1. BACK TO CHARACTER AND VIRTUE

The emergence of contemporary virtue ethics has had an invigorating effect on mainstream moral theory. It also has had a stimulating impact on the social sciences. There is a longstanding interest in character and traits in psychology—particularly trait theory in personality psychology (Allport, Reference Allport1937). This trend has been freshly revitalized by “positive psychology,” a recent movement that is focused on the study of the strengths and virtues that enable individuals and communities to flourish (Peterson and Seligman, Reference Peterson and Seligman2004). Organizational scholars have imported this concern from psychology; they aim to find the drivers of positive behavior in the workplace as a way of understanding, predicting, and explaining high performance (Cameron, Dutton, and Quinn, Reference Cameron, Dutton and Quinn2003). Similarly, business ethicists have emphasized the centrality of character and virtue concepts in business ethics (Solomon, Reference Solomon1997; Moberg, Reference Moberg1997, Hartman, Reference Hartman1998; Koehn, Reference Koehn1998; Moore, Reference Moore2005). And economists are following suit (McCloskey, Reference McCloskey2006; Bruni & Sugden, Reference Bruni and Sugden2013; Sugden, Reference Sugden2015).

2. VIRTUE AND THE IS-OUGHT GAP

Prominent topics a long time ago, character and virtue lost their reputation in philosophy and psychology. In philosophy, character lost its charm during the nineteenth century. Principles prevailed over character as scientists began to claim that science was about scientific principles explaining events rather than about the features of substances explaining the activities of things (Anscombe, 1958). In psychology, there was a shift of attention away from character towards the allegedly value-free notion of personality in the 1930s, when the social sciences decisively moved to split scientific facts from moral value, following a dominant trend in the development of the social sciences. Personality traits were separated from character traits, the former as scientific phenomena, the latter as normative statements. Consequently, character—a moral term—was gradually substituted by personality — an allegedly amoral, value-free term (Haidt, Reference Haidt2006). In the end, Anscombe (1958) and Maslow (1954) initiated the re-emergence of character in the 1950’s in Anglo-American philosophy and psychology.

Today, philosophers keep talking about character traits and psychologists about personality traits. But what is the difference? It might be helpful to first consider what we mean by the notion of a trait. A physical trait—say, “big”—describes a person’s physical features or physical abilities. A personality trait—say, “extrovert”—focuses on one’s personality. Personality traits, as studied by psychologists, are distinguished from character traits, which are the main concern of virtue ethicists. Character traits are a subset of personality traits that have bearing on moral matters (Gert, 1988; Brandt, Reference Brandt1970). This understanding is initially suggested by Aristotle when he distinguishes virtue and moral goodness from other aspects of our inner selves because only the virtues are matters of choice. While some virtue ethicists hold that the character traits of interest for virtue ethics supervene on the traits studied by personality psychologists (Solomon, Reference Solomon2005; Hartman, Reference Hartman1998; Moberg, Reference Moberg1999),^{Footnote 8} others stress the distinction between character and personality, arguing that two people may be radically different in personality yet quite alike in moral character (Kupperman, Reference Kupperman1991; Audi, Reference Audi1995).

Both social scientists and ethicists aim to contribute to the consolidation of the virtue approach, but they keep sending conflicting messages. The fact that philosophical theorizing parted ways with psychological research on the virtues takes us back to the old divide between facts and values, that is, between scientific psychology and moral psychology (Doris and Stich, Reference Doris, Stich, Jackson and Smith2005). And the split between social scientists and ethicists on the nature of virtue mirrors the dichotomy between empirical and normative approaches to business ethics research, with the social scientists describing the antecedents and consequences of certain traits and behavior and the philosophers advocating for the character traits a person needs in order to flourish as a human being (Alzola, Reference Alzola2011).

This lack of integration creates quandaries and ambiguities in the theoretical conceptualization and the empirical operationalization of virtue constructs (Bright et al., Reference Bright, Stansbury, Alzola and Stavros2011). Conflicting streams of research provide different and incompatible lists of virtues, different rank orders, and incompatible theories of the virtues, because they conceptualize and measure two different things using the same label. But, I shall argue, virtue is not polysemic.^{Footnote 9}

Why does this matter at all? At least for the following four reasons. First, we need an unambiguous conceptualization of what a virtue is in order to make progress in virtue theory (Von Wright, Reference Von Wright1963). Second, the conceptualization of virtue is the battlefield where competing moral theories struggle for the normative priority of deontic and aretaic notions (Williams, Reference Williams1985; Wallace, Reference Wallace1978). Third, the plausibility of empirically grounded objections against virtue ethics depends on an accurate understanding of what virtue is (Doris, Reference Doris2002; Harman, Reference Harman2003). Finally, the academic fields of management and psychology need an uncontroversial notion of virtue in order to identify plausible measures and indicators of “character strengths” and “virtuousness” in organizations (Cameron & Spreitzer, Reference Cameron and Spreitzer2011; Lopez & Snyder, Reference Lopez and Snyder2011).^{Footnote 10}

What, then, is virtue? Let us start with Aristotle’s position on the genus of virtue, arguably “one of the less contentious in the history of the concept” (Zagzebski, Reference Zagzebski1996: 102). Virtue, in Aristotle, is a condition of the soul. There are three conditions of the soul, namely, passions, capacities, and states. Aristotle argues that neither the virtues nor the vices are passions, because we are not called good or bad on the basis of our passions. We are neither praised nor blamed for our passions, but rather for our virtues and our vices. Virtues and vices are not capacities either, for we are not praised or blamed for the simple capacity of feeling the passions. Thus, Aristotle defines a virtue as a state of character (NE, 1106a13-14).

More needs to be said about this definition. What sort of “states of character” do the virtues comprise? How different are they from the “character strengths” measured by psychologists and organizational scholars?

The virtues have been conceptualized as habits (Dewey, Reference Dewey1922; Ryle, Reference Ryle1949; Nussbaum, Reference Nussbaum1999), skills or abilities (Sigdwick, Reference Sidgwick1981), sentiments (Rawls, Reference Rawls1971), tendencies (Wallace, Reference Wallace1978; Kamtekar, 2004), inclinations (Kant, Reference Kant and Gregor1996), dispositions (Doris, Reference Doris2002; Harman, Reference Harman2000), and character traits (von Wright, Reference Von Wright1963). Elsewhere, I have developed an account of virtue, which distinguishes its constitutive elements and dimensions.^{Footnote 11} For simplicity, here I shall define virtue as a character trait and provide a general characterization of its components.

To say that someone possesses a certain character means that he has appropriate and integrated framing capacities, beliefs, desires, feelings, motivations, and behavioral tendencies. Take the virtue of courage. The courageous person is “the person who fears the right things, from the right end, in the right way, at the right time, and is correspondingly confident,” Aristotle says (EN, 1115a17-18). A virtue, then, has four elements, namely, an intellectual,^{Footnote 12} an emotional,^{Footnote 13} a motivational,^{Footnote 14} and a behavioral^{Footnote 15} component. Attributing the virtue of courage to someone—saying that someone is a courageous person—entails the following four claims:

1. 1. He has developed appropriate framing capacities and deliberative skills to understand the presence of danger, he holds correct beliefs, and he deliberates carefully about how to respond to frightening conditions, including death (intellectual dimension).

2. 2. He feels the appropriate level of fear, neither more nor less of the right things. He fears the things that deserve to be feared and not just any bad thing that may possibly happen (emotional dimension).

3. 3. He stands firm against what is painful out of the right motivations and in service of the right ends, not because he fears the reproach of his fellows or expects a reward from the beneficiary of his actions (motivational dimension). And, finally,

4. 4. He successfully stands against frightening things; he is disposed to be unperturbed when facing things that are not too frightening, and he typically succeeds in responding appropriately to frightening conditions on the basis of the constituent understanding, feelings, and motivations (behavioral dimension).

It follows from this that it makes sense to distinguish—as in the title of this article—doing the virtuous action from having a virtuous character. In EN II. 4, Aristotle contrasts the value of acting from craft knowledge—which is purely instrumental, as a means to the right product—and the value of acting from virtue. The goodness of production is established by its usefulness for producing the right sort of product, but the value of virtue is not simply determined by its efficient production; it is intrinsic. A virtuous action has a characteristic motive; it is not simply a means to some further result. And while a product may be produced well even if not produced by a good craftsman, an action is virtuous only if carried out by a good agent (and from the right state of character).

This is where virtue concepts become especially helpful. Virtue attribution serves, I submit, three distinctive purposes, namely, descriptive, explanatory, and evaluative functions (Alzola, Reference Alzola2012):

• • Description: first, virtue attribution provides a short way to summarize how one person differs from another person, that is, a way to describe the state of a person’s soul. The descriptive function of virtue-attribution consists in classifying and categorizing a person’s psychological makeup. For example, ascribing the virtue of courage to a person summarizes a history of mental states and actions in the past as well as highlights individual differences.^{Footnote 16}

• • Explanation: second, virtue attribution has an explanatory role, that is, the possession (or lack) of a certain character trait explains (partially) the possessor’s beliefs, desires, emotions, motivations, and behavior. An explanatory ascription of virtue carries a commitment to the existence of some causal mechanism. One of the reasons why a person feels some apprehension about visiting the dentist—one of the reasons that explains why he or she is consistently putting off the needed appointment—is the virtue of courage, or the lack thereof. Traits are supposed to be explanatory in that it will, at least sometimes, be correct to explain actions in terms of character traits and not just in terms of situational features.^{Footnote 17}

• • Evaluation: third, a virtue is an acquired excellence; it is a matter of choice and achievement; it is something that we admire in a person. The value of a virtue is attached directly to the possessor and secondarily to his or her deeds. Hence, virtue attributions perform an evaluative function. Virtue attributions are forms of normative assessment.^{Footnote 18}

Virtues are thick ethical concepts in which facts and values are entangled, such that they can be and are employed in explanations and evaluations of agents and their actions (Putnam, Reference Putnam2002). As a thick ethical concept, when a virtue is used in judgments (by the right kind of agent), it has the feature that the judgment as a whole is both responsive to how the world is and gives the agent reasons for action (Williams, 1995; McDowell, Reference McDowell1998). We say that a man rescued a drowning child because he is courageous, or opportunistic (because, say, he is trying to impress his girlfriend), or greedy (because, say, he expects a big reward in return), or suicidal. And these are all ways of describing and explaining what he did as well as ways to assess his character.^{Footnote 19}

3. VIRTUE AND REDUCTION

A quick glance at what ethicists, positive psychologists, and management scholars have recently written about virtue may create the impression that they all constitute contributions to a unitary tradition about a single concept. But, as suggested above, they are not.

There are, at least, two basic strategies to conceptualize virtue, namely a reductive and a non-reductive approach.^{Footnote 20} Some philosophers—notably Ryle (Reference Ryle1949: 43) but also Brandt (Reference Brandt1970: 27), Frankena (Reference Frankena1973: 65), Beauchamp and Childress (Reference Beauchamp and Childress1983: 261-5), Becker (Reference Becker1986: 42), Carr (Reference Carr1988: 186), and Rawls (Reference Rawls1971: 192), among others—and most psychologists and organizational scholars (Ross and Nisbett, Reference Ross and Nisbett1991; Peterson & Seligman, Reference Peterson and Seligman2004; Cameron, Dutton, and Quinn, Reference Cameron, Dutton and Quinn2003; Cameron & Spreitzer, Reference Cameron and Spreitzer2011; Goodwin, Piazza, & Rozin, Reference Goodwin, Piazza and Rozin2014), explicitly or tacitly (and oftentimes unintentionally) hold a reductive account of the virtues, according to which virtues are, roughly, behavioral dispositions to act in conformity with certain rules of action.

Generally, we say that an object has a certain disposition if and only if the object would produce the associated manifestation when under the conditions of manifestation (Fara, Reference Fara2005; McKitrick, Reference McKitrick2005). Hence, a disposition is an entirely hypothetical property that makes references to counterfactual states of affairs (Armstrong, Reference Armstrong1969; Armstrong, Martin and Place, Reference Armstrong, Martin and Place1996; Bird, Reference Bird1998; Cross, Reference Cross2005).

The dispositional or reductive view holds a linear conception of virtue, whereby a virtue is a single linear trait that bears a one-to-one correspondence with certain behavior in any situation where it is relevant (Hampshire, Reference Hampshire1953; Bird, Reference Bird1954; Brandt, Reference Brandt1988). Whether a trait is indeed a virtue or a vice is established, the argument goes, by reference to the moral rule that determines the quality of the action associated with that trait. According to the dispositional view, the virtue of honesty can be reduced to a disposition to obey the moral rule that forbids someone, among other things, to tell lies or intentionally deceive others. Likewise, the virtue of benevolence is fundamentally a disposition to perform actions in accordance with the duty of benevolence.

There are two polemics over reductionism, namely, a philosophical polemic over the reduction of virtue to duty categories and a methodological polemic over the reduction of virtue to actions in empirical studies. The first polemic is concerned with the priority of deontic or aretaic concepts in the conceptualization of virtue: while some philosophers take deontic concepts such as “ought,” “right,” and “duty” to be the fundamental concepts of ethics (and so define virtue in terms of deontic notions), others take aretaic notions such as “admirable” and “excellent” to be the fundamental ones. The second polemic is about the role of character in the conceptualization of virtuous actions: while some psychologists define a virtuous action in terms of the qualities of the action, others are more concerned with the agent performing the action. So, the first raises the question whether the reduction of virtue to deontic categories fails to treat character seriously. And the second raises the question whether reducing virtue to actions may lead to either blindness to virtue—where it may actually be present—or seeing virtue where there is none, because factors behind relevant behavior are overlooked.^{Footnote 21}

The reductive view can be said to comprise two distinctive theses: a normative thesis and an observational thesis.

The normative thesis holds that virtues are dispositions to act in conformity with certain rules. For example, in the consequentialist camp, Sidgwick famously argues that the virtue of veracity involves “a settled endeavor to produce in the minds of others impressions exactly correspondent to the facts, whatever his motive may be for so doing” ([1907] 1981: 224). And Driver holds that virtue “is a trait of character that we value because we recognize its good-producing qualities (1996: vii). Likewise, deontologists claim that “moral virtues are dispositions to avoid unjustified violations of the moral rules” and that “to have a moral vice is to have a disposition to unjustifiably violate a moral rule.” (Gert, Reference Gert2005: 184) In sum, the reductive view defends the claim that “to have a moral virtue is to be disposed to act as moral rules direct.” (Gewirth, Reference Gewirth1978: 339)

The observational thesis holds that virtue entails an expectation of consistent behavioral manifestations of the virtue as a necessary condition of its existence. Pervin claims that “a kind person will act kindly even when there is no situational pressure or external reward for doing so, thus suggesting some kind of internal process or mechanism that is producing the behavior” (Pervin, Cervone, & John, Reference Pervin, Cervone and John2005). A virtue is a “relatively stable and long-term disposition to act in distinctive ways” (Harman, Reference Harman1999: 317). Or, as Doris approvingly quotes from Woods (Reference Woods1986: 149), it is “a disposition to act unfailingly in a virtuous manner” (2002: 176).^{Footnote 22} Character attribution is, according to this position, a means of saying that the person will act in a certain way across time and situations (Epstein, Reference Epstein, Zucker, Aronoff and Rabin1984; Ajzen, Reference Ajzen2005).^{Footnote 23}

Whether done purposely or not, psychologists in practice endorse the reductive view.^{Footnote 24} They do so presumably because the conceptualization of virtues as behavioral dispositions makes their job easier in terms of operationalizing the constructs that are supposed to be the virtues. Personality psychologists can be said to understand virtues as behavioral dispositions in that they are more focused on classifying behavior than theorizing about mental states. For instance, there is extensive research on a classificatory scheme in personality psychology (Cattell, Reference Cattell1943; Eysenck, Reference Eysenck1991; Costa and McCrae, Reference Costa and McCrae1985) that attempts to identify a minimum number of dimensions along which people differ by using factor analysis techniques on questionnaire data rather than by using a priori conceptualization.^{Footnote 25}

Positive psychologists also define character strengths in behavioral terms (Seligman and Csikszentmihalyi, Reference Seligman and Csikszentmihalyi2000, Seligman, Reference Seligman2011). For example, Peterson and Seligman define character strengths as “the psychological ingredients—process or mechanisms—that define the virtues” or “distinguishable routes to display one or another of the virtues” (Reference Peterson and Seligman2004:13). To illustrate this claim, they argue that the virtue of wisdom can be achieved through such strengths as creativity, curiosity, love of learning, and open mindedness. In their seminal work, they conclude that someone is of good character if he or she displays the strengths within a virtue group. A character strength, then, must meet some criteria in order to be considered a virtue.^{Footnote 26} The following five criteria are the most relevant features for the purpose of this section (Peterson and Seligman, Reference Peterson and Seligman2004: 17-24):

1. 1. Character strengths are like personal traits that a person owns, celebrates, and frequently exercises.^{Footnote 27}

2. 2. “Each strength is morally valued in its own right, even in the absence of obvious beneficial outcomes.”

3. 3. “The display of a strength by one person does not diminish other people in the vicinity.”

4. 4. “A strength needs to be manifest in the range of an individual's behavior… it should be trait-like in the sense of having a degree of generality across situations and stability across time.”

5. 5. Another criterion for a character strength is that people differ in their strengths.

Along the same lines, organizational scholars working on the Positive Organizational Scholarship tradition endorse the dispositional analysis of character strengths as virtues, although they seem to do so for pragmatic reasons. Scholars in this tradition are developing an array of instruments to measure character strengths (Cameron, Reference Cameron and Bakker2013), so they basically adopt the positive psychology’s framework without any conceptual analysis of the nature of such character strengths.

Both positive psychologists and POS scholars define virtue in behavioral terms. For example, the commendable work on the character strength of courage in the POS tradition (Worline and Quinn, Reference Worline, Quinn, Cameron, Dutton and Quinn2003; Worline, 2003) is primarily concerned with courageous behavior. In the absence of conceptual analysis, one has reasons to believe that POS scholars understand the virtue of courage as a disposition to act following a principle of courage, which would go along the lines of overcoming fear and voluntarily engaging in worthy action. In their words: “the effectiveness of each organizational form may only be viable over the long term if participants exercise the courage necessary to act from principle” (Worline and Quinn, 2002: 157). In the same vein, POS research on the character strength of compassion in the workplace is defined as “an expression of an innate human instinct to respond to the suffering of others” (Dutton, Worline, Frost, and Lilius, Reference Dutton, Worline, Frost and Lilius2006: 60), and POS research on the character strength of gratitude assumes that gratitude is just a behavioral disposition to act according to the principle of expressing thanks for the benefits one has received. As Emmons puts it, “in psychological parlance, gratitude is the positive recognition of benefits received” (Reference Emmons, Cameron, Dutton and Quinn2003: 82).

One can certainly understand the need for some degree of reductionism by empirical researchers. One of the most decisive arguments in favor of the reductive approach in positive psychology and POS is that it makes conducting empirical studies on character strengths easier. (Bright, Winn, & Kanov, Reference Bright, Winn and Kanov2014). The claim is that we do not need to provide a conceptual analysis of virtue because trait attributions are made—according to the reductive model—purely on the evidence of behavior. Why so? Because the classification of types is based on observable behavioral regularities, which are made for the purpose of predicting future behavior. Then, even if behavior is caused by internal factors, we do not need to know those factors; we can classify people and make predictions without recourse to them. Consequently, the reductive view holds that the only emphasis on explanation should be at the level of behavior, whatever the underlying causes of behavior. Given that trait attribution has a one-to-one correspondence with individual differences—so the reductive argument goes—there is no need to distinguish between trait attribution and the underlying individual differences that cause behavior. Peterson and Seligman suggest this point:

For the time being, such a theory (or theories) does not exist.

Now, what about the philosophers who advance a reductive theory of virtue? Unlike the psychologists who may be unaware of what is at stake here, philosophers do know what reduction amounts to. Why are these highly competent philosophers—who would not think of embracing behaviorism in the realm of perception and thought—behaviorists about virtue? We have reasons to suspect it is because they do not think virtues amount to anything interesting; they are committed to the claim that virtues do not add much to moral theory (to the sort of moral theory they advocate).

There are, in sum, at least two basic ways to conceptualize virtue, namely a reductive and a non-reductive approach. Most psychologists and organizational scholars, as well as a number of (non-virtue) ethicists, endorse the reductive (or dispositional) conception of virtue, which entails a twofold claim, namely:

1. (A) virtues are behavioral dispositions to act as moral rules direct (Gewirth, Reference Gewirth1978) and

2. (B) virtues reliably lead to trait-relevant behavior in the appropriate eliciting conditions (Doris, Reference Doris2002).

It follows that, under the reductive view, moral virtues derive their content and normative force from deontic concepts. In other words, virtue concepts have merely an instrumental, secondary role in moral theory (namely, to ensure conformity with the right moral rules).

4. WHAT IS WRONG WITH REDUCTION?

In spite of its ostensible practical advantages, there are four serious objections against the reductive account of virtue.

First, the reductive account has weak explanatory power and, hence, virtue attributions do not offer good explanations of behavior in organizations. The reductive account provides a simple, linear model of virtue. Under the reductive model, ascribing a virtue entails placing an individual on a ranking according to how much trait-relevant behavior the individual exhibits. Consider the virtue of honesty. In the famous studies on academic cheating, Hartshorne and May (Reference Hartshorne and May1928) measured cheating by counting how many points a child adds to his or her score when marking his or her own test. Each observation is supposed to provide a representative measure of the virtue, which must correlate with quantitative measurements of other behaviors that supposedly measure the same trait in other scenarios. According to the reductive view, a child who is honest will be consistent in not adding points. And the more dishonest a child is, the more points he or she will necessarily add. Ultimately, the number of points added is the measure of dishonesty, which is supposed to be consistent across situations and time for each child. The same sort of dispositional analysis is used to attribute virtues such as benevolence, compassion, and others in the experimental literature in social psychology (Isen and Levin, Reference Isen and Levin1972; Darley and Latane, Reference Darley and Latane1968; Latane and Rodin, Reference Latané and Rodin1969; Darley and Batson, Reference Darley and Batson1973; Milgram Reference Milgram1974) and organizational behavior (Reference Treviño and YoungbloodTrevino and Youngblood, 1990; Staw, Bell, and Clausen, Reference Staw, Bell and Clausen1986; Chatman and Barsade, Reference Chatman and Barsade1995; Barrick and Mount, Reference Barrick and Mount1991).

The reductive strategy has weak explanatory power because it does not have much to say about the interaction of more than one disposition. People are more complex than the reductive account suggests. Human beings have multiple dispositions that might lead to conflicting courses of action. But the reductive strategy does not take into account the possibility that two different traits, thought of as internal individual differences, may come into conflict in a particular situation. A number of different character traits may be associated with a particular behavior; for example, the virtue of honesty and the virtue of loyalty may be both associated with truth telling. Conversely, a single trait may explain inconsistent actions; for example, the virtue of integrity may regularly lead to truth telling but in other cases may require lying (e.g. in case of a serious emergency). Finally, there might be conflicting traits leading to a single behavior (Sabini & Silver, Reference Sabini and Silver2005). If two or more virtues conflict in a situation, behavior may appear inconsistent, but that is just an illusion resulting from assessing behavior with reference to one single trait (Alzola, Reference Alzola2012).^{Footnote 28}

Consider someone who is utterly fearless in two different situations. In the first, the person is being assaulted by an unarmed mugger; in the second, he is being assaulted by six armed muggers. That person’s state of mind is in some obvious sense identical in the two cases, but whereas in the first case he is showing courage—i.e., a virtue—in the second case he is showing foolhardiness—i.e. a vice. When we say that virtue is a mean, we are referring to the when, the how, the to whom, etc. in which an action may happen. So, virtues do not supervene merely on psychological states but rather on psychological states in contexts. A courageous person may experience fear in the case of the many muggers but not in the case of the one or may run in the first case but not in the second and still be a virtuous person exhibiting virtue in both cases.^{Footnote 29}

The second objection holds that a dispositional account of the virtues does not have much to say about the interaction of traits and other mental states in explanations of behavior. If a belief is a tendency to do a certain act given a certain desire, and a desire is a tendency to do a certain act given a certain belief, a reductive account does not provide a good answer for why a person performed (or failed to perform) such an action given his or her desire to perform that action and his or her beliefs that such an action leads to the desired outcomes (Kupperman, Reference Kupperman1991). To illustrate this claim, consider the previously mentioned experiments in social psychology. If a belief that one must help someone in need (if possible)^{Footnote 30} is a tendency to help someone in need given a certain desire to help someone in need, and a desire to help someone in need is a tendency to help someone in need given a certain belief that one must help someone in need (if possible), then, the subject’s failure to help someone in need does not provide any explanation of his or her behavior. The commonsense solution—that the agent can probably tell us what he or she wants and what he or she believes—is unattractive to the scientist who does not countenance inner states reportable by the agent.^{Footnote 31}

The third objection holds that a reductive account of virtue does not provide the resources to account for virtue attribution in the absence of behavior. Specifically, the reductive account is question-begging when it comes to assess virtues in the absence of the typical behavioral manifestations associated with the virtue. Aristotle argues that “(…) it seems possible for someone to possess virtue but be asleep or inactive throughout his life...” (EN, 1095b34-1096a2). The possession of a good character trait may not have an external manifestation, for example, when environmental conditions do not provide the opportunity for the trait to be displayed, regardless of the strength of the trait (Sonnentag & McDaniel, Reference Sonnentag and McDaniel2013). This claim will surely be resisted by reductivists.^{Footnote 32} As explained in the previous section, thesis (B) of the reductive account holds that it makes no sense to attribute a certain character trait X to a person who has never behaved in an X-like manner (Alston, Reference Alston1975; Doris, Reference Doris2002). The person’s failure to act according to virtue is enough to prove that he or she does not possess such a virtue (see footnote 23 above). And the person’s action, conforming to the moral rule or principle, is necessary as evidence of virtue.

Regrettably, the reductive account blurs the very important distinction between character attribution and the evaluation of actions, that is, between the possession of a virtue and an action in conformity with virtue. For example, positive psychologists, on the one hand, claim to be concerned with the notion of “a virtuous person” (Seligman, Reference Seligman2002: 137), but then they write “we are comfortable saying that someone is of good character if he or she displays but 1 or 2 strengths within a virtue group” (Peterson and Seligman, Reference Peterson and Seligman2004: 13). And POS scholars are satisfied with what they called virtuousness, which is a new label for what philosophers would plainly call “right action.” Even if it were possible to infer a person’s character from a single action,^{Footnote 33} a theory that is committed to the claim that virtues are just statistical measures of past behavior fails to capture the robust meaning of virtue as an excellence of character, which is the fundamental concept in the virtue ethics tradition.

The fourth objection is that if virtues are reduced to dispositions to follow certain ethical rules, the essence of virtue is lost, so we do not need any talk of moral virtue. The reductive view is serving the cause of rule-based normative theories and thereby misusing the language of virtue. If morality is just about rightness, if the fundamental question of morality is the question of what to do, then it follows that duties and principles have normative priority (and character has a secondary role in moral theory). The reductive strategy advises us to find and justify fundamental rules and principles. The moral virtues have, at most, an instrumental role. One ought to develop certain dispositions to reliably act in conformity with such rules and principles (Harman, Reference Harman, Byrne, Stalnaker and Wedgwood2001). As a result, the reductive account does not give us the tools to account for a critical distinction between acting from virtue and acting in conformity with virtue (Audi, Reference Audi1995). Acts that are merely in conformity with virtue may qualify as instances of what the POS scholar calls “virtuousness” (Cameron & Winn, Reference Cameron, Winn, Cameron and Spreitzer2011; Bright, Cameron, & Caza, Reference Bright, Cameron and Caza2006), but they are not genuinely virtuous. For only actions from virtue bespeak a feature of good character (Korsgaard, Reference Korsgaard, Engstrom and Whiting1996).

Even the most dishonest person could perform what looks like an honest action (for the sake of a reward, for example) and even a coward could perform a pseudo-act of courage—done for the sake of impressing bystanders (or done under the influence or by mistake). If the person’s motives behind his or her actions are irrelevant to the evaluation of his or her actions, then it follows that we do not need a talk of virtue at all. Neither do we need character. The language of rightness would be enough to do the job.

In a sense, the reductive account is the sort of “ethics without persons” that some scholars advocate (Doris, Reference Doris2009). But it was precisely as a reaction to the ethics-without-persons idea that the contemporary virtue ethics tradition has been developed since 1958! Virtue ethics is an ethics of being, of being good as a person.

5. REAL VIRTUE

When taken together, the four previous objections make a strong case against reduction. But discrediting the reductive account of virtue is only half the story. We need to examine what a non-reductive strategy of virtue has to offer and make a case for the claim that a comprehensive account^{Footnote 34} does not fall to the four aforementioned objections. I began articulating that account elsewhere (Alzola, Reference Alzola2012). Here I shall briefly outline the framework and extend it to distinguish virtue from “character strengths” and “virtuousness.”

A non-reductive account holds that a virtue is not a disposition to behave in accordance with certain rules of action. For an action to be from a state of virtue—for an action to bespeak a mark of good character—it must be expressive of appropriate inner states. The reductive account reduces virtue to its behavioral aspects, thereby neglecting its inner dimension. Virtue ethicists, on the contrary, highlight the understanding of virtue as “an inner quality of an agent and of his acts” (von Wright, Reference Von Wright1963), thereby integrating the cognitive, the emotional, and the motivational components of virtue. What constitutes an honest person, for instance, is not exhausted by any catalogue of possible honest actions he or she may perform. Being honest involves qualities of character proper to the honest person, and it is these qualities that establish what actions honesty requires.

If virtuousness is “conduct that conforms to an accepted standard of right and wrong,” then virtuousness cannot be “virtuous” because an action in conformity to virtue does not necessarily bespeak a feature of good character.

Virtue is not merely knowledge about what is courageous, just, or honest. Neither is it only the ability to align one’s desiderative states with what one believes. It is not only being motivated to do what is courageous, just, or loyal for the right reasons. Neither is it just a disposition to do what is courageous, just, or honest. Virtue is all of these or it is nothing. Thus, when seen through the notion of virtue articulated in Section Two, the reductive strategy is conceptually inaccurate and methodologically misleading.

My conception of virtue is non-reductive because it holds that the four elements of character are not reducible to each other. Although they are interrelated, none of them can be described solely in terms of one of the others. The virtues, in sum, cannot be plausibly reduced to dispositions to behave in conformity with certain principles and rules without losing the substance of the very idea of virtue.

Because it can accommodate the interplay of multiple virtues as well as the interaction between virtues and other mental states—by integrating the cognitive, emotional, and motivational components of virtue to understand the agent’s action—a non-reductive approach to virtue has greater explanatory power than the reductive view.^{Footnote 35} And because it is concerned not only with behavioral tendencies but also with the other three dimensions of virtue, a non-reductive account to virtue fares better than the dispositional theory when it comes to assessing cases of virtue attribution without behavior. For example, even if the situation does not provide the opportunity for behavioral manifestations, the agent may frame (or fail to frame) the situation appropriately, he or she may experience (or fail to experience) certain emotions (from anger to remorse) in response to the situation, he or she may deliberate and make proper judgments, etc. Virtue does not necessarily require “trait-relevant behavior in the appropriate eliciting conditions.”^{Footnote 36} Depending on the scope of “appropriate,” we might say that the absence of such behavior calls for an explanation. But an action—understood as the presence of such a behavioral manifestation—is not sufficient to infer the presence of a virtue (Aristotle, EN 1105a25). More needs to be said about the motives and intentions of the person, about his or her deliberation and reflection, about his or her emotions and values, and about his or her ends and motives (Nussbaum, Reference Nussbaum, French, Uehling and Wettstein1988). Moreover, the reductive account rests on a distinction between virtue and virtuous action that is unintelligible. And because the non-reductive account is definitionally character-based, it recognizes the normative priority of aretaic concepts over deontic notions. Acting virtuously, acting from virtue, is just a way to align the four dimensions of virtue identified in Section Three.

Can virtue, understood in this non-reductive way, be measured? A complex question requires a long answer. In principle, it depends on whether the four dimensions of virtue can be further analyzed independently of each other. If the comprehensive account of virtue defended in this section can be broken down into separate parts, and if it is methodologically feasible to examine virtue as a cluster of separate yet integrated components, then the answer is “Yes.”

Most virtue ethicists—especially those who endorse the unity of the virtues thesis (e.g. von Wright, Reference Von Wright1963; Irwin, Reference Irwin1988; Cooper, Reference Cooper1998, McDowell, Reference McDowell1998)³⁷—will be against this approach. Still, some virtue scholars—such as Robert Adams and Nancy Snow—may be more optimistic.

Adams (Reference Adams2006) favorably considers the possibility of micro-virtues that come in “modules,” that is, independent component parts from which a complex and more general structure is or can be assembled. Some domain-specific behavioral dispositions—for example, a disposition to behave honestly or courageously in only a single area of life such as one’s professional life, when motivated by a proper appreciation of the values of professional honesty or courage—can be, according to Adams, “modules of virtue.” Such modules might be added together to form a more complex and general disposition that may qualify as virtue.^{Footnote 38}

Snow (Reference Snow2010) suggests decomposing virtue into three measurable parts, which she calls virtue intelligence, dispositionality, and behavior. She claims that virtue intelligence can be operationalized through the measurement of the agent’s sensitivities and abilities to recognize and generate appropriate responses in the cognitive, affective, and behavioral domains (Snow, Reference Snow2014). These sensitivities can be measured using techniques that keep track of people’s perceptions of virtue-relevant stimuli in actual or fictional social interactions. One example is to ask people to watch a clip of another person’s social interactions and hit a button any time they observe a virtue-relevant stimulus. Upon completion, the investigator will ask the participants to go back through their hits and explain what they take the virtue-relevant stimuli to be and why. Another idea is to use diary methods whereby the investigator asks people to recall virtue-relevant stimuli from their past experiences and how they responded to such stimuli. Participants’ responses can be recorded using daily diary entries or portable electronic devices. The inspiration for these and other ideas to measure different components of virtue is the work of psychologist Matthias Mehl and his collaborators (Mehl and Conner, Reference Mehl and Conner2012). Mehl has pioneered the use of a computerized tape recorder, known as the Electronically Activated Recorder (Mehl & Pennebaker, Reference Mehl and Pennebaker2003), “to record and analyze virtue-relevant behavior in daily life.” Specifically, they are using the EAR to test “the stability, variability, and changeability of virtuous daily behavior relative to (non-virtuous) neutral and negative daily behavior using a novel, naturalistic observation sampling method” (Snow, Reference Snow2014: 26).^{Footnote 39}

Full elaboration of this point would deserve its own paper.^{Footnote 40} What we must retain is that in virtue ethics, there is no virtuous action without a virtuous person. We are to understand what it means to behave virtuously through the study of the nature and inclinations of the virtuous person, not through the observation of seemingly virtuous actions. In other words, we do not define virtuous actions as those that conform to certain moral rules and, then, define a virtuous person as one who reliably and characteristically perform such actions. Qualities of character are normatively more basic than rules and actions.

Being virtuous and acting like a virtuous person must be distinguished, because even the most vicious person may behave like the virtuous one. The reductive account does not have the resources for a negative evaluation of the vicious person’s actions. In contrast, according to the non-reductive view, a person who is virtuous has a standing commitment to acting from virtue, in the sense that her cognitive, emotional, motivational, and behavioral states are all well aligned. Importantly, under the non-reductive view virtue is more than just continence: the continent person actually fears what should not be feared—a trip to the dentist—and overcomes this fear but the courageous person simply does not fear it at all. To be virtuous is to be without fear where courage is called for; to be continent is to overcome one’s fear (EN 1119a10ff).

6. GRATITUDE AND THE GRATEFUL PERSON

One might object that much of my argument relies on distinctions that may not make a significant difference in practice. In this section, I shall illustrate the differences between the reductive and the non-reductive accounts by discussing an example of a well-regarded virtue: the virtue of gratitude. We could apply this framework to any other virtue—elsewhere I have examined honesty and compassion (Alzola, Reference Alzola2008, Reference Alzola2012) and in the section above, I have suggested how to analyze the virtues of courage and benevolence—but gratitude makes it easier to compare reductive and non-reductive approaches.

According to the “Values in Action Inventory of Strengths” (VIA), gratitude is defined as “being aware of and thankful for the good things that happen; taking time to express thanks” (Peterson and Seligman, Reference Peterson and Seligman2004: 30).^{Footnote 41}

Gratitude has a very important place in the POS literature. It is hypothesized to lead to positive individual and organizational outcomes, such as mood and prosocial behavior, and also to serve as an antidote to “toxic workplace emotions,” such as anger and envy (Emmons Reference Emmons, Cameron, Dutton and Quinn2003: 90).

Psychologists predominantly use two instruments to “scientifically measure gratitude,” namely, the GQ-6 (Gratitude Questionnaire-6) and the GRAT (Gratitude, Resentment, and Appreciation Test) (McCullough, Emmons, & Tsang, Reference McCullough, Emmons and Tsang2002; Watkins, Woodward, Stone & Kolts, Reference Watkins, Woodward, Stone and Kolts2003). The studies apparently indicate that gratitude is a reliable predictor of subjective well-being and other positive qualities felt by the beneficiary (Emmons and McCollough, Reference Emmons and McCullough2003; Algoe, Haidt, & Gable, Reference Algoe, Haidt and Gable2008; Bartlett, Condon, Cruz, Baumann, & Desteno, Reference Bartlett, Condon, Cruz, Baumann and Desteno2012). Designers and users of these instruments claim to have used “strong theoretical grounds to develop an a priori conception” of gratitude (Wood, Maltby, Stewart, and Joseph, Reference Wood, Maltby, Stewart and Joseph2008). Yet, according to a study of recent papers on gratitude in psychology, most of them merely rely on definitions by established scholars—notably, Emmons and McCollough—without providing a conceptual analysis of what gratitude is. And some studies do not even offer a definition of gratitude, although they examine its correlates and positive benefits! (Kristjansson, Reference Kristjánsson2013).

Following the reductive strategy, one would define the virtue (or character strength) of gratitude as a disposition to follow a rule of action, which would go along the lines of a duty of gratitude (Emmons, 2004). In turn, one could define the duty of gratitude as a Rossian prima facie duty: “We should express gratitude, in deed or at least in words of thanks, in a way that befits good things done for us by other people, where, other things equal, our obligation is stronger if what was done for us was not owed to us” (Ross, Reference Ross2002; Audi, Reference Audi2009). Or, in the positive psychologist’s version, the duty of gratitude is defined as requiring that anytime one is benefited by another person, one ought to recognize the benefits received and express gratefulness (Emmons, Reference Emmons, Cameron, Dutton and Quinn2003: 82). Then, the virtue of gratitude would be a disposition to act in conformity with the moral duty of gratitude (Emmons & McCullough, Reference Emmons and McCullough2004; Lambert, Graham & Fincham, Reference Lambert, Graham and Fincham2009).

The problem is that the reductive approach fails to describe gratitude in the way we colloquially speak of it and the way virtue ethicists understand the virtue of gratitude. For, even if it is true that a malevolent person may perform an act that looks like an act of benevolence, and even if a coward may perform an act that looks like a courageous act, in the case of gratitude it seems that the very notion of gratitude is tied to the inner component of the virtue (Walker, Reference Walker1988; McAleer, Reference McAleer2012). What constitutes an act of gratitude is indeed the motive from which it is done, that is, the intellectual, emotional, and motivational qualities of character of the person, besides his or her behavior (Berger, Reference Berger1975; Fitzgerald, Reference Fitzgerald1998).^{Footnote 42}

In other words, in order to perform an act of gratitude, one must be grateful. It is not sufficient to perform the action that is required by the duty of gratitude. Acting from such a duty cannot serve as a substitute motive without significantly altering the nature of an act of gratitude. Even if we agree that an act of insincere gratitude is better than manifesting no gratitude at all—although the case for this has yet to be made—the existence of a duty to act as if one is grateful fails to capture the essence of the virtue of gratitude (Card, Reference Card1988). Consequently, a disposition to act in conformity with a duty of gratitude is not going to do the work of gratitude as a virtue (Wellman, Reference Wellman1999).

In order to perform an act of gratitude, a person needs to be grateful, i.e., he or she needs to possess the virtue of gratitude. What makes an act into an act of gratitude—what makes us admire such an act as a praiseworthy act—is the state of character of the person who performs that act. That state of character is what we know as the virtue of gratitude. Virtue concepts have, in sum, priority over deontic concepts.

The reductive account of virtue is blind to the mental process underlying an act of gratitude (Weiss, Reference Weiss1985). Hence, the view that gratitude—as a character strength—is simply a behavioral disposition to behave according to the duty of gratitude does not properly describe our experience of gratitude, and it does not align with the best philosophical account of the virtue of gratitude (Carr, Reference Carr2013). Gratitude involves not only appreciating a favor and acting consequently but also appropriate inner states, including beliefs, desires, and attitudes towards one’s benefactor. The virtue of gratitude has intellectual, emotional, motivational, and behavioral dimensions. It entails some framing capacities to grasp the goodness and benevolence of the benefactor’s actions together with appropriate feelings and the willingness to be indebted to others (which requires other related virtues such as generosity and humility).^{Footnote 43} Gratitude also involves appropriate desires to return acknowledging tokens of benefit.

Ultimately, gratitude may not require an external manifestation. A person may be grateful without expressing thanks. A person may (wrongly) be grateful for things that harm him or her. And sometimes a person may have reasons not to accept some benefits and to refuse to express thanks to a benefactor whose intention is blatantly wicked (Fitzgerald, Reference Fitzgerald1998). For example, a benefactor may aim to cause harm to the beneficiary or third parties (Roberts, Reference Roberts1995).

Praise for gratitude and criticisms for failures of gratitude are invariably evaluations of the four dimensions of character. It is precisely this focus on good and bad character rather than unmet obligations and behavioral dispositions to act in conformity with such obligations that makes gratitude a virtue.

In sum, the virtue of gratitude does not conform to the reductive program, a form of behaviorism that fails to properly distinguish inner from outer states, cashing the former into the latter. An act of gratitude is characteristically defined as an action motivated by the virtue of gratitude. It cannot be performed from a sense of duty without significantly altering the nature of the action. As explained in Section Two, a virtuous action is virtuous precisely because of the intellectual, affective, motivational, and behavioral states from which it stems.