Understanding ‘Virtue’ Wayne D. Riggs
Published version in Intellectual Virtue: Perspectives from Ethics and Epistemology, Michael DePaul & Linda Zagzebski, eds. (Oxford: Oxford University Press, 2003)
Understanding ‘Virtue’ and the
Virtue of Understanding 1
Virtue theories are popular in epistemology these days, but they are a diverse lot. It is sometimes hard to see what they all have in common besides the appropriation of the term “virtue.” However, one can divide them very broadly into two groups: those theories that really put the analysis of epistemic virtues at the center of the epistemological enterprise, and those that make use of virtues or virtue-talk along the way to analyzing other terms and concepts that they take to be central to epistemology. Examples of the first kind are somewhat more rare than examples of the second. Jonathan Kvanvig, in his book on epistemic virtue theories2, makes a clear plea for epistemologists to take the idea of epistemic virtues seriously in their own right. He proposes the radical notion of abandoning what he calls the “Cartesian perspective” in favor of a different approach that places the intellectual virtues in the foreground of theoretical study. Where evaluations from the Cartesian perspective are synchronic in nature and focused on the individual, Kvanvig argues that social and cross-temporal factors are relevant to important epistemic evaluations.3 Kvanvig thinks that a genuine focus on intellectual virtues will bring such factors into the mix of epistemological theories.
Though Kvanvig does not develop such a theory of his own, Linda Zagzebski takes a few steps in the direction of the new perspective urged by Kvanvig. She does not cast off the Cartesian perspective entirely, but she does develop a detailed account of the intellectual virtues in their own right. Because she accepts a generally Aristotelian conception of virtue, her account includes some social and cross-temporal factors as relevant to having the various intellectual virtues. Virtues are generally acquired over time, and much of our success at being virtuous will depend upon facts about our parents, our society and culture, and other social factors. Thus, Zagzebski’s account of virtue makes some room for these influences.
James Montmarquet also presents an epistemological virtue theory in his Epistemic Virtue and Doxastic Responsibility.4 He, too, considers the nature of intellectual virtue in its own right, though his theory is much less Aristotelian and differs widely from Zagzebski’s view. He takes “epistemic conscientiousness” to be the primary intellectual virtue. His view differs most markedly from other theories of epistemic virtue in that he does not make reliability a necessary condition of virtue. One can be virtuous yet fail to get to the truth very often.
This approach contrasts most sharply with those of Ernest Sosa and John Greco, both of whom at one time or another have identified themselves as virtue theorists. Though each of these philosophers has his own distinct account of epistemic virtue, they share a common commitment to the importance of epistemic reliability.5 They hold that being epistemically virtuous is largely a matter of having cognitive equipment that reliably reports the truth. One possible explanation for this common feature is that each of these epistemologists takes epistemic virtue to be important primarily as a tool to offer an adequate account of knowledge. Epistemic virtues enter the picture by way of their theories of knowledge. Perhaps they began with a commitment to some form of reliabilism, and developing this theory led them to include virtues in their theories to account for certain intuitions about knowledge, or to avoid persistent counterexamples.
But despite the diversity of these approaches to "virtue epistemology," there is one point of agreement—namely, that epistemic virtues, whatever they may be, are defined either teleologically or instrumentally (or both) in terms of our epistemic ends. Accordingly, a character trait of an agent is an epistemic virtue only if, by the operation of that trait, one is either "aiming at" some epistemic end or one is actually likely to obtain it (or both). Thus, the definition of “epistemic virtue” one arrives at on this common picture depends on both (i) the specification of our epistemic ends, and (ii) the characterization of the relationship that must obtain between a cognitive trait and those ends for such a trait to count as an epistemic virtue. In this essay, I will speak to both of these issues. First, I will argue that one recently proposed relationship between cognitive traits and epistemic ends is not an appropriate condition on a trait’s being an epistemic virtue. Then I shall turn to the question of our epistemic ends. What ends must we accept in order to capture the traits we normally take to be epistemically virtuous? My answer will depart from the standard line among virtue theorists and other epistemologists alike. But first, …
2. A Methodological Prelude
Obviously, much of the literature in “virtue epistemology” is inspired, directly or indirectly, by Aristotle’s ethical theory and contemporary developments of similar “eudaimonistic” ethical theories. I think that this eudaimonistic theory of moral virtue offers a good model for accounting for intellectual virtues. For one thing, it is standard in the literature to speak in explicitly teleological terms about cognitive or epistemic “ends.” Indeed, until recently there was a fairly strong consensus about just what those ends were: maximizing truth while minimizing error. Though I think this particular account of our cognitive ends is mistaken, I agree that the best way to go about theorizing in epistemology is to specify the goals or ends that define the limits of the field. Having done so, it seems eminently reasonable to define intellectual virtues in terms of just those ends.
A highly simplified version of Aristotle’s view goes something like this: (1) eudaimonia is the highest good for humans. (2) Whatever contributes to a life of eudaimonia is good by virtue of that fact. (3) The members of the standard list of virtues (courage, justice, benevolence, etc.) contribute to a life of eudaimonia. (4) So, the virtues are good because they contribute to a life of eudaimonia.
On this schematic account of Aristotle’s view, the ultimate list of the virtues is driven by his conception of eudaimonia, often translated as “flourishing.” Contributing to human flourishing is both necessary and sufficient for a character trait to count as a moral virtue. One potential advantage this sort of view holds over non-eudaimonistic virtue theories is that once you’ve settled on a description of “flourishing,” you can go out into the world and see what character traits are necessary and sufficient for it. You don’t have to rely on unreliable bare intuition or on culturally relative conceptions of virtue to make one’s list of the virtues. This approach “grounds” the virtues in a conception of flourishing that, in principle at least, determines for us what can count as a virtue.6
I said that this is an advantage for this approach to a eudaimonistic virtue theory, but it is also a disadvantage. If we were willing to go wherever human flourishing takes us, so to speak, then this method of determining the moral virtues would be fine. However, in practice any theory that purports to account for the moral virtues simply must include certain specific virtues or it is ruled out from the start. For example, no one would be willing to accept a theory that failed to count courage as a moral virtue. Even if one could argue flawlessly that this followed from a reasonable conception of human flourishing, it would be assumed that the theory was seriously flawed simply because it failed to count courage as a moral virtue. Similarly, any theory of epistemic virtues that failed to count open-mindedness, say, as an epistemic virtue would be equally implausible.
In this regard, virtue theories in ethics are in a position no different from that of other moral theories. Any theory of right action, for example, simply must count the brutal acts of torture and rape committed by soldiers in the Bosnian war as wrong in order to be taken seriously. Theories in ethics, as well as in many other areas of philosophy, are constrained by certain commonly held intuitions that we hold dear. On the other hand, certain general principles are equally central to our theorizing. Any theory that did not count the killing of innocents as at least prima facie wrong will also get short shrift. Deeply held general moral principles as well as common intuitions about particular cases constrain ethical theorizing.
Similarly, I think it is unrealistic to propose a virtue theory that claims to derive the actual list of virtues from an account of flourishing. In the actual construction of the theory, the prior intuitions of the philosopher about what traits are really virtues will be driving the account of flourishing at least as much as the other way around. Thus, I will explicitly pursue a strategy of reflective equilibrium in developing the sketch of a theory of intellectual virtue that follows. I am committed to the existence and value of intellectual virtues, regardless of whether any particular theory of them, including my own, succeeds in making good sense of them. On the other hand, I am not so intuition-driven that just any account of what I will call “intellectual flourishing” will do so long as it captures my intuitions correctly. I will argue in the following sections that we can make sense of the nature and value of various intellectual virtues by placing the notion of “understanding” at the center of our account of human intellectual flourishing. This not only captures certain important intuitions about what the intellectual virtues are, but it also seems just the right sort of thing to be central to intellectual flourishing. As complications arise in fitting our intuitions about intellectual virtue to a rough and ready idea of “understanding,” the give and take of reflective equilibrium will guide our modifications as necessary.
3. Intellectual Virtue and Success
As I mentioned in the introduction, most current proponents of a “virtue theory” in epistemology take reliability at producing true beliefs to be a necessary condition for a character trait to be a virtue. This is most obvious with regard to “virtue reliabilists” like Ernest Sosa and John Greco. For example, Sosa defines a virtue as follows.
A subject S’s intellectual virtue V relative to an “environment” E may be defined as S’s disposition to believe correctly propositions in a field F relative to which S stands in conditions C, in “environment” E.7
Somewhat more simply, Sosa’s definition states that S’s intellectual virtue V is a disposition to have true beliefs about propositions in a field F when in environment E and conditions C. Thus, for example, properly functioning perceptual faculties (operating in a “normal” environment under “normal” conditions) are intellectual virtues according to Sosa.
The fact that Sosa’s definition of a virtue leaves it relative to a specified environment and conditions is quite significant. An actual “track record” of success is not necessary for one to have a virtue on his account. Nor is even a disposition to believe correctly in the environment and conditions one finds oneself in necessary to have an intellectual virtue. So, for example, if a cognitively and perceptually normal human being were transported to an evil demon world, this person would not cease to have whatever epistemic virtues she had before, despite the fact that those very dispositions no longer produce mostly true beliefs. The reliability of the disposition is indexed to our (presumably) “normal” world. Since this kind of reliability is less demanding than the kind to be considered in a moment, I shall call this “weak” reliability.
Though she does not label herself a reliabilist, Linda Zagzebski also thinks it important to build reliability into the definition of a virtue.
A virtue, then, can be defined as a deep and enduring acquired excellence of a person, involving a characteristic motivation to produce a certain desired end and reliable success in bringing about that end.8 (emphasis in original)
On this view, though the right sort of motivation is required to have a virtue, one fails to have the virtue unless such motivated action reliably yields the end toward which the motivation is directed. So, unlike Sosa, Zagzebski does require a successful “track record” of coming to hold true beliefs in order for a trait to count as an intellectual virtue. This is, of course, a much more demanding condition to satisfy. Consequently, I shall call this condition of Zagzebski’s above definition “strong” reliability.
Which version of reliability one builds into one’s definition of “virtue” has major implications for the tenability of one’s theory of the virtues, whether those virtues be moral or epistemic. As I shall argue below, strong reliability is indeed much too strong to serve as a condition of virtue—epistemic or moral. Weak reliability escapes these criticisms, and may well be a part of a complete account of epistemic virtue.
The Secret of “Success”
Julia Annas argues against what I have called a strong reliability or “success” component in theories of virtue ethics.9 I will briefly present her argument, and then propose that the same objections to a success component in virtue ethics hold against a reliability requirement in virtue epistemology.
Annas acknowledges that both the Aristotelian and the Stoic traditions of virtue ethics seem to hold that having a virtue requires that one be successful in achieving the aim of that virtue.10 However, these traditions come apart when one considers an ambiguity in the specification of the “aim” of a virtue.
But what is the virtuous person’s aim in acting? She has two, in fact. One is her telos or overall aim, of living virtuously and acting from motives of virtue. Virtue, after all, is a settled state of the person, with the overall aim of making the person’s life as a whole be one way rather than another, virtuous rather than evil or complacent… The virtuous person’s other aim is what the Stoics call her skopos or immediate target, which is what is aimed at in acting virtuously in any particular case. The target of a just distribution will be everyone’s getting what they are entitled to,… and so on.11
According to Annas, every virtuous act is aimed both at living a virtuous life overall, as well as whatever state of affairs constitutes the virtuous outcome in the particular instance. If a virtue ethics has a success component, it must specify which of the two kinds of aim (or both) one must reliably achieve. Much rides on this decision. Aristotle focused on the reliability of one’s achievement of the immediate target, the skopos, as necessary for virtue. The Stoics, Annas argues, chose rather the achievement of the overall aim, the telos, to be necessary for virtue. There turns out to be good reason to side with the Stoics on this matter.
Virtue ethics is concerned with the person’s life as a whole, with character and the kind of person you are. The right perspective on an action, therefore, will for virtue ethics be the one which asks about success in achieving the overall goal, not success in achieving the immediate target… To the extent that success in achieving the immediate target depends on factors over which the person has no control—moral luck of various kinds—it will be of less interest to virtue ethics.12
On this issue the Stoic view is not just clearer but more defensible than Aristotle’s. To require moral luck for the exercise of virtue is a very extreme view—far more so than the more familiar issue of whether moral luck is required for happiness. Of course it is often not up to me whether my action achieves the immediate target; but is it up to me whether I succeed or fail in acting virtuously—that is, with the right motives, from a developed disposition and with the right reasoning? If it is not, then it is not up to me whether or not I can become a moral person; and accepting this would, at the very least, require huge changes in the way we think of morality.13
Aristotle’s requirement that the virtuous person must reliably succeed in achieving the “immediate target” of a virtue in order to have that virtue places an onerous burden on the would-be virtuous agent. To modify one of Annas’ examples, a firefighter might have the virtue of courage as displayed by her continued willingness to enter burning houses to save the lives of the occupants. If, as it turns out, most of these people die later of complications, or even if she usually finds them already dead, the rescuer’s courage is not diminished, either in the individual actions themselves, or in her character more generally. The rescuer is courageous, despite generally failing to accomplish the immediate targets of her actions—saving individuals’ lives. Yet a virtue theory that has a success requirement on the achievement of the immediate target will fail to count such an individual (or her actions) as courageous.
J. L. A. Garcia has similar worries, though he illustrates the problem with an appeal to the familiar evil demon scenario.
Few of us will feel justified in withholding the classifications benevolent, compassionate, generous, just or honest from someone just because she is unlucky or even inept in her efforts. Consider some possibilities.
In a world wherein an evil demon systematically rendered all someone’s efforts to help ineffectual, we should still consider them (and her) to be … virtuous.
The claim that virtue has a necessary component of reliably successful behavior seems also to have a distasteful implication. If one must try to help others with reliable success in order to be benevolent, say, then it is hard to see how those severely incapacitated either physically or mentally can be virtuous … This conclusion is both counter-intuitive and morally repugnant.14
Here Garcia argues that the strong reliability condition on virtue not only has the counter-intuitive result that clear cases of virtuous actions and people would fail to make the grade, but also that virtue would be in principle inaccessible to whole classes of people who were physically incapable of certain kinds of actions. I’m not sure that this latter point is as obviously counter-intuitive as Garcia thinks it is, but his evil demon scenario complements nicely the arguments of Annas presented previously. And if one does find the latter point seriously wrong-headed, then so much the worse for the strong reliability condition on virtue that would imply it.
So far all this discussion has taken place within the sphere of moral virtue. Are there similar problem cases for a strong reliability requirement in one’s account of epistemic virtue? Certainly, as the following discussion will illustrate.
On the Shoulders of Virtuous Giants
The epistemic analogue to Annas’ objection to the incorporation of strong reliability into the notion of an intellectual virtue is stated nicely by James Montmarquet.
[I]f we are to appraise the relative worth or “virtue” of epistemic agents by the truth-conduciveness of their intellectual dispositions, then how are we to accommodate the approximate equality of epistemic virtue we find in such diverse agents as Aristotle, Ptolemy, Albertus Magnus, Galileo, Newton, and Einstein? From our current vantage point, we recognize these thinkers as differing greatly in the truth of their respective beliefs and systems of belief… How can such rough equality in virtue be reconciled with this verific diversity?15
Montmarquet puts his point in terms of the rough equality in intellectual virtue of our intellectual heroes of times past. I find it slightly more helpful to recast this thought in terms of a criterion of adequacy for a theory of intellectual virtue. I propose that any theory of intellectual virtue that does not clearly and definitively count the likes of Aristotle, Newton, Galileo, etc. as being intellectually virtuous does not capture what we mean by “intellectual virtue.” These individuals (among many others) are our exemplars of intellectual achievement and of intellectual virtue. It is hard to imagine a theory of intellectual virtue that could otherwise be so plausible that we would be willing to give up counting these individuals among the cognitive elite of our shared intellectual history.
And yet, as we now know, a great deal of Aristotle’s science and philosophy was mistaken. It may even be that he was wrong about more of these things than he was right. It is, of course, impossible to take such measurements now, but the mere significant possibility will suffice for our current purposes. For suppose we were somehow to discover that, overall, despite Aristotle’s careful observation, his meticulous study, his insightful explanatory hypotheses, the rigorous examination of his arguments, and so on, he nonetheless believed more falsehoods about the nature of reality (both physical and metaphysical) than truths. Would this unfortunate finding cause us to remove the mantle of intellectual virtue from Aristotle’s shoulders? I think the correct answer is “clearly not.”
Part of the reasoning behind this reaction is that Aristotle was hindered in his theory construction by the absence of the kinds of technological and theoretical innovations that we have had the benefit of in our own time. Had Aristotle had a microscope, a telescope, a better understanding of human psychology, and so on, we are confident that those very same intellectual character traits that led him astray in his own time (we are assuming for the moment) would have led him to theories much more like the more accurate (we hope!) theories of today.
This point can be generalized to all major figures in our intellectual history. The greatest and most virtuous intellects in our shared human history all labored under what we now know to be mistaken assumptions, inaccurate or imprecise measurements, faulty methods, and a whole host of other disadvantages. These factors limited what these figures were able to accomplish in terms of their immediate targets—understanding the nature of reality—but not the degree to which they could develop the intellectual character traits for which we rightly admire them. The very fact that we unhesitatingly ascribe intellectual virtue to these intellectual giants, despite their often spectacularly mistaken views, is eloquent testimony to the fact that success at accomplishing the immediate targets of cognition or inquiry, true belief, is not necessary for intellectual virtue.
This conclusion, if correct, rules out as non-starters any theory of intellectual virtue that requires strong reliability, or truth-conduciveness, of a trait before it can count as such. Such great thinkers as Aristotle and Newton were simply wrong about a great number of things they took to be true descriptions of reality. To deny them intellectual virtue because of this “failure” seems so outrageous as to constitute a reductio of any view that implies it.
Even if successfully “producing” truths cannot be made a requirement on intellectual virtues, it does not follow that considerations of truth and falsity are irrelevant to them. Far from it. Indeed, it might seem that the solution to this problem is simply to retract the “success condition” vis a vis the immediate target (true belief), and instead allow that one can have a virtue so long as one has a stable disposition to be appropriately motivated to acquire true beliefs (and avoid false ones). In other words, we could replace the instrumental relationship between intellectual virtue and true belief with a teleological one.
Garcia seems to suggest just such an account of moral virtue, according to which the courageous or benevolent person need not have a successful “record” of lives saved or suffering relieved, so long as she has a stable disposition to act appropriately in relevant circumstances. Similarly, one need not have a successful record of switching allegiance from false views to true ones heard about later in order to have the virtue of open-mindedness. In both these cases, luck plays too large a role in determining the particular circumstances that determine whether the right sort of action at the right time to the right degree will actually succeed in attaining the immediate target. If virtue theory is to survive at all, we must assume that people have sufficient control over the development of their own virtue, moral or intellectual, to be responsible for their moral and epistemic character. Thus, we cannot allow luck to play such a large role in determining the presence or absence of virtue. Something less than strong reliability must be found to serve its role in the correct account of virtue.
Does this pave the way for a definition of virtue in terms of weak reliability? Not necessarily. It may be that even weak reliability will be too strong for someone like Garcia. One may think that even the propensity to be successful in one’s putative virtuous acts under the right conditions and in the proper environment places too onerous a burden on the would-be virtuous. Even so, one won’t be able to make the same case against weak reliability that Garcia does against strong reliability. But my point is simply that there are alternative possible accounts of virtue that reject reliability altogether.