Everyone seems to believe that memory is biased in favor of the self (e.g., Greenwald, 1980), but there are few systematic studies to support this claim. The claim itself is rather vague: What kinds of mnemonic transfor-mations would count as examples? It is often suggested that people re-member "pleasant" experiences more readily than "unpleasant" ones, but early studies of the issue were so plagued by methodological prob-lems that it has been all but abandoned (Rapaport, 1959). Albright's dis-cussion of "subjectivity" in this volume makes it painfully clear that the real biases of memory are far too subtle to be captured by such a crude dichotomy. Another example of the same general claim (that memory is self-serving) is the casually held but common assumption that people remember their own contributions to a conversation better than those of other participants. In an empirical study of memory for seminar dis-cussions, Ira Hyman and I found that this claim is much too simplistic (Hyman & Neisser, 1992).
Recent diary studies of memory have opened up a new approach to this question. In what may be the most systematic of these studies, Willem Wagenaar (1986) did find that events rated as "pleasant" were somewhat better recalled than those rated "unpleasant." This finding seems straightforward enough, but it actually obscures an important distinction. In chapter 10, Wagenaar takes a crucial further step. In addition to classi-fying the events as "pleasant" or "unpleasant," he now also classifies them as "self-related" or "other-related." For example, an event in which he himself behaved badly (in his own judgment) would be scored as "unpleasant/self-related." In contrast, the funeral of a friend's daughter would be "unpleasant/other-related." The results were clear: Unpleas-ant/self-related events were remembered better (not worse) than anything else. Wagenaar's interpretation of these results is discussed by Alan Bad-deley in chapter 12, so I will not consider it here. I cannot help wonder-ing, however, whether his finding is widely generalizable. It may just apply to relatively self-critical individuals, such as Wagenaar himself.
No modern psychologist has addressed the problem of memory distor-tion more systematically than Michael Ross. Ross (1992; see also chap. 11 of this volume) has proposed and tested a very specific theory of one class of memory errors. He suggests that when people are asked to recall some past characteristic or trait - their political opinions 10 years ago, how bad their headaches used to be - they estimate the answer on the basis of two sources of information. The first of these is their present standing on the trait; the second is an implicit theory of how the trait is likely to have changed with time. Asked to recall a former political stance, for example,most people apply a "no-change" theory and remember it as similar to their attitude today. Asked what their reading speed was last month -before the start of the study skills course that they have just now com-pleted - they begin with the implicit hypothesis that the course must have been effective. Using that assumption, they recall their earlier reading speed as lower than the speed they have now.
This idea - that accounts of the past are driven by implicit theories of stability and transformation - has implications for many different do-mains. The abuse narratives discussed above are a case in point: They depend on the theory that adult life problems reflect histories of child-hood victimization. Similar implicit theories appear in other contexts. The unhappy childhoods often recalled by clinically depressed patients, for example, cannot be taken at face value. Ross and Buehler (chap. 11 of this volume) show that folk theories of the origin of depression contrib-ute to such accounts in predictable ways. In a different vein, they also consider the shared recollections of married couples - recollections that often reflect implicit theories of the relationship itself. (One interesting aspect of those theories is the belief that wives have better memories than husbands; cf. Ross & Holmberg, 1990.)
Like Albright, Bruner, and Gergen, Ross and Buehler know that recol-lections are shaped by narrative convention. They are also concerned with the functions of those recollections. For example, why does writing an account of a traumatic event sometimes produce positive psychological effects? Perhaps most interesting of all, they wonder why people find some narratives more believable than others. What makes a story cred-ible? Indeed, why do we credit our own stories in preference to accounts given by others? Ross and Buehler's explanations of this preference seem plausible - not only to me but also to Eugene Winograd, who comments on their contribution in chapter 13. Winograd's thoughtful commentary revisits many of the themes that have been touched on in this introduc-tion, and presents a further example.
www.psychspace.com心理学空间网
