As some of you are aware, I am currently working on a business book based on the virtue ethics framework. Knowing this, Samuel Hammond recommended his very favorite book on business ethics: Morality, Competition, and the Firm: The Market Failure Approach to Business Ethics by the formidable Joseph Heath. From the very beginning of the book, I suspected Sam was up to something—I’m no fan of Pareto-type reasoning in moral matters, and that is basically all of what Heath had originally intended to offer on the subject of business ethics. I did greatly enjoy the chapter (which was originally this paper) in which Heath argued that we should treat agency theory (and by implication, most of game theory and economics) as a branch of critical theory. That is, we need not treat it as revealed truth, but rather as a series of thought experiments from which we can draw valuable lessons. Thus, while there is far less opportunism than agency theory would predict, the “fault lines” of major corporate scandals fall right where agency theory would lead us to believe they would be. This is a profoundly McCloskeyan perspective.
But when I got to the chapter on virtue ethics, I realized that old Sam had laid a trap for me, and I had walked right into it. Heath is ruthless in his dismissal of virtue ethics, calling it “debunked” and essentially no better than “folk psychology”. Ensnared in Sam’s machinations, I feel at this point helpless to do anything but see his diabolical scheme to completion. That is, I’d like to offer a response.
Heath offers three critiques: first, that psychology has demonstrated that character as virtue ethics understands it does not exist. Second, that sociology has demonstrated that the process of moral learning virtue ethicists embrace is at odds with reality. And finally, that the notion of the good in virtue ethics is at odds with liberal neutrality.
The Situationist Critique
Heath’s criticism from psychology is largely the same situationist argument that has been lobbed at virtue ethics for years, pursued especially by John Dorris. The long and short of this argument is that hundreds of psychological experiments have shown that people’s behavior is highly sensitive to situations. Moreover, people who seem to have a virtue in one area show little correlation with displaying the very same virtue in another area—the clearest case of this being Harshorne and May’s study demonstrating that many students who are honest in one context are frequently deceitful in another.
Heath argues that good behavior in an area where we have settled habits don’t “correlate in any significant way with behavior in other types of situations, even ones that are only slightly different.” This is odd, given that Harshorne and May’s finding was that there was typically a .23 correlation between any two areas where honesty was called upon, a correlation which I should think would be considered quite large in most areas of social science research. Situationists have a tendency to focus on the fact that those apparently influenced by the situation comprise a supermajority, and neglect the fact that a substantial minority do not fit this narrative. Those who invoke the Milgram experiment, for instance, forget that 45% of the participants refused to complete the experiment.
These experiments, moreover, have a few serious defects if the goal is to “debunk” the psychology behind virtue ethics. First, other than the Harshorne and May experiment, most of these were one-time affairs. The virtue ethics of Aristotle, and of people such as Julia Annas who emphasis the skill analogy of virtue, has moral learning as one of the central components. The problem with one-shot experiments is that we don’t get to see whether those involved took anything away from their own actions, once the point of the experiment was revealed. Second, in order for these experiments to have “debunked” virtue ethics, it must be supposed that virtue ethicists believe that most people have achieved complete virtue.
Aristotle certainly did not believe such a thing. Deirdre McCloskey may believe something like that, but I follow Christian Miller in thinking that most of us have merely developed a variable range of local character traits, but only a minority manage to develop global character traits—that is, most character traits for most people are domain-specific. Nevertheless, there is nothing about this which suggests that the domain cannot be expanded, and eventually become global. Moreover, this possibility is suggested by the minority of participants in psychology experiments who do exhibit good character in a number of circumstances.
Local character traits also answer another point made by Heath—that if full virtue is indeed rare, then:
Of course, it is then incumbent upon the virtue theorist not only to show that such rare individuals do exist, but that other people are able to identify them reliably (so that emulation of these persons can serve some kind of useful function in everyday morality).
But if the path to complete virtue (which most people never finish) involves developing a lot of local character traits along the way, then everyday morality would by and large involve learning from virtue as it is observed in local contexts. And indeed Annas emphasizes just this, without using the terminology of local character traits. In Intelligent Virtue, she argues that we grow to recognize courage in a variety of specific contexts—on the one hand, seeing one’s parents chase off a vicious dog, on the other, watching a video of the passive resistance tactics used during the civil rights movement. Two very different types of courage; each possible to develop in a local way. But the locality of each does not preclude the possibility of developing them into a more global character trait, with time and experience.
Following Annas, I also think that Heath significantly misunderstands what she calls the “intellectualist” aspects of Aristotelian virtue. Once it is understand that practical wisdom unifies the virtues through a combination of habituated dispositions and reasoning, Heath’s argument loses a lot of its bite. To quote Annas:
Doris gives an example where a colleague invites you to dinner when your spouse is absent, and an attempt at seduction is clearly in the offing. Only the situationist, he thinks, will have the intelligence to avoid the dinner at the outset; the person who relies on character feels ‘secure in the knowledge of [their] righteousness’ and goes along – only for it to be probable that their reliance will turn out to be misguided. Doris misses the point that the virtuous person would have an intelligent understanding of what fidelity requires, and would do just what Doris says the situationist would do. Only somebody clueless about what virtue required would rely on the force of habit alone.
“Lead me not into temptation” is a prudent, time-tested strategy for preserving one’s virtues.
This is particularly relevant to Heath’s observation that how people behave often has more to do with construal of a situation than with anything objective about the situation itself. He suggests that business ethicists could more productively focus on influencing how their students construe particular situations. But if this is possible—that is, if a professor is capable of teaching a student how to construe future situations in such a way that makes them behave ethically when they would not have—how can this be described as anything but a persistent character trait in the student? How is such a thing in any way at odds with the traditional conception of practical wisdom? Aristotle’s phronesis is quite literally the ability to read a situation—to determine how to do the right thing, in the right way, in relation to the right people.
The student-teacher relationship brings us to the second part of Heath’s criticism.
Peers Are More Persuasive Than Teachers
On the subject of moral education, Heath argues:
The concern raised by criminological studies is that virtue theory, despite being social, may be social in the wrong way. It emphasizes the “vertical” dimension of behavioral transmission, from parents or authority figures to children, instead of the “horizontal,” from one peer to another. It also assigns greater importance to interactions in the past—on the grounds that they produce habits, which sediment to form character—over interactions in the present. Criminological research suggests that this puts the emphasis on the wrong set of social interactions. When it comes to determining criminality, horizontal interactions appear to be far more important than vertical ones.
To the extent that virtue ethicists have over-emphasized the role of parents, teachers, mentors, and authority figures in general in moral development, this is an appropriate criticism. Like Protagoras, I believe that we spend all our lives teaching and being taught how to be virtuous by everyone we encounter. Our peer group is always much larger than our mentor group, and since we are of the same status it is them that we are expected to emulate (though our mentors obviously prefer us to emulate only the best of them).
Annas observed of skills in general that “Skilled dispositions are not static conditions; they are always developing, being sustained or weakened” and further argued that it is all too common for skills to “ossify and decay” when allowed to. Given the important role that peers within our skilled community play in providing points of reference for, and sometimes constructive criticism and error correction, it makes sense that our skills may languish if we fall into a peer community with low standards and a poor ethic when it comes to the maintenance of the skill in question. Practical wisdom is the skill necessary for the exercise of the virtues, and a peer group that lacks wisdom can inhibit one’s ability to hone and maintain that skill, or acquire it in the first place. Again, this is implicit in the important role that Aristotle gives to friendship and community, and explicit in MacIntyre’s own formulation.
Nevertheless I don’t think that crimonology is the best basis for drawing conclusions about people in general. Heath does, because of the conclusions about character he drew from psychology (which I contested above) and because he buys into Hannah Arendt’s theory that evil is banal, inspired by a patently false understanding of Adolf Eichmann’s character. The Milgram experiment is frequently invoked in discussions of the banality of evil, this in spite of the fact that the participants who went along with their instructions were often agitated and shell-shocked by what they were doing. They were hardly simple yes-men who were blasé to the apparent suffering they were inflicting or the possibility that they were doing something wrong.
A Neutered Liberalism
Heath’s final argument is that virtue ethics flies in the face of modern democratic liberalism because it posits a substantive notion of the good. Heath embraces Rawls’ “reasonable pluralism” and calls for neutrality “with respect to all ‘big-picture’ views” in political matters.
This is perfect rubbish. Liberalism is a notion of the good, and it must be. Otherwise, why not be neutral with respect to people’s choice to form a community of pedophiles, or a community where murder laws are not enforced? Attempts to take the very historically specific compromise under which modern liberal states emerged and justify them in terms of neutrality is a fool’s errand. John Rawls failed in the task, and so will anyone else that attempts it. The idea that a political order must be justified by some sort of consensus—real or the result of some ethical autoCAD—is itself a substantive idea of the good of politics. Equality, which Heath holds up as a cardinal liberal political virtue, is as far from neutral as it gets—and this goes for both the ideal of equality before the law and the ideal of equality in outcomes.
There are certainly pragmatic reasons to argue for greater or lesser variety in the conceptions of good that people are allowed to pursue within the bounds of the law. But we need a substantive conception in order to draw those boundaries—or else we fall into the problems found in pure narratives of federalism, where we find ourselves with no rationale for opposing slave-states and other horrors.
If You’re So Smart, Why Aren’t You a Saint?
Heath concludes the sociology section with the following:
Again, it is always possible to design a more sophisticated, cognitivist version of virtue theory, which would avoid any commitment to these discredited common-sense ideas about bad upbringing and deviant values. Yet this is an uphill battle, because the traditional Aristotelian-Thomistic theory of virtue is the primary source of these common-sense ideas in Western societies. Rather than trying to rehabilitate the old-fashioned vocabulary, there is a lot to be said for starting from scratch, with the empirical evidence, and developing a vocabulary that is better suited to accounting for what we know and understand about norm-conformity and social deviance.
In business, when you work closely enough with someone on a regular enough basis, it doesn’t take long before you can anticipate how reliable they will be in a number of respects. Will they recognize moments where they can take initiative in a valuable way, and do so? Will they do the bare minimum of what is asked of them? Will they fail to even do the bare minimum, and simply come up with an excuse after the fact?
These are observable regularities. The fact that I cannot show Heath or anyone else these regularities in perfectly sterilized and controlled experimental conditions does not eradicate their existence. Evidence is important but like McCloskey I believe that our idea of what should count as evidence is often excessively narrow. If the only perspective that is allowed to count is one that self-consciously steps outside of life as it is lived and puts people in highly contrived scenarios that they frequently have little prior context for, then human nature is going to look very unrecognizable indeed—because that is not the perspective from which most human lives are lived.
In short, I think Heath significantly overplays the extent to which the core of virtue and character have been “discredited”, though I do not think we should ignore the enormously valuable contributions of the researchers that Heath draws upon. I also reject his remark that most virtue ethicist responses to situationists make virtue ethics unfalsifiable—though I hope he doesn’t have an overly positivist idea of what falsifiability entails. Christian Miller’s response opens the door to a number of possibly fruitful avenues of experimental investigation.
As a parting thought, I’d like to suggest that “starting from scratch” in ethics has rarely worked out very well, and that perhaps the “old-fashioned vocabulary” has persisted for a reason.