My brother died in my arms from dysentery. He faded away to nothing. A man who was a giant died a skeleton. I held him in my arms when he died. There was just nothing I could do. When I think about it, I sometimes blame myself. He did so much to keep me alive. I feel that had he saved some of that energy for himself he would have had a better chance to survive.55
In this moving example, an internal attribution is made despite apparently overwhelming evidence that the cause should be allocated elsewhere (his German tormentors), presumably because the speaker’s judgement is also influenced by his beliefs about himself (that he was not strong enough). In this way, beliefs about the self form an additional source of information when individuals attempt to determine the causal locus of an event. To take a more trivial example, if an individual already believes himself to be stupid, his failure to pass an exam will almost certainly be attributed internally, because there is an available self-representation that ‘fits’ the event. Hence, people who score low on global self-esteem measures typically make internal attributions for negative events56 whereas those with high self-esteem tend to make internal attributions for their successes.57
Readers who have followed this argument may be puzzled that I am now claiming that beliefs about the self influence attributions, whereas, only a few paragraphs ago, I was arguing that attributions affect beliefs about the self. Before resolving this gallinovular* problem, however, I want to consider two predictions that follow from my account of the way in which attributions are generated. The first concerns the time it takes to think of the likely cause of an event. If we assume that this process involves some kind of mental search that terminates when an appropriate explanation is found, and if we assume that the first step in this process involves searching current beliefs about the self, it follows that ordinary people who believe good things about themselves will more quickly generate attributions for positive events than for negative events (because they will more quickly find some thing good about themselves than something bad). On the other hand, people with psychiatric disorders who harbour negative beliefs about themselves should not show this bias, which may even be reversed if their negative beliefs about themselves outweigh their positive beliefs. This was the rationale for an experiment I carried out with Peter Kinderman and Kim Bowen-Jones in which we asked ordinary people, depressed patients and paranoid patients to read aloud descriptions of hypothetical events before saying what they thought their likely causes would be.58 This method enabled us to measure precisely the interval between the time the participants finished reading each scenario and the time they began to generate an attribution (see Figure 10.4). As expected, the ordinary people more quickly generated internal attributions for positive events than for negative events. The depressed patients, on the other hand, just as quickly generated internal attributions for negative events as for positive events, which is consistent with the idea that they harbour a mixture of positive and negative beliefs about themselves. (I will discuss paranoia in detail in a later chapter; suffice it to say for the moment that the data obtained from this experiment are consistent with the hypothesis that patients who are paranoid, like depressed patients, hold more negative beliefs about themselves than ordinary people.)
The second prediction involves the effects of making people believe bad things about themselves. This can be temporarily achieved in a variety of ways, but one fairly reliable method involves creating a
Figure 10.4 Response times needed to make internal attributions for positive and negative events by paranoid, depressed and normal participants (from Bentall, Kinderman and Bowen-Jones, 1999).
contrived failure experience.* If beliefs about the self influence attributions, people should make more pessimistic attributions about negative events as a consequence. In fact, this effect has been demonstrated by Stanford University psychologist Joseph Forgas in a series of experiments with ordinary people,59 and more recently by Sue Kaney and myself in a study that also included depressed and paranoid patients.60 In our version of the experiment, we asked our participants to make attributions for hypothetical negative events before and after they completed an anagram task, which included a mixture of some questions that could be solved (for example, IEPNCL, which can be solved as PENCIL) and some that could not be solved (for example, LSIDTL). The impact of this experience is shown in Figure 10.5. At the outset, the depressed patients made more internal attributions for hypothetical negative events than the ordinary people who acted as controls. The paranoid patients, on the other hand, made more external attributions, a phenomenon that we will discuss in detail later. However, both psychiatric groups made more internal attributions following the contrived failure experience than beforehand, presumably because this experience activated negative self-schemas, which in turn influenced
Figure 10.5 Normal, depressed and paranoid individuals’ internality scores for hypothetical negative events before and after completing the insoluble anagrams task (from Bentall and Kaney, in submission).
their subsequent attributions. In this study, no change was seen in the attributional responses of the healthy controls, undoubtedly because they were less readily influenced by the failure experience than the two psychiatric groups. However, Forgas was able to produce a pessimistic shift in the attributions of ordinary people, presumably because they used a more potent contrived failure experience.
I will have more to say about the psychological mechanisms involved in generating causal explanations when I discuss the role of attributions in paranoid beliefs. For the moment, I want to point out an important implication of what we have discovered. It seems that attributions not only influence self-representations, but that changes in self-representations affect future attributions. The two processes are cyclically coupled in an attribution-self-representation cycle, so changes in either inevitably lead to changes in the other.
As a consequence, an individual’s psychological response to a negative experience is likely to produce, not only an immediate change in mood, but also longer-term changes that may affect the way in which she appraises negative experiences in the future. Indeed, this cycle, shown in Figure 10.6, appears to have all the hallmarks of a dynamic and non-linear system, in which the various components (beliefs about the self, attributions) are likely to fluctuate, perhaps dramatically and unpredictably, over time.
If this account of the processes involved in generating attributions is
Figure 10.6 The attribution-self-representation cycle (highlighted).
correct, fluctuations in attributional responses and beliefs about the self should be observable in real life. In fact, we have already seen that this is so. Remember that, in studies in which attributions were extracted from recorded speech, it was discovered that Lyndon Johnson’s decisions during the Vietnam War were predicted by his current attributional style,61 and that changes in attributions during psychotherapy predicted changesinmood.62 The implication of these findings is that (in the case of American presidents and psychotherapy patients, at least) attributional biases change detectably over time. Similarly, we have also seen that, even in ordinary people, self-representations (at least as reflected by global self-esteem measures) sometimes fluctuate quite dramatically.63
The theory of the attribution–self-representation cycle helps us to understand better two observations noted earlier in this chapter. First, recall that most people take greater credit for success than failure. The self-serving bias seems to be one of a number of self-enhancing strategies that most of us employ in our everyday lives.64 (Another concerns the kinds of attributes we value most highly. People usually value those attributes on which they know themselves to excel. On the other hand, attributes that we do not possess are usually dismissed as relatively unimportant – overweight businessmen are rarely distressed by their inability to run marathons.) Ordinary people, it seems, actively strive to hold good opinions about themselves.
These strategies – which all involve self-serving appraisals – act as homeos
tatic mechanisms that maintain positive beliefs about the self in all but the most challenging conditions. Not surprisingly, research has shown that we tend to resort to these strategies when we feel that we are under threat,65 when pressures to maintain self-esteem are particularly intense,66 and when we are in the presence of other people who are relatively unknown to us and about whose opinions we are uncertain.67 Without self-serving appraisals, negative events would always lead to more negative beliefs about the self, and these negative beliefs would lead to more negative appraisals in the future–ultimately culminating in a horrible spiral of increasing dysphoria.
Of course, this is exactly what seems to happen in people who are vulnerable to depression. Their pessimistic appraisals of negative events lead to negative beliefs about the self, which increase the probability that they will make further pessimistic appraisals in the future. They have no psychological defences. Hence, when suffering the slings and arrows of outrageous fortune, their self-esteem goes into free-fall.
This effect helps us to understand an unresolved problem about the role of attributions in depression. Remember that the evidence that a pessimistic attributional style always precedes depression is rather weak. On the one hand, the results from the Temple–Wisconsin Cognitive Vulnerability to Depression Project demonstrated fairly clearly that a pessimistic attributional style confers vulnerability to future dysphoria. On the other hand, many longitudinal studies have revealed at best equivocal evidence that a pessimistic style precedes the onset of low mood in the majority of people who become depressed. The account I have just given of the attribution–self-representation cycle neatly accommodates these apparently contradictory findings. As negative experiences will tend to increase the magnitude of a vulnerable individual’s pessimism, highly pessimistic attributions need not be present at the onset of the descent into depression – it is sufficient that the individual’s self-serving bias is not quite strong enough to counter the damage to self-esteem inflicted by a distressing life event.
It’s Not What You Feel, it’s the Way that You Feel it
At this point I would like to introduce a final set of concepts that will help us understand the link between appraisals and negative mood, and which will be useful when we come to examine other symptoms later on. So far, I have focused almost exclusively on psychological processes that seem to play a role in the onset of dysphoria. However, processes that maintain a dysphoric state may be just as important as processes that create a negative mood in the first place. Most of us become depressed occasionally; factors that determine the rapidity with which we bounce back may separate those of us who react to adversity with extreme distress from those of us who do not.
This insight leads to two simple ideas that we will encounter so often in later chapters that, together, they may well constitute a fundamental principle of psychopathology. First, people who are experiencing psychological distress are rarely passive victims of their emotional turmoil; rather, they usually make active attempts to cope with it. Second, whereas some of these attempts may successfully ameliorate distress, other coping strategies may have the unfortunate effect of increasing the likelihood that the distress will persist. Stating these ideas formally as the reaction–maintenance principle:
People do not react passively to distressing behaviours and experiences; the way in which they react plays a role in determining the duration and nature of these behaviours and experiences.
It is probable that there are many different ways in which we can respond to the experience of negative mood. However, two particularly important coping strategies have been identified by psychologist Susan Nolen-Hoeksema of the University of Michigan.68 Some people, it seems, respond by ruminating; that is, by engaging in thoughts and behaviours that focus the individual’s mind on their distress. People with this style of coping spend much of their time thinking about how badly they feel and pondering such questions as, ‘Why am I in such a mess?’ and ‘Will I ever feel better?’ Others, however, distract themselves by focusing their attention away from their unhappiness and its causes and on to pleasant or neutral stimuli that are engaging enough to prevent their thoughts returning to the source of their distress. For example, they may throw themselves into their work, or seek out pleasant social activities. Like all good research psychologists, Nolen-Hoeksema has developed her own assessment tool, the Response Styles Questionnaire (RSQ), which measures the tendency to use these strategies. (Early versions of the questionnaire also measured two other strategies – the tendency to engage in active problem solving and the tendency to rush into dangerous activities such as reckless driving or the taking of illicit drugs. These subscales were abandoned, the former because it correlated highly with distraction strategies and the latter because–surprisingly – itcorrelated highly with rumination. However, other researchers, unconnected to Nolen-Hoeksema, have since shown that depressed patients, whether diagnosed as unipolar or bipolar, perform poorly on measures of problem-solving ability.)69
That ruminative coping leads to prolonged periods of depression was demonstrated by Nolen-Hoeksema first in simple experiments in which depressed mood was induced in ordinary people by asking them to read depressing stories while listening to sad music. The participants were then instructed either to ruminate or to engage in distracting activities. Those who ruminated experienced longer and deeper periods of dysphoria. In these experiments, it was also shown that dysphoric participants who ruminated, compared to those who distracted, recalled more negative memories about their lives,70 would interpret ambiguous situations more negatively, were more pessimistic about the future, and were less able to propose effective solutions to interpersonal problems.71 In other words, rumination seemed to exacerbate processes which, we have already seen, are thought to be important determinants of dysphoric mood.
A unique opportunity to test the impact of the ruminating and distracting strategies was presented by the Loma Prieta earthquake, which hit the San Francisco Bay area on 17 October 1989, with a force measuring 7.1 on the Richter scale. Sixty-two people were killed in the disaster, nearly 4000 were injured, and 12,000 people were left homeless. Horrific pictures were played on news programmes over the next few days, most notably of the upper deck of a two-deck highway in Oakland, which had collapsed on to the lower deck, killing the occupants of cars travelling below. By chance, Nolen-Hoeksema had administered the RSQ to a class of Stanford University students just two weeks earlier.72 She had also interviewed the students to determine whether they were experiencing symptoms of depression. Administering follow-up measures at ten days and at seven weeks after the earthquake, she found that those students who had been most exposed to danger, and who reported ruminating about the earthquake, were more likely to feel depressed afterwards. More importantly, she found that those students who were already depressed or suffering from stress symptoms, and who also reported a ruminative coping style on the RSQ, were especially likely to report depressive symptoms.
These findings have been confirmed in a more recent prospective study conducted under less dramatic circumstances by Nancy Just and Lauren Alloy, who found that reports of a ruminative response style in non-depressed college students predicted their later experience of dysphoric episodes over an eighteen-month follow-up period.73 Nor have these kinds of observations been confined to samples of college students: Nolen-Hoeksema showed that a ruminative coping style predicted longer and more severe periods of depressed mood in gay men whose partners had died of AIDS,74 and in the family members of recently deceased cancer victims.75
Overall, the weight of evidence about the impact of coping styles collected by Nolen-Hoeksema and her colleagues is impressive. However, two important questions remain to be answered. First, given the objections to the idea of basic emotions that I raised in the last chapter, it might be expected that rumination will affect a wide range of negative emotional states. This seems to be the case: Nolen-Hoeksema has recently shown that rumination can prolong episodes of anger76 and als
o episodes of anxiety.77 Second, although the negative impact of rumination has been well demonstrated, the effects of distraction remain relatively unexplored. We will return to this question in a later chapter.
Rhythms and Blues: from Fatigue to Depression
One obvious problem with the account I have so far given of the pathway from negative events to negative mood is that not everyone who is said to be depressed experiences negative beliefs about the self; indeed, we have seen that such beliefs may be relatively uncommon in some regions of the world. It is therefore time for me to make good my earlier promise to describe a second pathway that leads from negative experiences to the cluster of loosely related symptoms that we call ‘depression’. In the process, I will attempt to explain how somatic symptoms such as loss of appetite might fit into the general framework I have constructed. We will begin by briefly plunging into an area of research that, on first sight, appears to be completely unconnected to everything that we have discussed so far.
Madness Explained Page 31
