Character and Moral Psychology and the companion volume Moral Character: An Empirical Theory (Oxford University Press, 2013) represents the most sustained and comprehensive philosophical attempt to examine and evaluate the implications of contemporary empirical moral psychology for moral theory. Christian B. Miller has read just about all current philosophical papers, books and psychological studies concerned with “moral character” and uses his extensive knowledge of the field to provide novel, conceptually sophisticated and empirically informed discussions of character traits, of the dominant psychological models of personality and character and of the so-called “situationist critique of virtue ethics”.

The book is divided into four parts. Part 1 presents Miller’s own model of character and character traits, the “mixed trait framework”. The starting point for Miller’s account is the well-known discrepancy between philosophical theories of (moral) character on the one hand and empirical psychological studies of human behavior on the other.

Moral philosophers (and probably also many non-philosophers) typically assume that most people have relatively stable character traits, which play a significant role in determining how they act. In other words: People have a certain character, which influences, perhaps even determines, how they act and how they live their life. Unfortunately this intuitively plausible idea does not seem to be backed up by the data provided by empirical psychology. Hundreds of behavioral experiments concerning a number of different phenomena (cheating, lying, helping, harming) seem to suggest, that most people do not have the kind of stable moral character traits presupposed by most forms of moral theory. How people act seems to depend much more on particular, sometimes apparently insignificant, aspects of the situations the take themselves to be in than on inherent traits or dispositions.

Studies such as Milgram’s famous “Obedience experiment”, Darley and Batson’s “Princeton Theological Seminary experiment” and Isen and Levin’s “Helping for a Dime experiment” have thus lead at least some philosophers and psychologists to either entirely deny the existence of character and character traits, or to at least greatly downplay the importance of moral character for explaining behavior.

Miller provides a different take on these experiments (and of the countless other empirical studies he relies on in his books). What these experiments reveal, at least according to Miller, is not the non-existence of character traits per se, but rather the non-existence of those particular character traits we normally ascribe to ourselves and other; traits which in the moral domain goes by the name of virtues (and vices of course. However, for the sake of simplicity I will only refer to virtues in the following discussion).

Virtues are character traits which reliably dispose people to act morally appropriately in a wide range of situations. A fully virtuous person, it thus seems natural to say, is a person who possesses the character traits which enables her to act morally appropriately in all, or at least most, of the situations she might find herself in. It is the existence of this particular sort of character trait, traits which reliably leads to morally appropriate behavior, which the experiments of empirical moral psychologists seem to deny. Most people do not seem to be even weakly virtuous (or, for that matter, weakly vicious) in so far as they do not seem to possess any character traits that reliably lead them to act morally appropriate.

One obvious consequence to draw from this is of course to deny the existence of character traits as such. But as Miller convincingly argues (in chapter 2 of Character and Moral Psychology, and, at greater length, in Moral Character: An Empirical Theory) this would not only be premature; it would also contradict the available empirical evidence. While it is true that most people do not seem to possess virtues or vices, they do seem possess clusters of interrelated mental state dispositions which reliably result in consistent patterns of behavioral variability (p. 56). But this sort of “consistent behavioral variability” cuts across our traditional normative evaluations of behavior in ways that makes it impossible to characterize the underlying character traits as either virtues or vices.

More precisely; the character traits that people do seem to possess are neither virtues nor vices because they “have both significant morally positive aspects (hence precluding them from counting as vices) alongside significant morally negative aspects (hence precluding them from counting as virtues)” (p. xi-xii). These character traits are neither entirely good (because they sometimes lead to morally problematic behavior) nor entirely bad (because they sometimes lead to morally appropriate behavior). Our moral evaluation of these types of character traits are thus mixed (see e.g. p. 44); hence Miller’s favorite name for his own approach; the mixed trait framework.

Armed with the mixed trait framework Part 2 then discusses three alternative models for understanding character and character traits: situationism, the CAPS (Cognitive-Affective Personality System) model, and the Big Five (or “Five Factor”) model and argues that the mixed trait framework is both empirically and conceptually superior to these approaches. Miller’s argument in each case takes the same form. He first explicates the conceptual and theoretical basis of the model in question, trying to weed out theoretical inconsistencies and imprecisions. He then shows how the model builds on, deals with or incorporates the empirical data outlined in previous chapters. And finally he critically evaluates the conceptual and empirical adequacy of the model.

There is much of interest in Miller’s discussions. Here I will simply summarize his main conclusions. With regard to situationism Miller argues that the most plausible and consistent version of this approach has a somewhat surprising implication namely that “situationists can accept the existence of cross-situationally consistent moral character traits“ (p. 106, emphasis in original). Why? Because, as we have already seen, Miller believes, that philosophers and psychologists alike have typically overstated the implications of the available empirical data. What the experimental studies show, Miller argues, is not that people do not have any moral character traits, but rather that the character traits we do have seem to differ wildly from our traditional conceptions of virtue and vice.

As for the CAPS model Miller believes that this approach, at least as it is most typically stated, does not qualify as an empirical theory of character or character traits. According to Miller the best way to understand the CAPS model is as an attempt to use technical language to re-describe and find “supporting evidence for basic platitudes of commonsense folk psychology” (p. 108, emphasis in original). In other words: the model provides no new insights into the workings of the human psyche, but merely re-describes the psychological understanding of the common man in a more technical jargon. For this reason Miller believes that the CAPS model is best viewed “as more of a background framework or “meta-theory” with general principles from which to start developing an actual, detailed account of personality…[…] But it is not an actual theory of personality itself” (p. 127-128, emphasis in original).

Finally with regard to the Big Five framework Miller argues that the metaphysical status of the five “traits” identified by this approach is unstable. Either the Big Five traits are viewed simply as “descriptive labels for people without […] any underlying, metaphysical existence, causal powers or explanatory role to play” (p. 134). This, according to Miller, is probably the most plausible way to interpret the model. But in that case the Big Five model ends up in much the same position as the CAPS model; not as a theory of personality (it does not provide any explanation for why people behave as they do, nor does it aim to predict how they will act in the future) but rather as a handy way of organizing the phenomena to be explained by such a theory. Or the Big Five model is conceived as a theory, which aims to explain the basic psychological constituents of personality. However (or that at least is Miller’s claim) if one proceeds along this path, then the methodology behind the construction of the Big Five frameworks threatens to turn character traits into “mysterious entities, with their own ungrounded causal powers” (p. 140).

Part 3 of the book applies the mixed trait framework to discussions within meta-ethics and normative ethics, thus highlighting the implications of this approach for contemporary moral philosophy. Chapter 7 develops “an error theory about moral character” (p. 153) inspired by, but clearly different from, John Mackie’s well known error theory about (moral) value, and discusses the implications of adopting such a theory. From a philosophical point of view this is one of the most interesting parts of the book, in particular section 7.2, which examines the precise meaning of the  “the fundamental attribution error”; the human tendency to explain behavior by reference to (non-existent) character traits.

Chapter 8 focuses on John Doris’ and Gilbert Harman’s famous and influential criticism of virtue ethics. Miller agrees with Doris and Harman that empirical moral psychology raises serious trouble for most existing forms of virtue theory, but disagrees with their understanding of what precisely the problem consists in. Doris and Harman argues that empirical studies show (or at least strongly indicate) that most people do not possess any global character traits. And since virtues have traditionally been conceived as global character traits (they are presumed to reliably lead to morally appropriate behavior in a wide range of situations), moral psychology seems too undermine the empirical plausibility of virtue ethics.

According to Miller however this is the wrong way to the right conclusion. As already mentioned several times Miller believes that the empirical data supports the existence of some global and relatively stable character traits. The problem for virtue ethics according to Miller is thus not the non-existence of global character traits. No, the problem is that none of the character traits we do seem to possess qualifies as virtues, because they do not reliably lead to morally appropriate behavior (pp. 195-198).

Part 4 sketches how, in Miller’s opinion, philosophers and psychologists ought to proceed in the light of the insights provided by Miller’s two books. This is by far the weakest and most disappointing part of the book. Miller primarily focuses on how we might cultivate and develop the virtues, despite the fact that most people do not seem to possess any virtues: Not surprisingly Miller suggests that “we should turn to the psychology literature to see what, if any, psychological processes might help cultivate” virtuous character traits (p. 228).  Unfortunately the literature to which Miller appeals offers little more than suggestive hints. It would appear that we do not, at the present time, have adequate empirical data to develop a systematic and consistent account of moral development and character cultivation. Miller thus concludes by noting, that:

…the next question becomes devising the most promising, empirically informed strategies for reducing the gap between what our actual character looks like and what it should look like.

Clearly another book is need3ed here. This is one of the main areas I hope both philosophers and psychologists will explore in the future as they continue to better understand moral character. (p. 239)

Miller’s book (and the companion volume referred to above) is and ought to be required reading for everyone working on problems and questions in the interface between moral philosophy and (empirical) moral psychology. Not only does Miller provide the best and most up to date overview of the field that I know of. He also moves the discussion further, raises new and interesting questions and helps to steer the debate away from a number of dead ends. In short: Miller’s book shows both the value and importance of philosophical conceptual analysis for empirical philosophy, and the crucial significance of empirical moral psychology for systematic moral theory.

 

 

© 2014 Carsten Fogh Nielsen

 

Carsten Fogh Nielsen, Ph.D., post graduate student, University of Aarhus. Main interests: Kant, moral philosophy and the philosophy of popular culture.