Overall, therefore, there is some evidence that attributional style can predict future negative mood, as illustrated in Figure 10.2. How
Figure 10.2 Relationships between negative events, attributional style and dysphoria.
ever, we are not therefore entitled to conclude that attributional style is the sole, or even most important, psychological characteristic implicated in mood disorders. Nor should we rush to accept a simple stress-vulnerability model of dysphoria of the kind proposed by Seligman. Those who have debated the causal status of attributions have often assumed that attributional style is either a trait (an enduring personality characteristic) or a state (a transient phenomenon, perhaps a symptom of negative mood). As we will see shortly, the truth may be more complex than this.
The Self Worth Living with
A pessimistic appraisal of events probably has an impact on mood for several reasons. Aaron Beck, it will be recalled, suggested that people who feel depressed harbour negative thoughts about themselves, the world and the future. Lyn Abramson has focused on the last element in this negative cognitive triad, arguing that excessively stable and global attributions for negative events lead to a sense of hopelessness – a pervasive conviction that life cannot get better which in turn saps the individual’s motivation to cope with adversity.25 It is certainly true that many depressed people experience a profound sense of hopelessness about the future, and the greater this sense of hopelessness the greater the likelihood that they will attempt suicide.26 (Hopelessness has also been found to be a strong predictor of suicidal tendencies in schizophrenia patients.)27 However, at least as important must be the impact of pessimistic attributions on the first element in Beck’s triad: the individual’s view of the self.
In studies by Howard Tennen and his colleagues at the University of Connecticut, it was found that a pessimistic attributional style predicted low self-esteem better than it predicted scores on a more general measure of depressive symptoms.28 More recently, David Romney of the University of Calgary used a complex statistical technique known as path analysis to explore specific relations between attributional style, self-esteem and depressive symptoms. He found that the attributional dimensions of internality, stability and globalness did not affect depressive symptoms directly, but that each had an impact on self-esteem, which in turn had a strong impact on depression.29 On this evidence, then, pessimistic appraisals of events lead to problems of self-esteem, which in turn lead to negative mood.
How appraisals influence the self
In Chapter 8 we saw that the self is a dynamic, multifaceted phenomenon that is unlikely to be adequately captured by a single dimension varying between the extremes of grandiosity and self-loathing. Indeed, as British psychiatrist Philip Robson has lamented, self-esteem is ‘an idea rather than an entity and the term signifies different things to different people’.30 One problem is that it is not immediately obvious how we go about making global assessments of our self-worth. For example, it seems likely that positive self-esteem (the sum of all the good things an individual thinks about himself) and negative self-esteem (the sum of all the bad things that the individual thinks about himself) are not necessarily opposites – it is as easy to imagine that someone holds both extreme positive and negative opinions about herself as it is to imagine that someone holds exclusively positive or negative opinions, and, indeed, these types of beliefs can be measured separately.31 Although most ordinary people seem to hold beliefs about the self that are so predominantly positive that any objective assessment would deem them unrealistic, people who score low on self-esteem questionnaires rarely harbour exclusively negative opinions but, rather, tend to believe a mixture of positive and negative things about themselves.32
The specific beliefs about the self that we are immediately aware of have been described as the actual self. The depressed patient’s tendency to include more negative attributes in this description than ordinary people can be easily demonstrated by means of questionnaires, but it has also been shown using the self-reference effect (the general tendency to recall information that is directly relevant to the self), which I described in Chapter 8. When given a list of trait words that are either negative (for example, ‘stupid’, ‘unloved’, ‘weak’) or positive (‘successful’, ‘dynamic’, ‘confident’), and asked to say whether they are self-descriptive, unipolar depressed patients not only endorse more negative words as true of themselves in comparison with ordinary people, but later recall more of the negative words if given a surprise memory test.33 Helen Lyon, Mike Startup and I recently found exactly the same result with bipolar patients who were currently depressed.34
Of course, beliefs about the self are not always present in our minds. For much of the time, we direct our thoughts towards other things. Moreover, it appears that some people’s beliefs about themselves are more subject to change than others. When social psychologist Michael Kernis of the University of Georgia asked ordinary people to record their current feelings about themselves at random intervals over an extended period (usually about a week) he found that self-esteem remained relatively stable in some people whereas, in others, it fluctuated dramatically.35 Surprisingly, stability of self-esteem seems to be relatively independent of average level of self-esteem (that is, some people with unstable self-esteem have high self-esteem on most days whereas others have low self-esteem on most days). Kernis found that individuals with unstable self-esteem are more vulnerable to feelings of depression than individuals with stable self-esteem. He also found that individuals with unstable but typically high self-esteem are especially likely to react angrily to negative information about themselves.
Beliefs about the self change over time because they have to be provoked into consciousness (‘activated’ or ‘primed’ in the language of cognitive psychology) before they can affect us. This is why appraisals have such a powerful effect on mood – because they determine whether or not the events we experience activate positive or negative representations from our reservoir of stored knowledge and beliefs about ourselves. Not surprisingly, stability of self-esteem seems to be, at least in part, determined by the extent to which people make extreme appraisals for everyday events; individuals who make more extreme appraisals tend to have very unstable self-esteem.36 However, our beliefs about ourselves are also dependent on the information about ourselves that we have acquired and stored away earlier in our lives. The more negative information about ourselves we have stored in our long-term memory, the more vulnerable we will be to experiencing negative thoughts and ideas about ourselves. Conversely, if we have learned only positive things about ourselves, it is unlikely that even the most challenging of experiences will persuade us to think about ourselves in a very self-critical way.
If this account is correct, individuals who are depressed should have an abnormal tendency to recall negative information about themselves. This kind of memory bias has long been recognized by clinicians, for example Aaron Beck,37 who have observed that unipolar depressed patients organize their memories in a ‘depressogenic fashion’. In an early empirical study, G. W. Lloyd and W. A. Lishman at the Institute of Psychiatry in London asked depressed people and non-depressed people to think of either a pleasant or an unpleasant memory when they heard a series of cue words, and to signal to the researchers when suitable memories came to mind. They found that depressed patients tended to recall negative events more rapidly than positive events, but that this was not the case for non-depressed controls.38 This finding has been replicated many times. For example, John Teas-dale and Sarah Fogarty at Oxford University experimentally induced negative moods in normal volunteers by asking them to read depressing statements and found that they were slower to recall positive events as a consequence.39
Standards and discrepancies
Even if the account of the self I have so far given seems half plausible, a moment’s reflection will reveal that it must be incomplete. Although we have assumed that the discovery of undesirable aspects of the self will provoke fee
lings of distress, we have not explained how individuals know that particular attributes are undesirable. After all, it is quite possible to imagine a person who knows that he is vicious and heartless, but who believes that these characteristics are valuable, or at least acceptable (the Nazis deliberately cultivated this kind of self-image, and hence were able to commit atrocities against civilians while remaining emotionally unscathed).40 We therefore need to consider not only the actual self, but also the internal standards against which the self is evaluated.*
American social psychologist E. Tory Higgins of New York University has pointed out that most of us can describe how we think we would like to be (the ideal self) and how we think we ought to be (the ought self).41 According to Higgins, these self-standards help us to define our goals in life and are sources of two different types of motivation – the striving for desired outcomes in the case of the ideal self and the avoidance of undesired outcomes in the case of the ought self.† To further complicate matters, Higgins also points out that we can consider the self from different perspectives. For example, I can imagine how my mother thinks I actually am (the ‘mother-actual self’), how she would like me to be (the ‘mother-ideal self’) and how she thinks I ought to be (the ‘mother-ought self’). (Of course, I cannot read my mother’s mind to find out what she really thinks about me, but have to guess on the basis of the quality of our relationship.) In principle, there are as many perspectives on the self as there are people whose opinions we can imagine.
You can get a snapshot portrait of your actual self by simply writing down the first ten words that come to mind when thinking about ‘myself as I actually am’. Similarly, a snapshot of the ideal self can be recorded on paper by writing down ten words that come to mind when you think of ‘myself as I would like to be’. Next try ‘myself as I ought to be’, ‘how my mother/father thinks I am’, ‘how my mother/father would like me to be’ and so on. When carrying out this exercise, you will probably notice that many of the descriptions are very similar. However, you might also notice that some are different. For example, you might list the word ‘industrious’ when describing your ideal self and its antonym, ‘lazy’, when thinking about your actual self. Higgins refers to these kinds of differences as self-discrepancies, and has argued that it is our awareness of them that provokes negative mood.
A substantial body of research supports this idea. Tim Strauman of the University of Wisconsin has shown that ordinary people when depressed, as well as psychiatric patients with a diagnosis of unipolar depression, report a large number of actual–ideal discrepancies.42 Socially anxious people, on the other hand, typically report substantial discrepancies between the actual self and the ought self. Not surprisingly, given the substantial overlap between depression and anxiety symptoms, most researchers have reported a strong correlation between the two types of discrepancy, leading some to question whether they are qualitatively distinct.43
Strauman and Higgins have also shown that strong negative feelings can be provoked by making people think about self-discrepancies they have revealed at an earlier time.44 Not only does this lead to negative mood, but it also produces quite profound behavioural and physiological effects. For example, when people contemplate actual–ideal discrepancies, their speech becomes momentarily more sluggish, there is a change in the electrical conductivity of the skin (known as an ‘electrodermal response’ and caused by sweating, indicating that the autonomic nervous system is firing up in anticipation of stress) and, most dramatically of all, the efficiency of the immune system (responsible for protecting the individual from infection) is temporarily reduced.
Until recently, no attempt had been made to explore psychotic disorders from the perspective of Higgins’s theory. However, in a recent study Peter Kinderman, Kerry Manson and I measured self-discrepancies in patients who had received a diagnosis of bipolar disorder, and who were divided into three groups – patients who were currently depressed, those who were currently manic and those whose symptoms were in remission. (Patients in this last group had experienced both depressive and manic episodes in the past but were currently well.) Also participating was a group of ordinary people for comparison purposes. As expected, we found that the bipolar-depressed patients, like previously studied unipolar patients, reported considerable discrepancies between their actual selves and their ideal and ought selves. Currently manic patients, on the other hand, reported even fewer discrepancies between their actual selves and their self-standards than either the healthy controls or the patients in remission.45
The quality of ideals
Obviously, the more exacting our self-standards, the more likely it is that we will experience self-discrepancies. Aaron Beck has argued that many of the problems experienced by depressed patients can be attributed to their abnormally perfectionistic standards, which place them at constant risk of experiencing failure. On Beck’s view, such standards may lurk as barely articulated ‘schemas’, of which we are hardly aware until we are reminded that we have failed to meet them.46 Sticking to the terminology we have adopted so far in this chapter, we can say that failure experiences have the potential to activate these schemas, so that they become available in consciousness as representations of the ideal self and the ought self.
To test this idea, many researchers have used the Dysfunctional Attitudes Scale (DAS), a questionnaire that asks people to consider the extent to which they agree with statements such as ‘I should be able to please everybody’ and ‘My value as a person depends greatly on what other people think of me.’47 Patients suffering from depression consistently score highly on this and similar measures. Although many patients who recover from episodes of depression score normally on the DAS,48 there is also evidence that dysfunctional self-standards predict later episodes of mood disorder. In a study carried out by J. Mark Williams and his colleagues, it was observed that recovered patients who still scored highly on the scale were especially likely to relapse.49 More recently, in the Temple–Wisconsin Cognitive Vulnerability to Depression Project, the DAS was used together with the Attributional Style Questionnaire to select individuals who were later proven to be at very high risk of depression.50
Beck has suggested that self-standards can be divided into two main types, reflecting different fundamental needs. Sociotropy refers to a strong need for care and approval from others, whereas autonomy refers to a need for independence and goal-attainment.51 This idea that some people judge their self-worth in terms of the quality of their relationships, whereas others associate self-worth with freedom of choice and the achievement of goals, represents common ground between cognitive theorists such as Beck and some psychologists working from a neoFreudian perspective.52 According to Beck, sensitivity to different life stressors should reflect these different personality characteristics. Individuals high in sociotropy should feel especially miserable when rejected by others, whereas individuals high in autonomy should feel doubt, self-criticism and despondency when failing to live up to their goals and expectations. In general, researchers have found that individuals who are high in sociotropy are especially likely to become dysphoric when faced with interpersonally threatening events. However, the evidence for a specific vulnerability to failure experiences in individuals scoring high on autonomy has been more equivocal.53
The Attribution Self-Representation Cycle
The psychological structure of the self – the way in which different kinds of self-representation interact with each other – is the subject of enduring controversy, so the attempt I have made here to integrate the various findings may be disputed by some of my colleagues. It is, of course, possible that not only the content but also the structure of the self varies across cultures. Perhaps cultural differences help to explain why low self-esteem is less apparent in depressed patients from developing countries than in patients from the West. These doubts notwithstanding, we can take home two simple conclusions from the evidence reviewed thus far. First, most people who are prone to depression (including p
eople diagnosed as suffering from bipolar disorder) evaluate negative events with excessive pessimism. Second, in perhaps the majority of cases, it is the belief that the self is wanting that finally triggers negative mood. This belief, in turn, arises partly because the individual at risk of depression holds unrealistic standards. Building these observations on top of our previous findings (and including the pathway from life events to dysphoria via hopelessness proposed by Lyn Abramson) we arrive at the model of depression illustrated in Figure 10.3.
This model leaves one important question unanswered. Although we have assumed that the pathway from life experiences to dysphoria is mediated by attributions, we have not explained where attributions come from. Strangely, most researchers working on attributional models of psychopathology have neglected this problem. Because attributional style has been assumed to be a stable trait, the mechanisms
Figure 10.3 An elaborated model of dysphoria, showing the contributions of a pessimistic attributional style and dysfunctional self-standards. The pathway to dysphoria mediated by hopelessness, proposed by Abramson, is also sketched in.
involved in generating causal inferences have been almost completely ignored.
We can begin by acknowledging that most events include attributional signposts. More than thirty years ago, the American social psychologist Harold Kelley argued that information about the consistency (has it happened before?) and distinctiveness (does it only happen in particular circumstances?) of events, together with consensus information (does the event affect only me or everyone else?), may be particularly important in influencing causal judgements.54 For example, an apparently random assault experienced by a civilian during warfare may be attributed to causes external to himself because he knows that attacks have been happening for some time (high consistency), occur only when enemy troops are present (high distinctiveness) and are experienced by many other civilians (high consensus). However, even such well signposted events allow some freedom in the kind of attribution generated. Consider the following attempt by a Jewish concentration camp survivor to explain the death of his older brother during the final days of the Second World War:
Madness Explained Page 30