-esteem deficits will be to lack of ability; it can be to lack of effort: occur. In contrast, none of the external at- laziness (global), math always bores me (specific). 4 An internal-unstable attribution need not be to tributions will produce self-esteem deficits. lack of effort, it can be to (temporary) inabilities: Because so much real-life helplessness stems I have a cold, which makes me stupid (global); I from social inadequacy and rejection, Table have a cold, which ruins my arithmetic ability (spe- 2 illustrates a social example. Here a woman cific). An external-stable attribution need not be to task difficulty; it can be to lack of luck: Some is rejected by a man she loves. Her attribu- people are always unlucky on tests (global); people tion for failure will determine whether she are always unlucky on math tests (specific). An shows deficits in situations involving most external-unstable attribution need not be to bad other men (global) and whether she shows luck; it can be to task difficulty: ETS gave experi- mental tests this time that were difficult for every- deficits in the future with this particular man one (global); everyone's copy of the math test was or with other men (chronic). She might se- blurred (specific). So, ability and effort are logically lect any of four types of global attributions: orthogonal to internal-stable and internal-unstable attributions, and luck and task difficulty are or- I'm unattractive to men (internal, stable, thogonal to external-stable and external-unstable global); my conversation sometimes bores attributions. CRITIQUE AND REFORMULATION 59 pendent of chronicity and generality; it re- tion. So the person may find out by interven- fers to how strong a given deficit is at any ing successes that he was not as stupid as he one time in a particular situation. We believe thought, or he may gather evidence that that the intensity of the motivational and everyone obtained low scores on the math cognitive deficits increases with the strength GRE and so now ETS is under new manage- or certainty of the expectation of noncontin- ment. In such cases, the expectation need not gency. We speculate that intensity of self- be present across situations and time. On esteem loss and affective changes (see Im- the other hand, if the expectation is present, plications of the Reformulated Model for the then helplessness deficits must occur (see Helplessness Model of Depression below) will Weiner, 1972, for a related discussion of increase with both certainty and importance achievement motivation). of the event the person is helpless about. We also speculate that if the attribution is global or stable, the individual will expect to be Implications helpless in the distant future (both across The attributional account of the chronicity^ areas of his life and across time) as well as in and generality of the symptoms of helpless-1 the immediate future. The future will look ness explains why debriefing ensures that defi-l black. This expectation will increase the in- cits are not carried outside the laboratory. * tensity of the self-esteem and affective defi- The debriefing presumably changes the at- cits. If the attribution is internal, this may tribution from a global (and potential^ also tend to make these deficits more severe, harmful outside the laboratory) and possibly since internal attributions are often stable internal (e.g., I'm stupid) one to a more and global. specific and external one (e.g., psychologists are nasty: They give unsolvable problems to Attribution and Expectancy experimental subjects). Since the attribution for helplessness is to a specific factor, the ex- In general, the properties of the attribution pectation of uncontrollability will not recur predict in what new situations and across what outside the laboratory anymore than it would span of time the expectation of helplessness have without the experimental evidence. will be likely to recur. An attribution to These attributional dimensions are also rele- global factors predicts that the expectation vant to explaining when inappropriate, broad will recur even when the situation changes, generalization of the expectation of noncon- whereas an attribution to specific factors tingency will occur. Broad transfer of help- predicts that the expectation need not re- lessness will be observed when subjects at- cur when the situation changes. An attribu- tribute their helplessness in the training phase tion to stable factors predicts that the expecta- to very global and stable factors. Alterna- tion will recur even after a lapse of time, tively, attributing helplessness to very specific whereas an attribution to unstable factors and unstable factors predicts very little trans- predicts that the expectation need not recur fer of helplessness. after a lapse of time. Whether or not the ex- A final question concerns the determinants pectation recurs across situations and with of what particular attribution people make elapsed time determines whether or not the for their helplessness. Attribution theorists helplessness deficits recur in the new situation (e.g., Heider, 1958; Kelley, 1967; Weiner, or with elapsed time. Notice that the attribu- 1974) have discussed situational factors that tion merely predicts the recurrence of the ex- influence the sort of attribution people make. pectations but the expectation determines the In addition, Heider and Kelley pointed to sys- occurrence of the helplessness deficits. New tematic biases and errors in the formation of evidence may intervene between the initial attributions. Later, we discuss an "attribu- selection of an attribution and the new and tional style" that may characterize depressed subsequent situation and change the expecta- people. 60 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE

Validity of the Reformulated Model impossible) and stable enough to survive the brief time interval between tests. The data The validity of the new hypothesis must are ambiguous about whether the global, ultimately be assessed by its ability to gen- stable attribution is internal (e.g., I'm stupid) erate novel predictions that survive attempts or external (e.g., laboratory problems are at experimental disconfirmation. As it is a impossible); self-esteem changes would have new hypothesis, no results from such at- been relevant to this determination. Nonde- tempts are yet available. However, a mini- pressed students who escape noise, solve prob- mum requirement is that this hypothesis lems, or receive nothing as pretreatment do should be consistent with the available ex- not perceive response-outcome independence perimental evidence. Although such consist- and do not, of course, make any attribution ency can lend only limited support to the about such independence. hypothesis (as the available evidence has For a control procedure, subjects have been been one factor shaping the hypothesis), in- told to listen to noise (which is inescapable) consistency might seriously embarrass the but not to try to do anything about it hypothesis. (Hiroto & Seligman, 1975); similarly, sub- jects have been given a panic button that "will escape noise if pressed" but have been Is the Reformulated Hypothesis Consistent successfully discouraged from pressing ("I'd with the Experimental Evidence on Learned rather you didn't, but it's up to you"; (Glass Helplessness in Humans? & Singer, 1972). These subjects do not be- Three basic classes of evidence are covered: come helpless. Both the old and reformulated (a) deficits produced by learned helplessness, hypotheses hold that these subjects do not (b) attributional evidence, and (c) skill/ perceive noncontingency (in this latter case, chance evidence. they perceive potential response-outcome con- Deficits produced by learned helplessness. tingency; in the first case, they have no rele- Nondepressed students given inescapable noise vant perception) and so do not form the or unsolvable discrimination problems fail to relevant expectations and attributions. escape noise (Glass, Reim, & Singer, 1971; A number of studies on human helplessness Hiroto & Seligman, 1975; Klein & Seligman, have obtained findings that are difficult to 1976; Miller & Seligman, 1976), fail to solve explain with the old helplessness hypothesis. anagrams (Benson & Kennelly, 1976; Gatchel Examination of these studies suggests that in- & Proctor, 1976; Hiroto & Seligman, 1975; vestigators may have tapped into the at- Klein et al., 1976), and fail to see patterns tributional dimensions of generality and sta- in anagrams (Hiroto & Seligman, 1975; Klein bility. For example, Roth and Kubal (1975) et al., 1976). Escapable noise, solvable dis- tested helplessness across very different situa- crimination problems, or no treatment does tions: Subjects signed up for two separate ex- not produce these deficits. Both the old and periments that happened to be on the same the reformulated hypotheses explain these day in the same building. They failed on the deficits by stating that subjects expect that task in Experiment 1 (pretraining) and then outcomes and responses are independent in wandered off to Experiment 2 (the test task). the test situation. This expectation produces When subjects were told in Experiment 1 that the motivational deficit (failure to escape they had failed a test that was a "really noise and failure to solve anagrams) and the good predictor of grades in college" (impor- cognitive deficit (failure to see patterns). The tant), they showed deficits on the cognitive reformulated hypothesis adds an explanation problem of Experiment 2. When told that Ex- of why the expectation for the inescapability periment 1 was merely "an experiment in of the noise or the unsolvability of the dis- learning" (unimportant), they did better on crimination problems must have been global Experiment 2. In the case of "good predictor enough to transfer across situations (e.g., I'm of grades," subjects probably made a more unintelligent; problems in this laboratory are global, internal, and possibly more stable at- CRITIQUE AND REFORMULATION 61 tribution (e.g., I'm stupid enough to do badly ing attribution along the internal-external on this, therefore on college exams as well). dimension. Although the relevant data were The expectation therefore recurred in the new not collected, it is likely that real, but not situation, producing deficits. In the unimpor- recalled, success modifies attribution along tant condition, subjects probably made a more the global-specific dimension. Immunization specific and less stable attribution, so the ex- (Thornton & Powell, 1974; Dyck & Breen, pectation of failure was not present in Ex- Note 9) is explained similarly: Initial success periment 2. (See Cole and Coyne, 1977, for experience should make the attribution for another way of inducing a specific, rather than helplessness less global and therefore less a global, attribution for failure.) likely to recur in the new test situation. Similarly, Douglas and Anisman (1975) A number of human helplessness studies found that failure on simple tasks produced have actually shown facilitation in subjects later cognitive deficits but that failure on com- exposed to uncontrollable events (Hanusa & plex tasks did not. It seems reasonable that Schulz, 1977; Roth & Kubal, 1975; Tennen failure on simple tasks should produce a more & Eller, 1977; Wortman et al., 1976). While global and internal attribution (e.g., I'm such facilitation is not well understood (see stupid) whereas failure on the complex tasks Wortman & Brehm, 1975; Roth & Kilpatrick- could be attributed to external and more Tabak, Note 2, for hypotheses), it seems rea- specific factors (e.g., these problems are too sonable that compensatory attempts to reas- difficult). sert control might follow helplessness experi- An important advantage of the reformula- ences, once the person leaves the situations in tion is that it better explains the effects of which he believes himself helpless (see Solo- therapy and immunization than does the old mon & Corbit, 1973, for a relevant rebound hypothesis. The key here is the attributional theory). Such compensatory rebound might dimension of generality. Helplessness can be be expected to dissipate in time and be less reversed and prevented by experience with strong in situations very far removed from the success. Klein and Seligman (1976) gave non- original helplessness training. When the "fa- depressed people inescapable noise and then cilitation" effect of helplessness is brought did "therapy," using 4 or 12 cognitive prob- under replicable, experimental control, the lems, which the subjects solved. (Therapy compensatory rebound hypothesis can be was also performed on depressed people given tested. People may also show facilitation of no noise.) Therapy worked: The subjects performance in uncontrollable situations when (both depressed and nondepressed) escaped they cannot find a controlling response but noise and showed normal expectancy changes have not yet concluded that they are helpless. after success and failure. Following inescap- The reformulated hypothesis accounts for able noise the subjects presumably made an the basic helplessness results better than does attribution to a relatively global factor (e.g., the old hypothesis. The explanations given by I'm incompetent, or laboratory tasks are un- the reformulated hypothesis are necessarily solvable), which was revised to a more spe- post hoc, however. Relevant measures of the cific one after success on the next task (e.g., generality, stability, and internality of at- I'm incompetent in only some laboratory situ- tribution were not made. Helplessness studies • ations, or, only some laboratory tasks are diffi- can, in principle, test the hypothesis either by cult). The new test task, therefore, did not measuring the attributions and correlating evoke the expectation of uncontrollability. them with the deficits that occur or by induc- Teasdale (1978) found that real success ex- ing the attributions and predicting deficits. periences and recalling similar past successes We now turn to the few studies of helplessness were equally effective in shifting attribution that have induced or measured attribution. for initial failure from internal to external Attributional evidence. Dweck and her as- factors. Only real success, however, reversed sociates (Dweck, 1975; Dweck & Reppucci, helplessness performance deficits. This sug- 1973; Dweck, Davidson, Nelson, & Enna, gests success does not have its effect by shift- Note 10; Dweck, Goetz, & Strauss, Note 11) 62 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE have demonstrated the differential effects of universal and personal helplessness. In both attribution for failure to lack of ability versus cases people expect that outcomes are inde- lack of effort. When fourth-grade girls fail, pendent of their responses. In addition, at- they attribute their failure to lack of ability tribution of helplessness to specific or un- (consonant with their teachers' natural class- stable factors should be less likely to lead to room criticisms of girls) and perform badly performance deficits than attribution to stable on a subsequent cognitive test. Lack of ability or global factors. To date, four studies have is a global attribution (as well as internal manipulated attribution for helplessness in and stable) and implies failure expectation adults. In line with the reformulation, Klein for the new task. Fourth-grade boys, on the et al. (1976) found that relative to groups other hand, attribute failure to lack of effort receiving solvable problems or no problems at or bad conduct (also consonant with the all, nondepressed students did poorly on an teachers' natural classroom criticisms of boys) anagrams task following experience with un- and do well on the subsequent test. Lack of solvable discrimination problems regardless of effort is unstable and probably more specific whether they attributed their helplessness to (but also internal). Boys, having failed and internal factors (personal helplessness) or ex- attributed failure to lack of effort, put out ternal factors (universal helplessness). more effort on the test task and do adequately. Tennen and Eller (1977) attempted to ma- Similarly, when students are told to attribute nipulate attribution by giving subjects un- failure on math problems to not trying hard solvable discrimination problems that were enough, they also do better than if they at- labeled either progressively "easier" or pro- tribute it to lack of ability (Dweck, 1975). gressively "harder." The authors reasoned Effort is not only "unstable," but it is that failure on easy problems should produce readily controllable by the subject himself, attribution to lack of ability (internal, stable, unlike being bored, for example, which is also and more global) whereas failure on hard unstable, specific, and internal, or unlike lack problems should allow attribution to task dif- of ability. It should be noted that the dimen- ficulty (external, unstable, and more specific). sion of controllability is logically orthogonal Subjects then went to what they believed was to the Internal X Global X Stable dimensions a second, unrelated experiment (see Roth & (although it is empirically more frequent in Kubal, 197S) and tried to solve anagrams. In the internal and unstable attribution), and as line with the reformulation, attribution to in- such it is a candidate fora2X2X2x2 ability (easy problems) produced deficits. At- table of attributions. While we do not detail tribution to task difficulty (hard problems) such an analysis here, we note that the phe- resulted in facilitation of anagram solving. nomena of self-blame, self-criticism, and guilt The most likely explanation for lack of per- (a subclass of the self-esteem deficits) in formance deficits in the task-difficulty group is helplessness (and depression) follow from at- that their attributions for helplessness were tribution of failure to factors that are con- too specific to produce an expectation of non- trollable. Lack of effort as the cause of failure contingency in the test task. probably produces more self-blame than does Finally, two studies (Hanusa & Schulz, boredom, although both are internal and un- 1977; Wortman et al., 1976) found that rela- stable attributions. Similarly, a failure caused tive to a group exposed to contingent events, by not speaking Spanish attributed to lack of neither a group instructed to believe they ability to speak Spanish, which might have were personally helpless nor a group in- been corrected by taking a Berlitz course, structed to believe they were universally help- probably causes more self-blame than a less less on a training task showed subsequent correctable lack of ability, such as ineptitude performance deficits on a test task. While the for foreign languages, even though both are results appear contrary to the reformulation, internal and stable. they are difficult to interpret. The problem is According to the reformulation, perform- that in both studies, the typical helplessness ance deficits should occur in cases of both group (a group exposed to noncontingent CRITIQUE AND REFORMULATION 63

events in the training task but given no ex- changes when they believe outcomes are plicit attribution) did not show performance chance determined. deficits on the test task. Thus, the test task Recent developments in attribution theory may not have been sensitive to helplessness suggest that expectancy changes are not a di- deficits. (For a discussion of the relative sensi- rect index of people's expectations about re- tivity of tasks to helplessness in animals, see sponse-outcome contingencies. Weiner and his Maier and Seligman, 1976.) It is interesting colleagues (1971) argued that the attribu- that Wortman et al. (1976) found that per- tional dimension of stability rather than locus sonally helpless subjects showed more emo- of control is the primary determinant of tional distress than universally helpless sub- expectancy changes. According to Weiner jects. (Weiner, 1974; Weiner, Heckhausen, Meyer, Overall, then, the few helplessness studies & Cook, 1972) people give small expectancy directly assessing and manipulating attribu- changes when they attribute outcomes to un- tion provide some support for the reformula- stable factors and large expectancy changes tion. Because of the methodological problems when they attribute outcomes to stable fac- in some of these studies, future research that tors. The logic is that past outcomes are good manipulates attribution is necessary. Care predictors of future outcomes only when they must be taken to ensure that one attributional are caused by stable factors. dimension is not confounded with another. In the absence of knowledge about individ- Past studies, for example, have confounded ex- ual attributions, the reformulated helplessness ternality with specificity and internality with hypothesis cannot make clear-cut predictions generality. about expectancy changes and helplessness, Helpless subjects show dampened expec- since belief in response-outcome dependence tancy changes in skill tasks. In skill tasks, or independence is orthogonal to stable-un- expectancy for future success increases less stable. For example, suppose a person makes following success and/or decreases less fol- an internal attribution to lack of ability for lowing failure for helpless subjects than for his helplessness, i.e., he believes in response- subjects not made helpless (Klein & Seligman, outcome independence for himself. When con- 1976; Miller & Seligman, 1976; Miller, Selig- fronted with the skill task, he may show very man, & Kurlander, 1975; see also Miller & large expectancy changes after failure since Seligman, 1973, and Abramson, Garber, Ed- he believes he lacks the stable factor of ability wards & Seligman, 1978, for parallel evidence for the task. Alternatively, when confronted in depression). The old hypothesis interpreted with the 50% success rate typically used in these results as a general tendency of help- helplessness studies, he may maintain his be- less subjects to perceive responding and out- lief that he lacks the stable factor of ability comes on skill tasks as independent, and it but conclude that ability is not necessary for was assumed that this index measured the success on the task. After all, he succeeded central helplessness deficit directly. In other sometimes in spite of his perceived lack of words, it had been suggested that such sub- ability. Under such conditions, the person jects perceive skill tasks as if they were will believe outcomes are a matter of chance chance tasks. The rationale for this interpreta- (unstable factor) for himself but not for tion was derived from the work of Rotter and others. Accordingly, he will give small expec- his colleagues (James, 1957; James & Rotter, tancy changes. Moreover, a nonhelpless per- 1958; Phares, 1957; Rotter, Liverant, & son (who perceives response-outcome depen- Crowne, 1961). These investigators argued dency) may believe unstable factors, such as that reinforcements on previous trials have a effort, cause his outcomes and show little ex- greater effect on expectancies for future suc- pectancy change; alternatively, if he believes cess when the subject perceives reinforce- a stable factor is responsible for response- ment as skill determined than when he per- outcome dependence, he will show large shifts. ceives it as chance determined. According to Rizley (1978) similarly argued that ex- this logic, subjects will show large expectancy pectancy changes on chance and skill tasks do 64 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE not directly test the learned helplessness model of depression is that learning that out- model of depression. We agree. As argued in comes are uncontrollable results in the moti- the previous paragraph, small expectancy vational, cognitive, and emotional components changes need not imply belief in independence of depression (Seligman, 1975; Seligman et between responses and outcomes, and large al., 1976). The motivational deficit consists expectancy changes need not imply belief in of retarded initiation of voluntary responses, dependence between responses and outcomes. and it is reflected in passivity, intellectual Nor does belief in response-outcome inde- slowness, and social impairment in naturally pendence imply small expectancy changes, or occurring depression. According to the old belief in dependence imply large changes. The model, deficits in voluntary responding fol- fact that depressives often show smaller ex- low directly from expectations of response- pectancy changes than nondepressed people outcome independence. The cognitive deficit (Abramson et al., 1978; Klein & Seligman, consists of difficulty in learning that responses 1976; Miller & Seligman, 1973, 1976; Miller produce outcomes and is also seen as a conse- et al., 1975) is intriguing but provides only quence of expecting response-outcome inde- limited support for the learned helplessness pendence. In the clinic, "negative cognitive model. In order for expectancy changes to be set" is displayed in depressives' beliefs that used as a way of inferring perception of re- their actions are doomed to failure. Finally, sponse-outcome independence, the particular the model asserts that depressed affect is a attribution and its stability must also be consequence of learning that outcomes are un- known. None of the studies to date that mea- controllable. It is important to emphasize sured expectancy shifts also measured the that the model regards expectation of re- relevant attributions, so these studies do not sponse-outcome independence as a sufficient, tell us unambiguously that helpless (or de- not a necessary, condition for depression. pressed) people perceive response-outcome Thus, physiological states, postpartum condi- independence. They support the model only tions, hormonal states, loss of interest in re- in as far as these two groups show the same inforcers, chemical depletions, and so on may pattern of shifts, but the pattern itself can- produce depression in the absence of the ex- not be predicted in the absence of knowledge pectation of uncontrollability. According to about the accompanying attribution. the model, then, there exists a subset of de- To conclude this section, examination of pression—helplessness depressions—that is expectancy changes on chance and skill tasks caused by expectation of response-outcome is not a direct way of testing helplessness, independence and displays the symptoms of since such changes are sensitive to the attribu- passivity, negative cognitive set, and de- tional dimension of stability and not to ex- pressed affect. pectations about response-outcome contin- We believe that the original formulation of gencies. Recent failures to obtain small ex- the learned helplessness model of depression pectancy changes in depressed people (Mc- is inadequate on four different grounds: (a) Nitt & Thornton, 1978; Willis & Blaney, Expectation of uncontrollability per se is not 1978) are disturbing empirically, but less so sufficient for depressed affect since there are theoretically, since both depressed and help- many outcomes in life that are uncontrollable less subjects show the same pattern, albeit a but do not sadden us. Rather, only those un- different pattern from the one usually found. controllable outcomes in which the estimated probability of the occurrence of a desired out- come is low or the estimated probability of Implications of the Reformulated Model for the occurrence of an aversive outcome is high the Helplessness Model of Depression are sufficient for depressed affect, (b) Low- ered self-esteem, as a symptom of the syn- This reformulation of human helplessness drome of depression, is not explained, (c) The has direct implications for the helplessness tendency of depressed people to make in- model of depression. The cornerstone of pre- ternal attributions for failure is not explained, vious statements of the learned helplessness (d) Variations in generality, chronicity, and CRITIQUE AND REFORMULATION 65 intensity of depression are not explained. All the unavoidable outcome, and with the but the first of these shortcomings are di- strength or certainty of the expectation of rectly remedied by the reformulation of hu- uncontrollability. In addition, intensity of man helplessness in an attributional frame- depressed affect may depend on whether the work. person views his helplessness as universal or personal. Weiner (1974) suggested that fail- Inadequacy 1: Expectation of Uncontrolla- ure attributed to internal factors, such as bility Is Not Sufficient for Depressed Affect lack of abiity, produces greater negative af- fect than failure attributed to external fac- We view depression, as a syndrome, to be tors, such as task difficulty. The intensity of made up of four classes of deficits: (a) moti- cognitive and motivational components of vational, (b) cognitive, (c) self-esteem, and depression, however, does not depend on (d) affective (but see Blaney, 1977, for a whether helplessness is universal or personal, review that contends that only affective or, we speculate, on the importance of the changes are relevant to depression). Whereas event. the first three deficits are the result of un- Perhaps the expectation that one is receiv- controllability, we believe the affective ing positive events noncontingently contri- changes result from the expectation that bad butes indirectly to vulnerability to depressed outcomes will occur, not from their expected affect. Suppose a person has repeatedly uncontrollability. learned that positive events arrive indepen- Everyday observation suggests that an ex- dently of his actions. If the perception or ex- pectation that good events will occur with a pectation of response-outcome independence high frequency but independently of one's in future situations involving loss is facili- responses is not a sufficient condition for tated by such a set, then heightened vulnera- depressed affect (see Seligman, 1975 (p. 98), bility to depression will occur. versus Maier & Seligman, 1976 (p. 17), for previous inconsistent accounts). People do Inadequacy 2: Lowered Self-esteem as a not become sad when they receive $1,000 each Symptom of Depression month from a trust fund, even though the A number of theoretical perspectives (Beck, money comes regardless of what they do. In this case, people may learn they have no 1967, 1976; Bibring, 1953; Freud, 1917/ control over the money's arrival, become pas- sive with respect to trying to stop the money 5 One problem remains. It is a "highly desired" from arriving (motivational deficit), have outcome for us that the editor of this journal give trouble relearning should the money actually us each one million dollars, and we believe this to become response contingent (cognitive defi- have a very low probability and to be uncontrolla- ble. Yet, we do not have depressed affect upon real- cit), but they do not show dysphoria. Thus, izing this. Some notion, like Klinger's (1Q7S) "cur- only those cases in which the expectation of rent concerns," is needed to supplement our account. response-outcome independence is about the We feel depressed about the nonoccurrence of highly loss of a highly desired outcome 5 or about desired outcomes that we are helpless to obtain only when they are "on our mind," "in the realm of the occurrence of a highly aversive outcome possibility," "troubling us now," and so on. Inci- are sufficient for the emotional component of dentally, the motivational and cognitive deficits do depression. It follows, then, that depressed not need current concerns, only the affective deficit. affect may occur in cases of either universal We take this inadequacy to be general not only to or personal helplessness, since either can in- the theory stated here but to much of the entire psychology of motivation, which focuses on behav- volve expectations of uncontrollable, impor- ior, and we do not attempt to remedy it here. We tant outcomes. find Klinger's concept heuristic but in need of some- At least three factors determine the in- what better definition. We, therefore, use the notion tensity of the emotional component of depres- of "loss of a highly desired outcome" rather than "nonoccurrence." Loss implies that it will probably be sion. Intensity of affect (and self-esteem defi- a current concern. Since this is only part of a suf- cits) increases with desirability of the unob- ficiency condition, we do not deny that nonoccur- tainable outcome or with the aversiveness of rence can also produce depressed affect. 66 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE

1957) regard low self-esteem as a hallmark helplessness model of depression. Similarly, symptom of depression. Freud has written, aware that depressives often blame them- "The melancholic displays something else be- selves for bad outcomes, Abramson and sides which is lacking in mourning—an extra- Sackeim (1977) asked how individuals can ordinary diminution in his self-regard, an possibly blame themselves for outcomes about impoverishment of his ego on a grand scale" which they believe they can do nothing. Al- (p. 246). A major shortcoming of the old though the reformulation does not articulate model of depression is that it does not explain the relation between blame or guilt and help- the depressive's low opinion of himself. Our lessness, it clearly removes any contradiction analysis of universal and personal helplessness between being a cause and being helpless. De- suggests that depressed individuals who be- pressed individuals who believe they are per- lieve their helplessness is personal show lower sonally helpless make internal attributions for self-esteem than individuals who believe their failure, and depressed individuals who believe helplessness is universal. Suppose two indi- they are universally helpless make external at- viduals are depressed because they expect that tributions for failure. A personally helpless in- regardless of how hard they try they will re- dividual believes that the cause of the failure main unemployed. The depression of the per- is internal (e.g., I'm stupid) but that he is son who believes that his own incompetence helpless (No response I could make would is causing his failure to find work will feel low help me pass the exam). self-regard and worthlessness. The person who What are the naturally occurring attribu- believes that nationwide economic crisis is tions of depressives? Do they tend to at- causing his failure to find work will not think tribute failure to internal, global, and stable less of himself. Both depressions, however, will factors, and success to external, specific, and show passivity, negative cognitive set, and unstable factors?6 sadness, the other three depressive deficits, Hammen and Krantz (1976) looked at since both individuals expect that the prob- cognitive distortion in depressed and nonde- ability of the desired outcome is very low and pressed women. When responding to a story that it is not contingent on any responses in containing "being alone on a Friday night," their repertoire. depressed women selected more depressed-dis- It is interesting that psychoanalytic writers torted cognitions ("upsets me and makes me have argued that there are at least two types start to imagine endless days and nights by of depression, which differ clinically as well as myself"), and nondepressed women selected theoretically (Bibring, 1953). Although both more nondepressed-nondistorted cognitions types of depression share motivational, cogni- ("doesn't bother me because one Friday night tive, and affective characteristics, only the alone isn't that important; probably everyone second involves low self-regard. Further paral- has spent nights alone"). Depressed people leling our account of two types of depression seem to make more global and stable attribu- is recent empirical work (Blatt, D'Afflitti, & tions for negative events. When depressed Quinlan, 1976) suggesting that depression can women were exposed to failure on an inter- be characterized in terms of two dimensions: personal judgment task, they lowered their dependency and feelings of deprivation, and self-rating more than did nondepressed women. low self-esteem and excessively high stan- dards and morality. aThe literature on the relation between internal locus of control and depression might be expected to Inadequacy 3: Depressives Believe They yield direct information about internal attribution in Cause Their Own Failures depression. It is, however, too conflicting at this stage to be very useful. Externality, as measured by Recently, Blaney (1977) and Rizley (1978) the Rotter scale, correlates weakly (.2S-.30) with have construed the finding that depressives at- depression (Abramowitz, 1969; Miller & Seligman, 1973), but the external items are also rated more tribute their failures to internal factors, such dysphoric and the correlation may be an artifact as lack of ability, as disaffirming the learned (Lamont, 1972). CRITIQUE AND REFORMULATION 67

This indicates that the depressed women are sion and helplessness is necessary from the systematically generating mpre internal as standpoint of our reformulated hypothesis. well as global and stable attributions for failure.7 Rizley (1978) caused depressed and nonde- Inadequacy 4: Generality and Chronicity oj Depression pressed students to either succeed or fail on a cognitive task and then asked them to make The time course of depression varies greatly attributions about the cause. Depressed stu- from individual to individual. Some depres- dents attributed failures to incompetence (in- sions last for hours and others last for years. ternal, global stable), whereas nondepressed "Normal" mourning lasts for days or weeks; students attributed their failures to task dif- many severe depressions last for months or ficulty (external, specific, stable). Similarly, years. Similarly, depressive deficits are some- depressed students attributed success to the times highly general across situations and ease of the task (external, specific, stable), sometimes quite specific. The reformulated whereas nondepressed students attributed their helplessness hypothesis suggests that the chro- success to ability (internal, global, stable). nicity and generality of deficits in helplessness Although inconsistent with the old model, depressions follow from the stability and Rizley's results are highly consistent with the globality of the attribution a depressed per- reformulation. son makes for his helplessness. The same logic Klein et al. (1976) assessed the attribution we used to explain the chronicity and gen- depressed and nondepressed college students erality of helplessness deficits above applies made for failure on discrimination problems. here. Whereas depressed students tended to at- The reformulation also sheds light on the tribute failure to internal factors, nonde- continuity of miniature helplessness depres- pressed students tended to attribute failure sions created in the laboratory and of real-life to external factors. These findings parallel depression. The attributions subjects make those of Rizley on attribution in achievement for helplessness in the laboratory are pre- settings. sumably less global and less stable than at- Garber and Hollon (Note 12) asked de- tributions made by depressed people for fail- pressed and nondepressed subjects to make ure outside the laboratory. Thus, the labora- predictions concerning their own future suc- tory-induced depressions are less chronic and cess as well as the success of another person less global and are capable of being reversed in the skill/chance situation. The depressed by debriefing, but, we hypothesize, they are subjects showed small expectancy changes in relation to their own skilled actions; however, not different in kind from naturally occurring when they predicted the results of the skilled helplessness depressions. They differ only actions of others, they showed large expec- tancy changes, like those of nondepressives 7 Alloy and Abramson (Note 7) also examined rating themselves. These results suggest that distortion, not in attributions but in perception of depressives believe they lack the ability for contingency between depressed and nondepressed the skill task but believe others possess the students. The subjects were exposed to different rela- ability, the internal attribution of personal tions between button pushing and the onset of a green light and were asked to judge the contingency helplessness. between the outcome and the response. Depressed Taken together, the studies examining de- students judged both contingency and noncontin- pressives' attributions for success and failure gency accurately. In contrast, nondepressed students suggest that depressives often make internal, distorted: When the light was noncontingently re- lated to responding but occurred with a high fre- global, and stable attributions for failure and quency, they believed they had control. So there may make external, specific, and perhaps less was a net difference in perception of contingency by stable attributions for their success. Future depressed and nondepressed subjects, but the distor- tion occurred in the nondepressed, who picked up research that manipulates and measures at- noncontingency less readily (see also Jenkins & Ward, tributions and attributional styles in depres- 1965). 68 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE quantitatively, not qualitatively, that is, they 3. The generality of the depressive deficits are mere "analogs" to real helplessness de- will depend on the globality of the attribution pressions. for helplessness, the chronicity of the depres- Do depressive deficits occur in situations sion deficits will depend on the stability of the that have nothing to do with the expectation attribution for helplessness, and whether self- of noncontingency? After failing a math GRE, esteem is lowered will depend on the inter- the student goes home, burns his dinner, cries, nality of the attribution for helplessness. has depressive dreams, and feels suicidal. If 4. The intensity of the deficits depends on this is so, there are two ways our reformula- the strength, or certainty, of the expectation tion might explain this: (a) He is still in the of uncontrollability and, in the case of the af- presence of the relevant cues and expectations, fective and self-esteem deficits, on the im- for even at home the expectation that he will portance of the outcome. not get into graduate school is on his mind, We suggest that the attributional frame- and (b) the expectation, present earlier but work proposed to resolve the problems of absent now, has set off endogenous processes human helplessness experiments also resolves (e.g., loss of interest in the world, catechola- some salient inadequacies of the helplessness mine changes) that must run their course. Re- model of depression. member that expectations of helplessness are held to be sufficient, not necessary, conditions Vulnerability, Therapy, and Prevention of depression. Finally, does the attributional reformula- Individual differences probably exist in at- tion of helplessness make depression look too tributional style. Those people who typically "rational"? The chronicity, generality, and in- tend to attribute failure to global, stable, and tensity of depression follow inexorably, "ra- internal factors should be most prone to gen- tionally" from the attribution made and the eral and chronic helplessness depressions with importance of the outcome. But there is room low self-esteem. By the reformulated hypoth- elsewhere for the irrationality implicit in de- esis, such a style predisposes depression. Beck pression as a form of psychopathology. The (1967) argued similarly that the premorbid particular attribution that depressed people depressive is an individual who makes logical choose for failure is probably irrationally dis- errors in interpreting reality. For example, the torted toward global, stable, and internal fac- depression-prone individual overgeneralizes; a tors and, for success, possibly toward specific, student regards his poor performance in a unstable, and external factors. It is also pos- single class on one particular day as final sible that the distortion resides not in attribu- proof of his stupidity. We believe that our tional styles but in readiness to perceive help- framework provides a systematic framework lessness, as Alloy and Abramson (Note 7) for approaching such overgeneralization: It is have shown: Depressed people perceive non- an attribution to a global, stable, and internal contingency more readily than do nonde- factor. Our model predicts that attributional pressed people. style will produce depression proneness, per- haps the depressive personality. In light of In summary, here is an explicit statement the finding that women are from 2 to 10 of the reformulated model of depression: times more likely than men to have depression 1. Depression consists of four classes of defi- (Radloff, Note 13), it may be important that cits: motivational, cognitive, self-esteem, and boys and girls have been found to differ in at- affective. tributional styles, with girls attributing help- 2. When highly desired outcomes are be- lessness to lack of ability (global, stable) and lieved improbable or highly aversive outcomes boys to lack of effort (specific, unstable; are believed probable, and the individual ex- Dweck, 1976). pects that no response in his repertoire will The therapeutic implications of the refor- change their likelihood, (helplessness) depres- mulated hypothesis can now be schematized. sion results. Depression is most far-reaching when (a) the CRITIQUE AND REFORMULATION 69

Table 3 Treatment, Strategies, and Tactics Implied by the Reformulated Hypothesis

A. Change the estimated probability of the relevant event's occurrence: Reduce estimated likelihood for aversive outcomes and increase estimated likelihood for desired outcomes. a. Environmental maniuplation by social agencies to remove aversive outcomes or provide desired out- comes, for example, rehousing, job placement, financial assistance, provision of nursery care for children. b. Provision of better medical care to relieve pain, correct handicaps, for example, prescription of analgesics, provision of artificial limbs and other prostheses.

B. Make the highly preferred outcomes less preferred. a. Reduce the aversiveness of highly aversive outcomes. 1. Provide more realistic goals and norms, for example, failing to be top of your class is not the end of the world—you can still be a competent teacher and lead a satisfying life. 2. Attentional training and/or reinterpretation to modify the significance of outcomes perceived as aversive, for example, you are not the most unattractive person in the world. "Consider the counter- evidence" (Beck, 1976; Ellis, 1962). 3. Assist acceptance and resignation. b. Reduce the desirability of highly desired outcomes. 1. Assist the attainment of alternative available desired outcomes, for example, encourage the dis- appointed lover to find another boy or girl friend. 2. Assist reevaluation of unattainable goals. 3. Assist renunciation and relinquishment of unattainable goals.

C. Change the expectation from uncontrollability to controllability. a. When responses are not yet within the person's repertoire but can be, train the necessary skills, for example, social skills, child management skills, skills of resolving marital differences, problem-solving skills, and depression-management skills. b. When responses are within the person's repertoire, modify the distorted expectation that the responses will fail. 1. Prompt performance of relevant, successful responses, for example, graded task assignment (Burgess, 1968). 2. Generalized changes in response-outcome expectation resulting from successful performance of other responses, for example, prompt general increase in activity; teach more appropriate goal- setting and self-reinforcement; help to find employment. 3. Change attributions for failure from inadequate ability to inadequate effort (Dweck, 1975), causing more successful responding. 4. Imaginal and miniaturized rehearsal of successful response-outcome sequences: Assertive training, decision-making training, and role playing.

D. Change unrealistic attributions for failure toward external, unstable, specific; change unrealistic at- tributions for success toward internal, stable, global. a. For failure 1. External: for example, "The system minimized the opportunities of women. It is not that you are incompetent." 2. Unstable: for example, "The system is changing. Opportunities that you can snatch are opening at a great rate." 3. Specific: for example, "Marketing jobs are still relatively closed to women, but publishing jobs are not" (correct overgeneralization). b. For success 1. Internal: for example, "He loves you because you are nurturant not because he is insecure." 2. Stable: for example, "Your nurturance is an enduring trait." 3. Global: for example, "Your nurturance pervades much of what you do and is appreciated by every- one around you." 70 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE estimated probability of a positive outcome is examined the effectiveness of therapies that low or the estimated probability of an aver- appear to modify the depressive's cognitive sive outcome is high, (b) the outcome is highly style. One study found that forcing a depres- positive or aversive, (c) the outcome is ex- sive to modify his cognitive style was more pected to be uncontrollable, (d) the attribu- effective in alleviating depressive symptoms tion for this uncontrollability is to a global, than was antidepressant medication (Rush, stable, internal factor. Each of these four Beck, Kovacs, & Hollon, 1977). A second aspects corresponds to four therapeutic strat- study found cognitive modification more ef- egies. fective than behavior therapy, no treatment, 1. Change the estimated probability of the or an attention-placebo therapy in reducing outcome. Change the environment by reducing depressive symptomatology (Shaw, 1977). Fu- the likelihood of aversive outcomes and in- ture research that directly tests the thera- creasing the likelihood of desired outcomes. peutic implications of the reformulation is 2. Make the highly preferred outcomes less necessary. preferred by reducing the aversiveness of un- The reformulation has parallel preventive relievable outcomes or the desirability of un- implications. Populations at high risk for de- obtainable outcomes. pression—people who tend to make stable, 3. Change the expectation from uncontrol- global, and internal attributions for failure— lability to controllability when the outcomes may be identifiable before onset of depres- are attainable. When the responses are not sion. Preventive strategies that force the per- yet in the individual's repertoire but can be, son to criticize and perhaps change his at- train the appropriate skills. When the re- tributional style might be instituted. Other sponses are already in the individual's reper- factors that produce vulnerability are situa- toire but cannot be made because of distorted tions in which highly aversive outcomes are expectation of response-outcome indepen- highly probable and highly desirable outcomes dence, modify the distorted expectation. When unlikely; here environmental change to less the outcomes are unattainable, Strategy 3 pernicious circumstances would probably be does not apply. necessary for more optimistic expectations. A 4. Change unrealistic attributions for failure third general factor producing vulnerability toward external, unstable, specific factors, and to depression is a tendency to exaggerate the change unrealistic attributions for success to- aversiveness or desirability of outcomes. Re- ward internal, stable, global factors. The ducing individuals' "catastrophizing" about model predicts that depression will be most uncontrollable outcomes might reduce the in- far-reaching and produce the most symptoms tensity of future depressions. Finally, a set to when a failure is attributed to stable, global, expect outcomes to be uncontrollable—learned and internal factors, since the patient now helplessness—makes individuals more prone to expects that many future outcomes will be noncontingently related to his responses. Get- depression. A life history that biases indi- ting the patient to make an external, unstable, viduals to expect that they will be able to and specific attribution for failure should re- control the sources of suffering and nurtur- duce the depression in cases in which the ance in their life should immunize against original attribution is unrealistic. The logic, of depression. course, is that an external attribution for failure raises self-esteem, an unstable one cuts the deficits short, and a specific one makes the Reference Notes deficits less general. 1. Miller, I. Learned helplessness in humans: A re- Table 3 schematizes these four treatment view and attribution theory model. Unpublished strategies. •manuscript, Brown University, 1978. Although not specifically designed to test 2. Roth, A., & Kilpatrick-Tabak, B. Developments in the study of learned helplessness in humans: the therapeutic implications of the reformu- A critical review. Unpublished manuscript, Duke lated model of depression, two studies have University, 1977. CRITIQUE AND REFORMULATION 71

3. Tennen, H. A. Learned helplessness and the per- learned helplessness and depression. Unpublished ception of reinforcement in depression: A case of doctoral dissertation, University of Pennsylvania, investigator misattribution. Unpublished manu- 1977. script, State University of New York at Albany, Abramson, L. Y., Garber, J., Edwards, N. B., & 1977. Seligman, M. E. P. Expectancy changes in depres- 4. Eisenberger, R., Mauriello, J., Carlson, J., Frank, sion and schizophrenia, Journal of Abnormal Psy- M., and Park, D. C. Learned helplessness and in- chology, 1978, 87, 102-109. dustriousness produced by positive reinforcement. Abramson, L. Y., & Sackeim, H. A. A paradox in Unpublished manuscript, State University of New depression: Uncontrollability and self-blame. Psy- York at Albany, 1976. chological Bulletin, 1977, 84, 838-8S1. Bandura, A. Self-efficacy: Toward a unifying theory 5. Hirsch, Kenneth A. An extension of the learned of behavioral change. Psychological Review, 1977, helplessness phenomenon to potentially nega- tively punishing and potentially positively rein- 84, 191-215. Beck, A. T. Depression: Clinkal, experimental and forcing stimuli non-contingent upon behavior. theoretical aspects. New York: Hoeber, 1967. Unpublished manuscript, 1976. (Available from Beck, A. T. Cognitive therapy and emotional dis- K. A. Hirsch, Galesburg Mental Health Center, orders. New York: International Universities Galesburg, Illinois.) Press, 1976. 6. Nugent, J. Variations in non-contingent experi- Benson, J. S., & Kennelly, K. J. Learned helplessness: ences and test tasks in the generation of learned The result of uncontrollable reinforcements or un- helplessness. Unpublished manuscript, University controllable aversive stimuli? Journal of Person- of Massachusetts. ality and Social Psychology, 1976, 34, 138-145. 7. Alloy, L. B., & Abramson, L. Y. Judgment of Bibring, E. The mechanism of depression. In P. contingency in depressed and nondepressed stu- Greenacre (Ed.), Affective disorders: Psychoana- dents: A nondepressive distortion. Unpublished lytic contributions to their study. New York: In- manuscript, University of Pennsylvania, 1977. ternational Universities Press, 1953. 8. Weiner, B. Personal communication to M. E. P. Blaney, P. H. Contemporary theories of depression: Seligman, 1977. Critique and comparison. Journal of Abnormal 9. Dyck, D. G., & Breen, L. J. Learned helpless- Psychology, 1977, 86, 203-223. ness, immunization and task importance in hu- Blatt, S. J., D'Afflitti, J. P., & Quinlan, D. M. Ex- mans. Unpublished manuscript, University of periences of depression in normal young adults. Manitoba, 1976. Journal of Abnormal Psychology, 1976, 85, 383- 389. 10. Dweck, C. S., Davidson, W., Nelson, S., & Burgess, E. P. The modification of depressive be- Enna, B. Sex differences in learned helplessness: haviors. In R. D. Rubin & C. M. Franks (Eds.), (II) The contingencies of evaluative feedback Advances in behavior therapy. New York: Aca- in the classroom and (III) An experimental demic Press, 1968. analysis. Unpublished manuscript, University of Clark, K. B., & Clark, M. P. The development of Illinois at Urbana-Champaign, 1976. consciousness of self and the emergence of racial 11. Dweck, C. S., Goetz, T., & Strauss, N. Sex dif- identification in Negro preschool children. Journal ferences in learned helplessness: IV. An experi- of Social Psychology, 1939,10, 591-599. mental and naturalistic study of failure general- Cohen, S., Rothbart, M., & Phillips, S. Locus of ization and its mediators. Unpublished manu- control and the generality of learned helplessness script, University of Illinois at Urbana-Cham- in humans. Journal of Personality and Social Psy- paign, 1977. chology, 1976, 34, 1049-1056. 12. Garber, J., & Hollon, S. Depression and the ex- Cole, C. S., & Coyne, J. C. Situational specificity of pectancy of success for self and for others. Un- laboratory-induced learned helplessness. Journal of published manuscript, University of Minnesota, Abnormal Psychology, 1977, 86, 61S-623. 1977. Douglas, D., & Anisman, H. Helplessness or expecta- 13. Radloff, L. S. Sex differences in helplessness— tion incongruency: Effects of aversive stimulation with implication for depression. Unpublished on subsequent performance. Journal of Experi- manuscript, Center for Epidemiologic Studies, mental Psychology: Human Perception and Per- National Institute of Mental Health, 1976. formance, 1975,1, 411-417. Dweck, C. S. The role of expectations and attribu- tions in the alleviation of learned helplessness. Journal of Personality and Social Psychology, References 1975, 31, 674-685. Dweck, C. S. Children's interpretation of evaluative Abramowitz, S. I. Locus of control and self-reported feedback: The effect of social cues on learned help- depression among college students. Psychological lessness. In C. S. Dweck, K. T. Hill, W. H. Reed, Reports, 1969, 25, 149-150. W. M. Steihman, & R. D. Parke. The impact of Abramson, L. Universal versus personal helplessness: social cues on children's behavior. Merrill-Palmer An experimental test of the reformulated theory of Quarterly, 1976, 22, 83-123. 72 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE

Dweck, C. S., & Reppucci, N. D. Learned helplessness Jenkins, H. M., & Ward, W. C. Judgment of con- and reinforcement responsibility in children. Jour- tingency between responses and outcomes. Psy- nal of Personality and Social Psychology, 1973, 25, chological Monographs, 1965, 79(1, Whole No. 109-116. 594). Ellis, A. Reason and emotion in psychotherapy. New Kelley, H. H. Attribution theory in social psychol- York: Lyle Stuart, 1962. ogy. In D. Levine (Ed.), Nebraska Symposium on Festinger, L. A theory of social comparison pro- Motivation (Vol. 15). Lincoln: University of Ne- cesses. Human Relations, 1954, 7, 117-140. braska Press, 1967. Fosco, F., & Geer, J. H. Effects of gaining control Klein, D. C., Fencil-Morse, E., & Seligman, M. E. over aversive stimuli after differing amounts of no P. Learned helplessness, depression, and the at- control. Psychological Reports, 1971, 29, 1153-1154. tribution of failure. Journal of Personality and So- Freud, S. Mourning and melancholia. In J. Strachey cial Psychology, 1976, 33, 508-516. (Ed. and trans.), Standard edition of the complete Klein, D. C., & Seligman, M. E. P. Reversal of per- psychological works of Sigmund Freud (Vol. 14). formance deficits in learned helplessness and de- London: Hogarth Press, 1957. (Originally pub- pression. Journal of Abnormal Psychology, 1976, lished, 1917.) 85, 11-26. Frumkin, K., & Brookshire, K. H. Conditioned fear Klinger, E. Consequences of commitment to and dis- training and later avoidance learning in goldfish. engagement irom incentives. Psychological Review, Psychonomic Science, 1969, 16, 159-160. 1975, 82, 1-25. Gatchel, R. J., & Proctor, J. D. Physiological cor- Krantz, D. S., Glass, D. C., & Snyder, M. L. Help- relates of learned helplessness in man. Journal of lessness, stress level, and the coronary prone be- Abnormal Psychology, 1976, 85, 27-34. havior pattern. Journal of Experimental Social Glass, D. C., Reim, B., & Singer, J. R. Behavioral Psychology, 1974, 10, 284-300. consequences of adaptation to controllable and un- Lament, J. Depression, locus of control, and mood controllable noise. Journal of Experimental Social response set. Journal of Clinical Psychology, 1972, Psychology, 1971, 7, 244-257. 28, 342-345. Glass, D. C., & Singer, J. E. Urban stress: Experi- Maier, S. F., Albin, R. W., & Testa, T. J. Failure to ments on noise and social stressors. New York: learn to escape in rats previously exposed to in- Academic Press, 1972. escapable shock depends on nature of escape re- Golin, S., & Terrell, F. Motivational and associative sponse. Journal of Comparative and Physiological aspects of mild depression in skill and chance. Jour- Psychology, 1973, 85, 581-592. nal of Abnormal Psychology, 1977, 86, 389-401. Maier, S. F., & Seligman, M. E. P. Learned helpless- Goodkin, F. Rats learn the .relationship between re- ness: Theory and evidence. Journal of Experi- sponding and environmental events: An expansion mental Psychology: General, 1976,105, 3-46. of the learned helplessness hypothesis. Learning and Maier, S. F., Seligman, M. E. P., & Solomon, R. L. Motivation, 1976, 7, 382-393. Pavlovian fear conditioning and learned helpless- Griffiths, M. Effects of noncontingent success and ness. In B. A. Campbell & R. M. Church (Eds.), failure on mood and performance. Journal of Per- Punishment. New York: Appleton-Century-Crofts, sonality, 1977, 45, 442-457. 1969. Hammen, C. L., & Krantz, S. Effect of success and Maier, S. F., & Testa, T. J. Failure to learn to failure on depressive .cognitions. Journal of Ab- escape by rats previously exposed to inescapable normal Psychology, 1976, 85, 577-586. shock is partly produced by associative interfer- Hanusa, B. H., & Schulz, R. Attributions! mediators ence. Journal of Comparative and Physiological of learned helplessness. Journal of Personality and Psychology, 1975, 88, 554-564. Social Psychology, 1977, 35, 602-611. Masserman, J. H. The principle of uncertainty in Heider, F. The psychology of interpersonal relations. neurotigenesis. In H. D. Kimmel (Ed.), Experi- New York: Wiley, 1958. mental psychopathology. New York: Academic Hiroto, D. S. Locus of control and learned helpless- Press, 1971. ness. Journal of Experimental Psychology, 1974, McNitt, P. C., & Thornton, D. W. Depression and 102, 187-193. perceived reinforcement: A reconsideration. Jour- Hiroto, D. S. & Seligman, M. E. P. Generality of nal of Abnormal Psychology, 1978, 87, 137-140. learned helplessness in man. Journal of Personality Miller, W. R., & Seligman, M. E. P. Depression and and Social Psychology, 1975, 31, 311-327. the perception of reinforcement. Journal of Ab- James, W. H. Internal versus external control of re- normal Psychology, 1973, 82, 62-73. inforcement as a basic variable in learning theory. Miller, W. R., & Seligman, M. E. P. Depression and Unpublished doctoral dissertation, Ohio State Uni- learned helplessness in man. Journal of Abnormal versity, 1957. Psychology, 1975, 84, 228-238. James, W. H., & Rotter, J. B. Partial and one hun- Miller, W. R., & Seligman, M. E. P. Learned help- dred percent reinforcement under chance and skill lessness, depression, and the perception of rein- conditions. Journal of Experimental Psychology, forcement. Behaviour Research and Therapy, 1976, 1958, S5, 397-403. 14, 7-17. CRITIQUE AND REFORMULATION 73

Miller, W. R., Seligman, M. E. P., & Kurlander, H. velopment, and death. San Francisco: Freeman, M. Learned helplessness, depression, and anxiety. 1975. Journal of Nervous and Mental Disease, 1975, 161, Seligman, M. E. P., & Beagley, G. Learned helpless- 347-357. ness in the rat. Journal oj Comparative and Physi- Morse, S., & Gergen, K. J. Social comparison, self- ological Psychology, 1975, 88, 534-541. consistency, and the concept of self. Journal of Seligman, M. E. P., Klein, D. C., & Miller, W. R. Personality and Social Psychology, 1970, 16, 148- Depression. In H. Leitenberg (Ed.) Handbook oj 156. behavior modification and behavior therapy. Engle- Overmier, J. B., & Seligman, M. E. P. Effects of in- wood Cliffs, N.J.: Prentice-Hall, 1976. escapable shock upon subsequent escape and avoid- Seligman, M. E. P., & Maier, S. F. Failure .to escape ance learning. Journal of Comparative and Physio- traumatic shock. Journal oj Experimental Psychol- logical Psychology, 1967, 63, 28-33. ogy, 1967, 74, 1-9. Padilla, A. M. Effects of prior and interpolated shock Seligman, M. E. P., Maier, S. F., & Geer, J. The al- exposures on subsequent avoidance learning by leviation of learned helplessness in the dog. Journal goldfish. Psychological Reports, 1973, 32, 451^56. of Abnormal and Social Psychology, 1968, 73, 256- Padilla, A. M., Padilla, C., Ketterer, T., Si Giacolone, 262. D. Inescapable shocks and subsequent avoidance Seligman, M. E. P., Maier, S. F., & Solomon, R. L. conditioning in goldfish (Carrasius auratus). Psy- Unpredictable and uncontrollable aversive events. chonomic Science, 1970, 20, 295-296. In F. R. Brush (Ed.) Aversive conditioning and Phares, E. J. Expectancy change in chance and skill learning. New York: Academic Press, 1971. situations. Journal of Abnormal and Social Psy- Seligman, M. E. P., Rosellini, R. A., & Kozak, M. chology, 1957,54,339-342. Learned helplessness in the rat: Reversibility, time Racinskas, J. R. Maladaptive consequences oj loss or course, and immunization. Journal oj Compara- lack of control over aversive events. Unpublished doctoral dissertation, Waterloo University, Ontario, tive and Physiological Psychology, 1975, 88, 542- Canada, 1971. 547. Rizley, R. Depression and distortion in the attribu- Seward, J. P., & Humphrey, G. L. Avoidance learn- tion of causality. Journal of Abnormal Psychology, ing as a function of pretraining in the cat. Journal 1978, 87, 32-48. of Comparative and Physiological Psychology, Rodin, J. Density, perceived choice, and response to 1967, 63, 338-341. controllable and uncontrollable outcomes. Journal Shaw, B. F. Comparison of cognitive therapy and oj Experimental Social Psychology, 1976, 12, 564- behavior therapy in the treatment of depression. 578. Journal oj Consulting and 'Clinical Psychology, Rosenberg, M. Society and the adolescent self-image. 1977, 45, 543-551. Princeton, N.J.: Princeton University Press, 1965. Solomon, R. L., & Corbit, J. D. An opponent-process Roth, S. The effects of experimentally induced ex- theory of motivation: II. Cigarette addiction. pectancies oj control: Facilitation of controlling Journal oj Abnormal Psychology, 1973, 81, 158- behavior or learned helplessness? Unpublished doc- 171. toral dissertation, Northwestern University, 1973. Teasdale, J. D. Effects of real and recalled success on Roth, S., & Bootzin, R. R. Effects of experimentally learned helplessness and depression. Journal of Ab- induced expectancies of external control: An in- normal Psychology, 1978, 87, 155-164. vestigation of learned helplessness. Journal oj Per- Tennen, H., & Eller, S. J. Attributional components sonality and Social Psychology, 1974, 29, 253-264. of learned helplessness and facilitation. Journal oj Roth, S., & Kubal, L. Effects of noncontingent rein- Personality and Social Psychology, 1977, 35, 265- forcement on tasks of differing importance: Fa- 271. cilitation and learned helplessness. Journal oj Per- Thomas, E., & Dewald, L. Experimental neurosis: sonality and Social Psychology, 1975, 32, 680-691. Neuropsychological analysis. In J. D. Maser & Rotter, J. B. Generalized expectancies for internal M. E. P. Seligman (Eds.), Psychopathology: Ex- versus external control of reinforcement. Psycho- perimental models. San Francisco: Freeman, 1977. logical Monographs, 1966, 80(1, Whole No. 609). Thornton, J. W., & Jacobs, P. D. Learned helpless- Rotter, J. B., Liverant, S., & Crowne, D. P. The ness in human subjects. Journal oj Experimental growth and extinction of expectancies in chance Psychology, 1971, 87, 369-372. controlled and skilled .tasks. Journal of Psychol- Thornton, J, W., & Powell, G. D. Immunization and ogy, 1961, 52, 161-177. alleviation of learned helplessness in man. American Rush, A. J., Beck, A. T., Kovacs, M., & Hollon, S. Journal oj Psychology, 1974, 87, 351-367. Comparative efficacy of cognitive therapy and Weiner, B. Theories oj motivation: From mechanism pharmacotherapy in the treatment of depressed out- to cognition. Chicago: Rand McNally, 1972. patients. Cognitive Therapy and Research, 1977, 1, Weiner, B. (Ed.). Achievement motivation and at- 17-37. tribution theory. Morristown, N.J.: General Learn- Seligman, M. E. P. Learned helplessness. Annual Re- ing Press, 1974. view of Medicine, 1972, 23, 407-412. Weiner, B., Frieze, I., Kukla, A., Reed, L., Rest, S., Seligman, M. E. P. Helplessness: On depression, de- & Rosenbaurn, R. M. Perceiving the causes oj sue- 74 L. ABRAMSON, M. SELIGMAN, AND J. TEASDALE

cess and failure. Morristown, N.J.: General Learn- Wortman, C. B., & Brehm, J. W. Responses to un- ing Press, 1971. controllable outcomes: An integration of reactance Weiner, B., Heckhausen, H., Meyer, W., & Cook, R. theory and the learned helplessness model. In L. E. Causal ascriptions and achievement behavior: Berkowitz (Ed.), Advances in experimental social A conceptual analysis of locus of control. Journal psychology (Vol. 8). New York: Academic Press, of Personality and Social Psychology, 1972, 21, 1975. 239-248. Wortman, C. B., Panciera, L., Shusterman, L., & Welker, R. L. Acquisition of a free operant appetitive Hibscher, J. Attributions of causality and reac- response in pigeons as a function of prior experi- tions to uncontrollable outcomes. Journal of Ex- ence with response-independent food. Learning perimental Social Psychology, 1976,12, 301-316. and Motivation, 1976, 7, 394-405. Willis, M. H., & Blaney, P. H. Three tests of the learned helplessness model of depression. Journal Received August IS, 1977 of Abnormal Psychology, 1978, 87, 131-136. Revision received October 27, 1977 •

Instructions to Authors Authors should prepare manuscripts according to the Publication Manual of the Ameri- can Psychological Association (2nd ed.). Instructions on tables, figures, references, metrics, and typing (all copy must be double-spaced) appear in the Manual. Also, all manuscripts must include an abstract of 100-175 words typed on a separate sheet of paper. Authors should submit manuscripts in triplicate and should keep a copy to guard against loss. Authors submitting a manuscript previously considered for publication in another APA journal should inform the Editor, who will then seek to obtain independent reviews, thus avoiding the possibility of repeated reviewing by the same consultant. When statistical anal- yses such as ANOVA or chi-square are used, appropriate tables should accompany the manuscript. These will not ordinarily be printed but are for the convenience of the editorial consultants to help in evaluating the article. Therefore, when significant differences are reported, the appropriate F, df, p, and other values should be inserted in the text.