Darley and Batson designed an experiment aimed at uncovering which of these differences might be most relevant to explaining the differences in behavior. Subjects in this experiment were students at Princeton Theological Seminary. As each subject arrived, he was informed that he was to give a talk that would be recorded in another building. Along the way to the place for the talk, the subject encountered a "victim" slumped in a doorway. The question was under what conditions would a subject would stop to help the victim.
Half of the subjects were assigned to talk on the Good Samaritan Parable; the others were assigned a different topic. Some of the subjects were told they were late and should hurry; some were told they had just enough time to get to the recording room; and some were told they would arrive early. Judging by their responses to a questionnaire, they had different religious and moral orientations.
The only one of these variables that made a difference was how much of a hurry the subjects were in. 63% of subjects that were in no hurry stopped to help, 45% of those in a moderate hurry stopped, and 10% of those that were in a great hurry stopped. It made no difference whether the students were assigned to talk on the Good Samaritan Parable, nor did it matter what their religious outlook was.
Standard interpretations of the Good Samaritan Parable commit the fundamental attribution error of overlooking the situational factors, in this case overlooking how much of a hurry the various agents might be in.
1 6 Direct Empirical Challenges to Character Traits
But don't we know from ordinary experience that people differ in character traits? Here it is useful to consider related issues.
Psychoanalysts acquire a considerable experience in treating patients and can cite many instances in which psychoanalytic treatment is successful. However, emprical studies of psychoanalytic treatment as compared with no treatment have found no objective benefit. (Dawes, 1994)
Some diagnosticians have used Rorschach inkblot tests to make psychological diagnoses. It seemed to those using these tests that they had abundant evidence that certain characteristics of the test results were diagnostic of certain disorders. Empirical studies showed there was no correlation between those characteristics and the test results. (Nisbett & Ross, 1980, pp. 93-97)
Many employers are convinced that useful information can be gained from interviewing potential employees. However, for the most part, interviews simply add noise to the decision process. Empirical studies indicate that decisions made on information available apart from an interview are more reliable than decisions made when an interview is added. (Ross & Nisbett, 1991, pp. 136-138)
Discovery of such errors in reasoning has encouraged research into why people are subject to such errors (Tversky & Kahneman, 1974, Nisbett & Ross, 1980). One suggested reason is confirmation bias. Given a hypothesis, one tends to look for confirming evidence. Finding such evidence, one takes it to support the hypothesis. Evidence against the hypothesis tends to be ignored or downplayed.
Ross & Nisbett suggest that the initial source of the fundamental attribution error may have to do with Gestalt considerations of figure and ground. Where we distinguish figure from ground, we pay more attention to figure and less to ground and we try to explain what happens in terms of features of the figure rather than in terms of features of the ground. Typically the actor is figure and the situation is ground, so we seek an explanation of the action in features of the actor in the foreground rather than in features of the background situation. The suggested explanation is then subject to confirmation bias. Additional support comes from the fact that other people give explanations in terms of dispositional features of agents rather than in terms of aspects of their situations.
When investigators have looked for objective evidence that people differ in character traits, the results have been much as with psychoanalysis, Rorschach tests, and interviews. People take themselves to have lots of evidence that agents differ in character traits. Yet empirical studies have not found any objective basis for this confidence. Summarizing a number of studies, Ross & Nisbett (1991, p. 95) report that the "average correlation between different behavioral measures designed to tap the same personality trait (for examples, impulsivity, honesty, dependency, or the like) was typically in the range between .10 and .20, and often was even lower." These are very low correlations, below the level which people can detect. Using such correlations to make predictions yields hardly any improvement over guessing. Even if predictions are limited to people one takes to be quite high on a particular trait, the correlations are still very low.
Ross & Nisbett observe that people have some appreciation of the role of situation in the way they understand such stories as The Prince and the Pauper or the movie Trading Places. But for the most part, people are quick to infer from specific actions to character traits.
It is true that there are better correlations for very specific situations. "Hartshorne and May (1928) found that the tendency to copy from an answer key on a general information test on one occasion was correlated .79 with copying from an answer key on a similar test six months later. Newcomb (1929) found that talkativeness at lunch was a highly stable attribute; it just was not very highly correlated with talkativeness on other occasions..." (Ross & Nisbett, 1991, p. 101).
Surprisingly, Flanagan (1991) argues that this shows there really are character traits, "albeit not traits of unrestricted globality or totally context- independent ones." I guess he means such character traits as "being disposed to copy from an answer key on a certain sort of test" and "being talkative at lunch." But, first, no reason has been given for thinking that these specific narrow regularities in behavior reflect dispositions or habits rather than, for example, skills or strategies that have worked in the past. Second, and more importantly for our purposes, ordinary thinking about personality and character attributes is concerned with more global traits like honesty and talkativeness.
Flanagan concludes: "Yes, there are character traits. The language of character traits picks out psychologically real phenomena." But I do not see that he has cited any empirical evidence for this claim.
Flanagan also seems to think that it inconsistent to argue against character traits by appeal to the fundamental attribution error. He says, "It is telling against the situationalist who is also an eliminativist that he will have extreme difficulty (indeed he courts inconsistency) in positing attributional biases of any sort if by these he means to refer to what he must be taken to want to refer to, namely, dispositions to think in certain ways" (305). But this is true only if a "situationalist" is someone who denies that there are any dispositions at all, or who (perhaps like Skinner, 1974) denies that it is useful to explain anything in terms of dispositions. The issue we have been concerned with is whether people differ in certain particular dispositions--character traits. To deny that people differ significantly in character traits is not to deny that they have any dispositions at all. People might well all share certain dispositions, such as a disposition to make the fundamental attribution error. Secondly, they might differ in various dispositions that do not constitute character traits, such as personality disorders and other mental illnesses. (So, for example, to deny that there are character traits is not to accept the view in Laing,1960, that schizophrenia is simply a rational response to a difficult family situation.)
7 Benefits of Appreciating the Fundamental Attribution Error
There are various benefits to a proper appreciation of ways in which ordinary moral thinking rest on the fundamental attribution error.
1 7.1 Philosophy
7.1.1 Virtue ethics
Character based virtue ethics may offer a reasonable account of ordinary moral views. But to that extent, these ordinary views rest on error.
It is worth mentioning that there are variants of virtue ethics that do not require character traits in the ordinary sense. For example, Thomson (1996) tries to explicate moral thinking by appeal to judgments about whether particular actions are just or courageous or whatever. To the extent that such judgments are concerned entirely with the action and not with any presumed underlying trait of character, Thomson's enterprise is unaffected by my discussion.
Maria Merritt (forthcoming) has been developing a version of virtue ethics that emphasizes the role of the situation in maintaining relevant regularities in behavior.
7.1.2 Better understanding of Moral luck.
Adam Smith (1759) wrote about the influence of fortune on our moral judgments, giving nice examples. Someone carelessly throws a brick over a wall. His companion may complain about this even if no harm is done. But if the brick does hit someone, much greater condemnation ensues. Nagel (1979) gives a similar example of a driver who takes his eyes off the road for a second. That's bad, but suppose in that second a child darts into the street and is hit. Then much worse condemnation seems appropriate.
Smith and Nagel note that from a certain point of view, our moral judgment of the act should be based entirely on the motives of the agent and the agent's epistemic situation, so that from that point of view there should be no difference between two cases that are the same in those respects in one of which someone is hit by the brick (or car) and in the other of which no one is hit. Yet, it is clear that we will judge the cases differently.
Perhaps these are simply further instances of the fundamental attribution error. This bad thing has happened and we attribute it to the bad character of the agent in the foreground.
1 7.2 Real Life7.2.1 Moral Education
If there is no such thing as character, then there is no such thing as character building.
7.2.2 Tolerance
When things go wrong, we typically blame the agent, attributing the bad results to the agent's bad character. Even when things do not go bad, we are quick to interpret actions as expressive of character traits, often hostile traits. For example, a person with poor vision may fail to recognize an acquaintance, who then attributes this to coldness in that person.
A greater understanding of the agent's situation and how it contributed to the action can lead to a greater tolerance and understanding of others.
7.2.3 Better understanding of ethnic hatred
Recent terrible events in the former Yugoslavia are often attributed to historical "ethnic hatreds". Yet it is possible to explain these events in rational terms (Hardin, 1995). Suppose there are limited resources and a successful coalition will benefit its members more than those excluded from the coalition. Such a coalition is possible only if insiders can be distinguished from excluded outsiders and only if it is possible to keep members from defecting to other groups. Coalitions formed around ethnic or religious lines might succeed. The threat that one such coalition may form can lead other groups to form competing coalitions and to struggle against each other. If stakes are high enough, such struggles can become violent. If we attribute the resulting violence to ethnic hatred, we may very well doubt that there is anything we can do. If we understand the way the violence arises from the situation, we may see more opportunities to end the conflict.
1 8 Summary
We very confidently attribute character traits to other people in order to explain their behavior. But our attributions tend to be wildly incorrect and, in fact, there is no evidence that people differ in character traits. They differ in their situations and in their perceptions of their situations. They differ in their goals, strategies, neuoses, optimism, etc. But character traits do not explain what differences there are.
Our ordinary views about character traits can be explained without supposing that there are such traits. In trying to explain why someone has acted in a certain way, we concentrate on the figure and ignore the ground. We look at the agent and ignore the situation. We are naive in our understanding of the way others view a given situation. We suffer from a confirmation bias that leads us to ignore evidence against our attributions of character.
It is very hard to do studies that might indicate whether or not people differ in character traits, but the few studies that have been done do not support this idea. We must conclude that, despite appearances, there is no empirical support for the existence of character traits.
Furthermore, it is clear that ordinary thinking about character traits has deplorable results, leading to massive understanding of other people, promoting unnecessary hostility between individuals and groups, distorting discussions of law and public policy, and preventing the implementation of situational changes that could have useful results.
Bibliography
Aristotle, (1985). Nicomachean Ethics, translated by T. Irwin. Indianapolis, Indiana: Hackett.
Bennett, W. J. (1993). The Book of Virtues. New York: Simon & Schuster.
Brandt, R. B. (1988). "The structure of virtue." Midwest Studies in Philosophy XIII, pp. 64-82.
Darley, J. M., & Batson, C. D. (1973). "`From Jerusalem to Jericho': A Study of Situational and Dispositional Variables in Helping Behavior," Journal of Personality and Social Psychology 27.
Dawes, R. M. (1994). House of Cards. New York: Free Press.
Doris, J. M. (forthcoming). People Like Us: Personality and Moral Behavior. New York: Cambridge University Press.
Fischer, J. M., & Ravizza, M., editors, (1992). Ethics: Problems and Principles. Fort Worth: Harcourt Brace Jovanovich.
Flanagan, O. (1991). Varieties of Moral Personality. Cambridge, Massachusetts: Harvard University Press.
Hardin, R. (1995). One for All: The Logic of Group Conflict. Princeton: Princeton University Press.
Harman, G. (1983). "Human flourishing, ethics, and liberty," Philosophy and Public Affairs 12, pp. 307-322.
Harman, G. (1996). "Moral relativism," part I of Harman & Thomson, 1996.
Harman, G. (forthcoming). "Moral philosophy and linguistics," Proceedings of the 20th World Congress of Philosophy (Philosophy Documentation Center).
Harman, G., & Thomson, J. J. (1996). Moral Relativism and Moral Objectivity. Oxford: Blackwell.
Hartshorne, H., & May, M. A. (1928). Studies in the Nature of Character, I: Studies in Deceit. New York: Macmillan.
Holland, J. H., Holyoak, K. Jl, Nisbett, R. E., & Thagard, P. R. (1986). Induction: Processes of inference, learning and discovery. Cambridge, MA: Bradford Books/M.I.T.
Hursthouse, R. (1996). "Normative virtue ethics." Crisp, R., How Should One Live? Essays on the Virtues. Oxford: Oxford University Press, pp. 19-36
Kagan, S. (1989). The Limits of Morality. Oxford: Oxford University Press.
Laing, R. D. (1960). The Divided Self. Chicago: Quadrangle.
Lewin, K. (1935). Dynamic Theory of Personality. New York: McGraw-Hill.
Merritt, M. (forthcoming). "Virtue Ethics and the Social Psychology of Character," Ph. D. Dissertation, University of California, Berkeley.
Milgram, S. (1963). "Behavioral study of obedience." Journal of Abnormal and Social Psychology 67.
McCloskey, M. (1983). "Intuitive physics." Scientific American 248, pp. 122-130.
Nagel, T. (1979). "Moral luck." In Mortal Questions. Cambridge, England: Cambridge University Press.
Newcomb, T. M. (1929). The Consistency of Certain Extrovert-Introvert Behavior Patterns in 51 Problem Boys. New YOrk: Columbia University Teachers College Bureau of Publications.
Nisbett, R. E., & Ross, L. (1980). Human Inference: Strategies and Shortcomings of Social Judgment. Englewood-Cliffs: Prentice-Hall.
Railton, P. (1997). "Made in the Shade: Moral Compatibilism and the Aims of Moral Theory," Canadian Journal of Philosophy, Supplementary Volume 21.
Ross, L. (1977). "The intuitive psychologist and his shortcomings." In L. Berkowitz (ed.), Advances in Experimental Social Psychology, Vol 10. New York: Academic Press.
Ross, L., & Nisbett, R. (1991). The Person and the Situation: Perspectives of Social Psychology. New York: McGraw-Hill.
Skinner, B. F. (1974). About Behaviorism.. New York: Knopf.
Smith, A. (1759). "Of the Influence of Fortune upon the Sentiments of Mankind, with regard to the Merit or Demerit of Actions." Theory of the Moral Sentiments, Part II, Section III.
Taylor, R. (1991). Virtue Ethics. Interlaken, New York: Linden Books.
Thomson, J. J. (1996). "Evaluatives and directives," chapter 8 of Harman & Thomson, 1996.
Tversky, A., & Kahneman, D. (1974). "Judgment under uncertainty: heuristics and biases," Science 185: 1124-31.
--------------------------------------------------------------------------------
[1 "The relation between lay personology and a more correct theory of personality is analogous to the relation between lay and scientific physics," Ross and Nisbett (1991) pp. 161, citing earlier work including Lewin (1935).
2 An alternative view (Harman, forthcoming) is that children no more require moral instruction in order to acquire morality than they require instruction in their first language in order to acquire that language.
3 But see 7.1.1 below, where I mention two versions of "virtue ethics" that do not treat virtues as traits of character.
