The Nonexistence of Character Traits

Gilbert Harman
Princeton University

(View this page in Romanian courtesy of azoft web)

Athanassoulis (1999) objects as follows to Harman (1999). `What the Milgram experiments challenge is not the assumption that people can have character traits, but rather the assumption that most people will act compassionately under pressure'.

However, although the Milgram experiment does not by itself challenge the assumption that subjects of the experiment have robust character traits, it does illustrate the tendency of observers to infer wrongly that actions are due to distinctive robust character traits rather than to aspects of the situation. In other words, it illustrates the way in which observers are subject to a `fundamental attribution error.' Furthermore, it is only one illustration. The psychological literature contains all sorts of other examples, as is evident from any contemporary textbook in social psychology, e.g. Ross and Nisbett, 1991.

This raises the question whether there is any evidence that people differ in character traits. One might suppose that such differences are evident in ordinary experience. But these ordinary opinions can be completely explained without any supposition that there really are character traits, as noted in Harman (1999), summarizing Ross and Nisbett (1991). Furthermore, studies of actual individual differences do not support ordinary assumptions about character traits.

In other words, although it may seem perfectly obvious, at least to someone who is unfamiliar with social psychology, that people differ in character traits, such an opinion is evidentially on a par with the opinion of a practicing psychoanalyst about the therapeutic value of psychoanalysis or the opinion of an employer that it is obvious that interviews improve hiring decisions. Such opinions are firmly held quite independently of their truth (they are known to be false) and can be explained in terms of confirmation biases of various sorts. Similarly for ordinary opinions about character traits. There is no reason at all to believe in character traits as ordinarily conceived.

Suppose that there are no such things as character traits as ordinarily conceived. What are the implications for virtue ethics? Perhaps, it does not matter. `Indeed, the virtuous agent is often discussed as an idea which we aim towards, but do not necessarily ever achieve. ... Virtue ethicists do not and need not argue that most people are indeed virtuous or could in principle become virtuous' (Athanassoulis, 1999). But if we know that there is no such thing as a character trait and we know that virtue would require having character traits, how can we aim at becoming a virtuous agent? If there are no character traits, there is nothing one can do to acquire character traits that are more like those possessed by a virtuous agent.

Of course, it depends on what sort of virtue ethics one has in mind. If there are no such things as character traits, one might still imagine what it would be for there to be character traits and one might then try to act in the way that a virtuous person would act if it were possible for there to be one. This would be contrary to one theme in virtue ethics, but in accord with another. (Of course, there are standard problems with this: What should I do if I am in a situation no virtuous person would ever be in? Should I make a promise I know I won't be able to keep if an ideally virtuous person would make the promise and be able to keep it? Harman, 1983.)

Other ideas are possible. Thomson (1997) outlines a kind of virtue ethics that appeals in the first instance to virtuous actions rather than to character. Merritt (1999) argues persuasively in favor of a Humean virtue ethics that can allow for nonrobust character traits that are supported by the social situation, in contrast with an Aristotelian virtue ethics that requires robust character traits.

Although there is clearly much of value in these last two ideas, I myself think it is better to abandon all thought and talk of character and virtue. I believe that ordinary thinking in terms of character traits has had disastrous effects on people's understanding of each other, on their understandings of what social programs are reasonable to support, and of their understandings of international affairs. I think we need to get people to stop doing this. We need to convince people to look at situational factors and to stop trying to explain things in terms of character traits. We need to abandon all talk of virtue and character, not find a way to save it by reinterpreting it.

One minor point. Harman (1999) does not 'suppose that either moral philosophers or virtue ethicists are unaware of the conclusions of social psychology or the Milgram experiments in particular' (Athanassoulis, 1999). The Milgram experiment is the most famous of contemporary psychological experiments and philosophers have certainly been thinking about its implications for more than twenty-five years. Although fewer philosophers have paid attention to the ensuing skepticism in social psychology about character traits, some certainly have. Harman (1999) cites Flanagan (1991), Railton (1997), Doris (forthcoming), and Merritt (1999).

Athanassoulis (1999) cites Kupperman (1991) and Cullity (1995) as examples of philosophers aware of the conclusions of social psychology. But neither speaks directly to the issue discussed in Harman (1999). Kupperman (1991) does discuss some of the relevant psychological literature, but does not examine the case against the existence of character traits in any detail.

Cullity (1995) challenges the idea endorsed by Thomson (1997) that virtue ethics might be based on acting viciously or virtuously rather on having a vicious or virtuous character. Cullity argues that whether an action is callous can depend on the agent's attitude toward a range of actions, 'specifically, attitudes of willingness to make a certain maximum sacrifice in response to a given collective need. To evaluate attitudes of this kind as callous is not yet to evaluate action; but it is to evaluate an element of the agent's character' (299). This last part does not follow, however, since there clearly can be attitudes of this sort even if there are no such things as robust character traits.

Finally, an even more minor point. Recall that Athanassoulis says, 'What the Milgram experiments challenge is not the assumption that people can have character traits, but rather the assumption that most people will act compassionately under pressure'. Notice the phrase, 'most people', which I have emphasized. Similarly, Kupperman says, 'At the very least, his experiments together with their replication in various countries prove that most people have weak characters' (170, my emphasis again). But in Milgram (1963) not just most, but every subject was willing to apply shocks of up to 300 volts, twice what was expected ahead of time to be the norm.


Athanassoulis, N., (1999). `A response to Harman: Virtue Ethics and Character Traits.'

Cullity, G., (1995). `Moral Character and the Iteration Problem,' Utilitas 7, pp. 289-99.

Doris, J. M. (forthcoming). People Like Us: Personality and Moral Behavior. New York: Cambridge University Press.

Flanagan, O. (1991). Varieties of Moral Personality. Cambridge, Massachusetts: Harvard University Press.

Harman, G., (1983). `Human flourishing, ethics, and liberty,' Philosophy and Public Affairs 12 (1983) pp. 307-322.

Harman, G., (1999). `Moral philosophy meets social psychology: virtue ethics and the fundamental attribution error.' Proceedings of the Aristotelian Society 1998-99, 99, pp. 315-331.

Kupperman, J., (1991). Character. Oxford, Oxford University Press. Appendix A.

Merritt, M. (1999). `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.

Railton, P. (1997). "Made in the Shade: Moral Compatibilism and the Aims of Moral Theory," Canadian Journal of Philosophy, Supplementary Volume 21.

Ross, L., & Nisbett, R. (1991). The Person and the Situation: Perspectives of Social Psychology. New York: McGraw-Hill.

Thomson, Judith Jarvis, (1997). `The Right and The Good,' Journal of Philosophy 94, 273-298