Situationism and Confucian Virtue Ethics
- First Online:
- Cite this article as:
- Mower, D.S. Ethic Theory Moral Prac (2013) 16: 113. doi:10.1007/s10677-011-9312-9
- 583 Views
Situationist research in social psychology focuses on the situational factors that influence behavior. Doris and Harman argue that this research has powerful implications for ethics, and virtue ethics in particular. First, they claim that situationist research presents an empirical challenge to the moral psychology presumed within virtue ethics. Second, they argue that situationist research supports a theoretical challenge to virtue ethics as a foundation for ethical behavior and moral development. I offer a response from moral psychology using an interpretation of Xunzi—a Confucian virtue ethicist from the Classical period. This Confucian account serves as a foil to the situationist critique in that it uncovers many problematic ontological and normative assumptions at work in this debate regarding the prediction and explanation of behavior, psychological posits, moral development, and moral education. Xunzi’s account of virtue ethics not only responds to the situationist empirical challenge by uncovering problematic assumptions about moral psychology, but also demonstrates that it is not a separate empirical hypothesis. Further, Xunzi’s virtue ethic responds to the theoretical challenge by offering a new account of moral development and a ground for ethical norms that fully attends to situational features while upholding robust character traits.
KeywordsSituationismConfucianVirtue ethicsMoral developmentXunziCharacter traitsMoral psychologyRitualDorisHarman
Situationist research in social psychology focuses on the situational factors that influence behavior.2 Doris and Harman argue that this research has powerful implications for ethics, and virtue ethics in particular. First, they claim that situationist research presents an empirical challenge to the moral psychology presumed within virtue ethics. Second, they argue that situationist research supports a theoretical challenge to virtue ethics as a foundation for ethical behavior and moral development. Virtue ethicists did not take such challenges lightly, and a vast literature developed in response. Many responses address the conception of character or virtue assumed by Harman and Doris, while others target the implications drawn from and the quality of situationist research. Doris refined the situationist challenges in his recent book Lack of Character and argued that “divesting ethical reflection of an empirically discredited psychology of character will facilitate emotional, evaluative, and deliberative habits that are more defensible, more sensitive, and more conducive to ethically desirable behavior” (2002, p. 2). The debate rages on at near impasse, both sides fully entrenched.
Unsurprisingly, the debate is framed via Western conceptions of character and virtue;3 after all, the protagonists are Western scholars embroiled over the implications of psychological research for modern ethical theory. In contrast, I offer a response from moral psychology using a Confucian conception of virtue ethics. Xunzi is an early Confucian philosopher who is very aware of situational effects on behavior, and builds this awareness into his account of moral development and program of moral education. I suggest that Xunzi holds an “embeddedness model” of moral development,4 in which deliberation focused on and embedded in the determinative features of situations is the normative ground for ethics and the apex of moral development. Consequently, one project of this paper is to present and argue for this interpretation of moral development in Xunzi’s account.
Secondly, I argue that this account serves as a lens to critique key ontological and normative assumptions about moral psychology and moral development within the arguments posed by the situationists Harman and Doris. In a sense, Doris and I share the same project: we each want to divest virtue ethics of problematic assumptions and then address the implications of rejecting those assumptions for moral education. However, we part company in the end: my project is to divest Harman and Doris of their problematic assumptions using an interpretation of Xunzi’s model of moral development. Although my argument stems from an account of Confucian virtue ethics, the scope of this paper is quite broad in that it not only deflates the situationist challenges, but it also invites a revision of any account of virtue ethics that relies on these unwarranted and problematic assumptions.5
Finally, I argue that Xunzi’s account offers a unique response to the situationist challenges and a new avenue to conceive of virtue ethical moral psychology and moral education. Xunzi’s account responds to the situationist empirical challenge by uncovering problematic assumptions about moral psychology, and to the theoretical challenge by offering a new account of moral development that fully attends to situational features. What is so unique and promising about Xunzi’s account is that it not only provides a normative ground for deliberative situational embeddedness without making the normative and ontological assumptions that plague situationism (and indeed which, situationists charge, plague other accounts of virtue ethics), but also offers a way to move forward by putting situationist research from social psychology in the service of detailing a Xunzian account of virtue ethics.
2 The Challenges of Situationism
Doris and Harman mount their empirical challenge to Western accounts of virtue ethics by critiquing globalism: an account of moral psychology committed to the theses that virtue or character traits are stable, consistent, and integrated. On a dispositionalist account, virtues or character traits such as courageousness or honesty produce behavior because they are psychological dispositions to behave in specified ways. Doris explains that character traits are stable when they are “reliably manifested in trait-relevant behaviors over iterated trials of similar trait-relevant eliciting conditions” (2002, p. 22). For example, if I am honest, I would return extra money to a cashier whether (a) in Starbucks, (b) in the grocery store, or (c) at the bank. He defines character traits as consistent when they are “reliably manifested in trait-relevant behavior across a diversity of trait-relevant eliciting conditions that may vary widely in their conduciveness to the manifestation of the trait in question” (Doris 2002, p. 22). For example, if I am honest, I would (a) return money to the cashier upon receiving extra change, (b) report all my income to the IRS, and (c) register for all philosophy conferences attended. Character traits are integrated when “the occurrence of a trait with a particular evaluative valence is probabilistically related to the occurrence of other traits with similar evaluative valences” (Doris 2002, p. 22). For example, if I think my friend is trustworthy, I may also assume that she is honest. Doris explains that to attribute a virtue is to assert a conditional: “If a person possesses a trait, that person will exhibit trait-relevant behavior in trait-relevant eliciting conditions with markedly above chance probability p” (2002, p. 19). However, this honesty-attribution does not merely express a correlational claim that my friend will display honesty in honesty-eliciting situations, but that the reason she returns the money is because she is honest; honesty, as a psychological trait, is causally efficacious for or productive of her behavior. Attributing virtues to others not only helps to explain their behavior, but it also allows for fairly reliable predictions. When a cashier accidentally returns an extra $20 in change to my honest friend, I can predict with some confidence that she will return the money despite the fact that she recently lost her job.
Doris explains that stability, consistency, and integration are the three key theses of globalism. However, he notes that they are three separate theses: stability and consistency are theses about the nature of traits, while integration is a thesis about personality and trait organization (2002, p. 23). My interests in this paper focus on the first two theses of globalism: the characterization of virtues as stable and consistent psychological traits. For Doris and Harman, the lack of stability and consistency in character traits provides strong evidence against virtue ethics, whether dispositionalist or intellectualist. On a dispositionalist account, as described above, character traits are robust when they are stable through time, consistent across situation types, causally efficaious, and allow for reliable predictions and explanations of behavior. On an intellectualist account, virtue is the correct perception of and response to the morally salient features of a situation (e.g., perceiving a situation as one which requires honesty, and then performing the honest action, such as returning the extra change to the cashier).6 Doris claims that the attempt to explain virtue as a perceptual capacity or sensitivity still requires that such traits be stable and consistent in order to be connected to behavior. Consequently, he holds that the same arguments can be leveled against both accounts of virtue ethics.
Although a virtue ethicist might point to regular exhibitions of honest behavior as evidence of the psychological trait or virtue of honesty, Harman and Doris begin the empirical challenge by arguing that there is a competing hypothesis: the behavioral regularity we observe may be due to the stability and consistency of situations rather than stable and consistent character traits (Doris 2002, p. 26; Harman 1999, p. 317). Situationist research, Harman argues, shows that “aspects of a particular situation can be important to how a person acts in ways that ordinary people do not normally appreciate, leading them to attribute certain distinctive actions to an agent’s distinctive character rather than to subtle aspects of the situation” (2003, p. 91). Situationist research illustrates the myriad ways that situations produce or elicit behavior, indicated most clearly by changes in behavior as a consequence of situational variation. As Doris notes, “the problem is not that substantial situational factors have substantial effects on what people do, but that seemingly insubstantial situational factors have substantial effects on what people do” (2002, p. 29). Given this, Harman concludes that “despite appearances, there is no empirical support for the existence of character traits” (1999, p. 330, my emphasis). For Harman, the fact that researchers can manipulate behavior by varying situational features demonstrates that situational regularity, not robust psychological character traits, yields behavioral regularity.7
Doris adopts a more temperate position: given that situationist research casts serious doubts on virtues as globally stable and consistent psychological traits, the empirical evidence favors situational regularity over robust character traits as a behavioral explanation (2002, p. 26). Although Harman abandons character traits, Doris suggests that a modified conception of character traits may have some explanatory value.8 In contrast to robust or global character traits which are stable through time and consistent across a wide variety of situations, Doris proposes that character traits may be local or limited to narrowly specified situations. He explains that the disposition for “battlefield physical courage” would yield courageous behavior whenever one engaged in a physical contest on a battlefield, but not during storms on the high seas, while being tortured, or upon stepping into a den of snakes (2002, p. 62). The fact that one does not behave courageously when confronted with snakes or when tortured is not a mark of instability or inconsistency; rather, it is simply not a disposition that is productive of behavior in such cases. Although Doris doubts the empirical adequacy of global character traits, he thinks that local character traits may offer an explanatory alternative more compatible with situationist research. However, local character traits offer no solace to virtue ethicists. Doris explains that “local traits are not likely to produce the patterns of behavior expected on broad trait categories like ‘compassionate’ or ‘courageous:’ even seemingly insignificant variations in situation may ‘tap’ different dispositions, effecting inconsistent behavior” (1998, p. 507). Local character traits, subject to situational fluctuations, are of no use to accounts of virtue ethics that depend on the situational invariance of global character traits.9
The empirical challenge to robust character traits supports Harman’s sweeping theoretical challenge to virtue ethics as a foundation for ethical behavior and moral development. He argues that an empirically suspect psychological model of character traits undercuts an account of moral behavior defined in terms of virtue. As Harman notes, “if we know that there is no such thing as a character trait and we know that virtue would require having character traits, how can we aim at becoming a virtuous agent?” (2000, p. 224). If virtues do not exist, then the moral ideals assumed by virtue ethics are untenable, and the attempt to develop virtue and character is a waste of time and effort.10 Harman holds no punches.
Harman’s challenge, if it succeeds, goes for the jugular, but Doris’s challenge is no less formidable. Doris charges that “the approach to moral psychology suggested by situationism enjoys certain advantages…as a foundation for normative thought” (1998, p. 505). He suggests that rejecting the assumption of robust global character traits may yield more ethical behavior—a counter-intuitive consequence to many virtue ethicists. Rather than getting into situations likely to elicit unethical behavior and mistakenly relying on the “strength” of one’s character, he argues that situationism suggests attention to the features and avoidance of enticing but sinful situations. For example, quaffing wine while attending a private dinner at the home of a handsome and flirtatious colleague while his spouse is away is a situation loaded with sinfully enticing features (Doris 1998, p. 516). Relying on one’s character is foolhardy. Doris argues that “the way to get things right more often…is by attending to the determinative features of situations,” and that “the implication of this is that our duties may be surprisingly complex, involving…a sort of ‘cognitive responsibility’ to attend, in our deliberations, to the determinative features of situations” (1998, p. 517). His challenge, if it succeeds, undercuts virtue ethics by offering a new foundation for ethical behavior and moral development. In the next few sections, I introduce Xunzi and the importance of ritual, explain how his account of moral development and program of moral education develops attention to situational features as a moral action, and show how his virtue ethic can withstand these empirical and theoretical challenges.
3 Xunzi and the Importance of Ritual
Although Confucius (551–479 BCE) is as well known as Aristotle (384–322 BCE), many Western scholars are less familiar with the philosophers that follow Confucius. Mencius (372–289 BCE) and Xunzi (approx. 310–215 BCE)11 are philosophers of the Classic period of Confucianism. The Classic period is delineated by the Warring States period (479–221 BCE), in which a number of independent feudal states vied for supremacy. The Classic period ended in 221 BCE, when the state of Qin conquered all to create the first unified state and dynasty in China.
As was common for many philosophers of his time, Xunzi served as a political figure, holding posts in several states. In each, he expounded Confucian views on morality, education, and political policy. He also was a leading figure at the prestigious Jixia Academy, akin to a modern “think-tank” for the intellectuals of the day, in which he came into contact with proponents of all the major Sinitic schools of philosophical thought (such as Daoism and Mohism). In contrast to texts like the Analects and the Mencius, which capture aphoristic quotations or record conversations, the Xunzi is a treatise composed of essays on specific topics. In this text, Xunzi develops focused arguments against other schools of thought and other interpretations of Confucianism, and builds a systematic argument with integrated theses on metaphysics and human nature, moral education, political philosophy, philosophy of language, action theory, and aesthetics.12
Mencius and Xunzi offered competing interpretations of Confucianism, and Xunzi’s emphasis on the importance of ritual is often seen as direct response to Mencius and his views on human nature. Mencius famously claimed that human nature was “good,” or that the components of morality are innate. (For a modern Western parallel: one might claim that an empathic ability is innate.) Xunzi is often credited with the contrasting claim that human nature is “bad.” However, his position is not that human nature is evil, but that it is crude, unformed, and contains possibilities for both good and evil: to claim that human nature is “good” is to account for only part of our nature. Xunzi thought Mencius’s views were dangerously simplistic because they implied that morality could develop naturally and minimized the importance of moral education. For Xunzi, moral education is necessary to shape the myriad innate capacities of human nature: without moral education, individuals will not develop into moral persons and society will devolve into chaos (akin to a Hobbesian state of nature).
For Xunzi, ritual forms an ineliminable part of moral education: rituals provide models for individuals and establish order and structure in society.13 Although Western theorists may not be accustomed to thinking in terms of ritual, rituals are ubiquitous. For example, the context of listening to a speech or lecture is a common modern American ritual. Without ever pausing to consider how to act while attending a lecture, one knows not only what to do, but also what to expect; one knows how to behave (e.g., that one sits in the chair and does not stand and dance on it, that one does not speak unless it is necessary and then only at a whisper to a colleague), which establishes expectations for how others will behave as well (e.g., that one should not interrupt the speaker). This implicit knowledge that we all share coordinates our collective action (e.g., comments and concerns of audience members will be addressed in a question and answer session after the lecture, each person will speak in turn and not in a group shouting match, etc.), and our collective action is ordered and orderly (i.e., we each attain what we desire—to hear the lecture—because we engaged in the ritual properly). Although this is just one example, rituals pervade our lives. As Tu notes, Confucians conceive of ritual within “personal conduct, social relations, political organizations, and religious behavior….[Ritual] includes virtually all aspects of human culture: psychological, social, and religious” (1972, p. 198).
Much like the medieval European practice of covering a corpse in flowers to mask the smell, or the modern Western practice of arranging hair, applying makeup, and dressing the corpse in her “Sunday best,” the rituals surrounding Chinese ornamentation in funeral rites ensure that there is no diminution or distortion of natural emotions that should arise when regarding the deceased. Rather than cultivating disgust, funeral rituals provide a supportive context to nurture positive emotional responses such as sympathy, love, and respect.15 Further, the funeral is structured in order to remember and honor the departed: eulogies, short stories, or moments of silence are common ways to honor the deceased and to comfort the living. Because funeral rituals are structured in such a way to elicit positive emotional responses and to provide avenues for remembrance and consolation, they provide interpretive frameworks by making aspects of events salient; although one may be overcome with personal grief, the structure of the funeral (the placement of the family at the front, speeches made by family members, prayers offered in consolation and support to family members, etc.) makes the family’s grief salient and redirects one’s attention accordingly. In addition to providing models for proper behavior, rituals provide models for proper thought by shaping one’s emotional responses, expectations, interpretations, etc.
the standard practice of funeral rites is that one changes the appearance of the corpse by gradually adding more ornamentation, one moves the corpse gradually further way, and over a long time one gradually returns to one’s regular routine. Thus, the way that death works is that if one does not ornament the dead, then one will come to feel disgust at them, and if one feels disgust, then one will not feel sad. If one keeps them close, then one will become casual with them, and if one becomes casual with them, then one will grow tired of them. If one grows tired of them, then one will forget one’s place, and if one forgets one’s place, then one will not be respectful. (Hutton Ch.19: 290–300)14
As comprehensive models for clothing, food, bearing, countenance, emotion, mood, and thought, rituals provide standards for and regulate our individual and collective action. Ritual pervades all aspects and all details of our lives by shaping our emotions, moods, and desires, framing our thoughts, interpretations, and perceptions, structuring our behavior, and mediating our interactions with others. By taking the raw material afforded within human nature (our innate capacities and natural responses) and shaping it, ritual develops individuals into social beings capable of engaging with each other, promoting social order, and seeking personal moral development.17 Consequently, the study and practice of ritual is a necessary component to maintain social order and to develop individuals through moral education.
accord with ritual, they will be ordered and effective. If they do not accord with ritual, they will be disorderly and unproductive. If your meals, clothing, dwelling, and activities accord with ritual, they will be congenial and well-regulated. If they do not accord with ritual, you will encounter dangers and illnesses. If your countenance, bearing, movements, and stride accord with ritual, they will be graceful. If they do not accord with ritual, they will be barbaric, obtuse, perverse, vulgar, and unruly. (Hutton Ch. 2: 35–50)16
4 Moral Development: Rituals as Type-Level Models
Xunzi holds that we all have the same human nature, and hence, are capable of developing through moral education, but he is quite realistic: many will achieve some level of moral development, but not all will. Although the progression is gradual, I argue that Xunzi holds a three-stage “embeddedness model” of moral development,18 in which one progresses ethically by transitioning from type-level models inherent in ritual, to the reflective analysis of and adherence to the “logic” within ritual, to careful evaluative deliberation focused on and embedded in the determinative features of situations or token events. Xunzi explains that “those who take ritual as their model and find sufficiency in it are called men of standards,” while those “who nevertheless do not take ritual as their model nor find sufficiency in it are called standardless commoners” (Hutton Ch. 19:170, my emphasis).19 The “standardless commoners” are immoral persons; they are neither willing, nor can be made, to follow ritual and thereby reject and violate the very models necessary for moral development and social order itself.20 Consequently, those persons who are shaped by and live according to ritual have already achieved the first stage of moral development; they are “men of standards” or “well-bred” because they accept and follow the standards for appropriate behavior and thought. Others go beyond mere practice and approval of ritual, and dedicate themselves to a rigorous program of moral education following the guidance of a teacher.21 These are gentlemen; truly virtuous persons that consciously strive to embody the models in their every thought and action, and continually reflect on them in the course of their training and study. Xunzi explains that “he who focuses his intentions upon…[ritual] and embodies it is a gentleman. He who completely understands it and practices it without tiring is a sage” (Hutton Ch. 2:160, my emphasis).22 Sages are moral exemplars, few and far between.23
In the first stage of moral development, one learns and performs rituals appropriate to a variety of situations. Although there is a vast array of cultural, religious, and interpersonal events, most individuals learn how to behave appropriately in the course of growing up in a given society. One learns these general models through observing others, explicit instruction, and correction, and one practices implementing the models in the course of living: engaging in relationships (as a friend, neighbor, lover, spouse, parent, child, or citizen) as well as attending social events (family dinners, religious services, public lectures, weddings, funerals, etc.). Such continual practice and reinforcement instills these general models as implicit knowledge about a huge variety of contexts.
Although one may be completely unreflective about and consciously unaware of the models, they inform our thoughts, interpretations, evaluations, emotions, moods, and actions. The well-bred man strives to order his external behaviors and internal emotions and attitudes in accordance with ritual as a marker or guideline for moral behavior. Xunzi explains that “when ritual is at its most perfect, the requirements of inner dispositions and proper form are both completely fulfilled” (Hutton Ch. 19:122).24 By informing our practice, rituals also inform our evaluative standards: the well-bred man not only guides his actions according to the models, but also implicitly approves of the values within them as appropriate. However, Xunzi does not naively expect perfection; one may perform rituals well or badly. Insofar as some individuals perform rituals well, they exemplify the model: they are “courageous” because they behave and respond according to the typified standards, while others do not warrant this description or commendation. At this first stage of moral development, ritual instills comprehensive models for action, which enables identification and evaluative judgments of behavior according to how well one exemplifies the model.25
Xunzi’s view on the role of ritual has several implications for the situationist empirical challenge. Doris and Harman argue that the substantial impact of seemingly insignificant situational features on behavior demonstrates that the behavioral regularity we observe is due to the stability and consistency of situations rather than global character traits like traditional Western virtues of courage or honesty (Doris 2002, p. 26–29; Harman 1999, p. 317). However, Xunzi’s emphasis on ritual makes it clear that he is aware of the fact that situational features affect behavior and that the significance of those features is often not apparent. As detailed above, rituals provide the filter through which we experience the world. They pervade every aspect of our lives and shape our emotional responses, moods, expectations, interpretations, evaluations, and actions through the minutiae of clothing, bearing, countenance, placement, sounds, etc.26 Rituals establish social order and are effective for moral education precisely because they capitalize on the influence of situational minutiae.27 As Hutton notes, the power of ritual “is attributed to various sounds and clothing, and this implies that people are susceptible to being seriously influenced by certain situational factors that both we and many ancient Chinese…would not normally think of as especially relevant to explaining their behavior” (2006, p. 45). And much like the ancient Chinese, few of us are aware of how situations elicit or prohibit our emotional responses, or why some situational features are made salient and others are not. Regardless of our conscious awareness of the influence of situational features or our judgments about the significance or insignificance of such features, the power of ritual stems from its comprehensive reliance on such minutiae. Situationist research seems full of promise because it appears to reveal the influence of insignificant situational features. But given an account like Xunzi’s, one should predict that seemingly insignificant situational features have powerful behavioral effects. To Xunzi, it is obvious that such situational features have powerful behavioral effects, and he capitalizes on the power of ritual as an ineliminable part of moral education. The situationists provide no challenge here; rather, they elucidate a Xunzian account.
Doris and Harman also charge that irregular behavior, as demonstrated in situationist research, provides reason to doubt or reject traditional Western virtues such as courage or honesty as global character traits. Xunzi would be nonplussed. At this first stage of moral development, it is certainly true that rituals provide comprehensive models for behavior and thought, and that ritual practice instills these as psychological models. But while Xunzi would allow that such psychological models effect regular behavior, he does not posit them as robust, global character traits. For Doris and Harman, global character traits are stable and consistent psychological entities causally operative across situation types. In contrast, for Xunzi, the psychological models for behavior and thought are type-level models: patterns or standards for appropriate action and thought indexed to particular situation types, and hence, stable but not consistent. For example, the psychological models for appropriateness in funerals are stable insofar as they yield behavioral regularities at funerals, but they are not operative in other contexts; while the standards for appropriate behavior within funerals are applicable to all funerals, they have no role for lectures, family dinners, participation in psychological experiments, etc. Clearly these type-level psychological models are stable in that they yield behavioral regularity within situation types, but Xunzi would reject any attempt to cast them as virtues or global character traits; indeed, to assume that these psychological models must be causally operative across situation types is nonsensical—it violates their nature as models.28 However, this is not to say that he rejects global character traits simpliciter, but merely that he does not posit these first stage psychological models as virtues or global character traits. Doris and Harman assume more than is required for behavioral regularity, which leads them to mistakenly suppose that the “irregularity” of behavior across situation types has evidential warrant against the efficacy of psychological models for behavioral regularity.
Consequently, we should be careful not to “overattribute” when making mental state attributions. For Xunzi, traditional Western virtues such as “courage” or “honesty” are descriptive and normative labels for actions that exemplify stable type-level psychological models for behavior and thought, but no further inference to global character traits is warranted at this stage of moral development. When my friend returns extra money to a cashier in Starbucks, I can rightly attribute “honesty” to her: she exemplified29 the psychological models that clearly inform her practice in such types of situations, and hence earned this descriptive and normative label. What I cannot do is to attribute “honesty” to her as a virtue or global character trait: despite our loose practice and sloppy language in making mental state attributions, our ontological postulates need to be modest.30
Finally, Xunzi rejects Doris and Harman’s proposal of situational regularity as a competing explanatory hypothesis: his emphasis on ritual makes it clear that the regularity of a situation as an instance of a situational type is not separable from the regularity of individuals’ action and thought as an instance of type-level psychological models. His notion that rituals establish and maintain social order is an important insight: because rituals function as a common cause for the regularity of situations and the regularity of individuals’ action and thought, each reinforces the regularity of the other.31
Recall the earlier example of attending a lecture. When attending a lecture, the lecture event unfolds as it does because we follow a particular pattern or model specified for a given type of circumstance. In this case, we adhere to the standards for “lecture events”: rather than boisterously playing touch football in the stadium, we expect the event to begin and so take our seats; rather than disrupting the speaker continually with questions, we evaluate that there will be enough time left for a question and answer session; rather than a cacophony of questions, we value the opportunity for individual questions in turn. But the lecture event also unfolds as it does because the ritual for “lecture events” structures the situation itself: there is a specific starting and ending time, components of the lecture (e.g., the introduction, the speech, the Q and A), a sequence of those components, etc. Individuals’ actions and thoughts such as emotional responses, moods, expectations, interpretations, evaluations, etc.—insofar as they adhere to the model for lecture events—are consistent with, allow the unfolding of, and reinforce the structure of the situation. Correspondingly, the structure of the situation itself as an instance of a situational type, insofar as it replicates a psychological model for lecture events, reinforces the appropriateness of that model. Consequently, there is no such thing as the regularity of situations apart from the regularity of individuals’ thought and action, insofar as the regularity of a given situation is partially constituted by the thoughts and actions of individuals and the collective thoughts and actions of others.32 Situational regularity cannot explain behavioral regularity sans psychological type-level models for proper thought and action; situationism is not a separate empirical hypothesis.33
5 Moral Development: Reflective Deliberation
In the first stage of moral development, one strives to order his behavior and thought in accordance with ritual, implicitly accepting the rituals as good and proper models by which to act and live. In the second stage of moral development, one desires to enact rituals as models of moral action and engages in a rigorous program of moral education following the guidance of a teacher. Xunzi notes that “those who are transformed by teachers and proper models, who accumulate culture and learning, and who make ritual and yi [“righteousness”] their path, become gentlemen” (Hutton Ch. 23:35, my emphasis).34 These are virtuous persons that consciously strive to embody the models, and to perfect every thought and action.
As Xunzi notes, rituals contain no explanation of or rules for their proper application, so this second stage of moral development involves not only learning to perform rituals correctly when and where they are appropriate, but also understanding when and how they should be altered. Teachers develop and perfect this nuanced ritual practice into moral action: “ritual is that by which to correct your person. The teacher is that by which to correct your practice of ritual” (Hutton Ch. 2:170).37
in learning, nothing is more expedient than to draw near to the right person. Rituals and music provide proper models but give no precepts. The Odes and History contain ancient stories but no explanation of their present application….However, if you imitate the right person in his practice of the precepts of the gentleman, then you will come to honor those things for their comprehensiveness. (Hutton Ch. 1:180, my emphasis)36
Such a single-minded devotion to the study and performance of ritual goes far beyond the mere proper performance of ritual in the first stage of moral development.
The gentleman knows that whatever is imperfect and unrefined does not deserve praise. And so he repeatedly recites his learning in order to master it, ponders it over in order to comprehend it, makes his person so as to dwell in it, and eliminates things harmful to it in order to nourish it. He makes his eyes not want to see what is not right, makes his ears not want to hear what is not right, makes his mouth not want to speak what is not right, and makes his heart not want to deliberate over what is not right. He comes to the point where he loves it…and his heart considers it more profitable than possessing the whole world. (Hutton Ch. 1:245)39
The rituals provide type-level models of proper action and thought, but as general type-level models, they are not always applicable to particular token situations. As Xunzi notes, simply following the ritual (like those in stage one do) may cause harm: “if one follows a model but does not debate over it, then cases where the model does not reach will surely be botched. If one holds a position but is not open-minded about things, then cases which one’s assigned position does not cover are sure to be let slip (Hutton Ch. 9:50).”40 In order to avoid harm or misapplication, it may be necessary sometimes to depart from the letter of the ritual in order to follow its spirit.41
To grasp the “spirit” or what is central to rituals is to understand their underlying “logic”: how they function, why they are effective, and their purpose. For example, dispensing with ornamentation in the funeral rite (e.g., the medieval European use of flowers to cover the smell or modern day Western practices of embalming, applying makeup, and dressing the deceased in “Sunday best” clothing), would result in revulsion and thereby diminish respect. Consequently, the emotions evoked and one’s attitude toward the deceased are what is essential to the rite, and the specific cultural and historical differences (e.g., the fabric used in the covering, the flowers, a Sunday-best suit, etc.) matter only insofar as they contribute to that purpose. Xunzi explains that if one’s consideration of ritual is myopic on the “letter of the law,” one will fail to grasp the spirit of the ritual: “if he has the proper model but does not fix his intentions on its true meaning, then he will act too rigidly. If he relies on the proper model and also deeply understands its categories, only then will he act with comfortable mastery of it” (Hutton Ch. 2:160).42
This continual focus on the underlying purpose of ritual develops a particular type of reflective deliberation about the essence and proper application of ritual. Xunzi notes that “to be able to reflect and ponder what is central to ritual is called being able to deliberate. To be able not to diverge from what is central to ritual is called being able to be firm” (Hutton Ch.19:170).43 By following the nuanced practice of the teacher as a guide, gentlemen examine rituals as models to understand when or if they are applicable, analyze how they should be altered, and reflect on how to engage in a sophisticated moral performance that simultaneously captures the underlying purpose or essence of the ritual while attending to the variation between situations.
As gentlemen proceed in their course of study and practice reflective deliberation, they develop the attitudes of benevolence (ren) and righteousness (yi). Ren, understood as benevolence, or caring for others, is cultivated by the repeated consideration of the application of ritual and its effects. Caring for others involves being concerned about potential harms to another and ensuring that one considers and treats others well. Xunzi explains this concern for others and being attentive to their needs as public-spiritedness and harmoniousness in all things (Hutton Ch. 9:60). Yi, understood as righteousness, is cultivated by the repeated consideration of the underlying essence or purpose of ritual and the desire to engage in the proper moral performance of ritual in each situation. Righteousness involves the perpetual commitment to the essence of ritual (the Dao) and attention to the details of situations for the proper application of ritual. Ren and yi enable moral action over the mere proper performance of the well-bred man through the sensitive application of ritual to the significant and complex details of the situation, which, as Hutton notes, “helps to bring about a kind of heightened sensitivity to certain features of circumstances” (2001, p. 204). At this stage of moral development, ren and yi are not yet full-fledged virtues, but they enable one “to follow a model and debate over it, and hold a position while being open-minded about it, so that there is…no goodness left undone, and out of a hundred affairs no errors are made—none but the gentleman is capable of this” (Hutton Ch. 9:50).44 The attitudes of ren and yi encourage one to consider and care about the details of each situation so as to cause no harm and leave no good left undone. As Cua notes, “successful moral training entails perceptiveness or appreciation of concrete situations” (1992, p. 50).45
Understanding the moral development of the gentleman yields a further interesting response to the empirical challenge. As type-level models, rituals may not be applicable to all token situations of a given type. The well-bred man, at his stage of moral development, only knows to follow the ritual. The gentleman’s task in learning is to reflect and deliberate on the essence of ritual to understand its appropriate application for a given situation. Accordingly, one should expect the gentleman to depart from the explicit ritual in cases in which he judges that the model is not applicable or needs alteration. Because gentlemen follow the explicit ritual in cases in which it is appropriate and depart from it to uphold its essence in others, they act morally in every case. Their behavior is highly regular when one evaluates moral performances (not the mere proper performance of the well-bred man). But if one evaluated their actions from the standpoint of performing explicit rituals, their actions would look more irregular than those who are well-bred. But of course, there is no reason to evaluate the actions of gentlemen using standards slated for those at lower levels of moral development.46 Such an evaluation would be a gross mistake. Contrary to Doris and Harman’s challenge, failure to behave “consistently” or with regularity is not an objection, but is an expected outcome for individuals at middle stages of moral development. Note further that such irregular behavior is not an indication of confusion or a moral failing, but of moral success: the gentleman has developed a cognitive understanding of the logic of rituals which will enable their appropriate alteration and selective application.47
6 Moral Development: Deliberative Situational Embeddedness
Although the gentleman is a virtuous person who performs moral actions, he can still be perfected. Xunzi notes that “the gentleman is the beginning of ritual and yi. Practicing them, habituating oneself in them, accumulating great regard for them, making oneself fond of them—these are the beginning of becoming a gentleman” (Hutton Ch. 9:305).48 By continually engaging in reflective deliberation and developing ren and yi, over time a gentleman may become a sage. Whereas a gentleman’s moral actions require labor and conscious effort in his course of learning,49 a sage acts with ease. Xunzi explains that “one who is truly sublime is a perfected person. For the perfected person, what forcing oneself, what steeling oneself, what precariousness is there?” (Hutton Ch. 21:310).50 A sage not only understands the Way (the Dao), but also loves it; performing moral action does not require wrenching effort, but flows from his developed character.51
The deliberation of the sage, guided by the Dao and sensitive to the situation, is not fixated but balanced and fair, focused on and embedded in the determinative features of the situation or token event.56 This kind of deliberation, which I call Deliberative Situational Embeddedness, is highly nuanced and appropriately flexible. Xunzi explains:
sees the disaster of being fixated and blocked up on one’s thinking. So, he is neither for desires, nor for dislikes, is neither for the origins, nor for the end results, is neither for what is near, nor for what is far away, is neither for what is broad, nor for what is shallow, is neither for the ancient past, nor is for the present. He lays out all the myriad things and in their midst suspends his scales. For this reason, the various different things are unable to become fixating and so disorder his categories of judgment. (Hutton Ch. 21:145)55
Because the judgments of the sage co-vary with the fine details of each situation, they are reliably accurate.58
the character of a true king [sage] is that he ornaments his every move with ritual and yi. He hears and decides cases in accordance with their proper kinds. He holds up for clear inspection the fine points of things. His policies adapt to changes endlessly. This is called having a proper source of action. Such is the character of a true king. (Hutton Ch. 9:225)57
Xunzi’s account of the sages also has some interesting implications for the empirical challenge: although ren and yi are undoubtedly virtues, they are virtues that enable situation-specific behaviors. They support and are operative within a process of deliberation about each situation, and so are clearly consistent across a variety of trait inducing situations. But it is not clear that ren and yi would count as stable traits—at least, in the sense that Doris intends—because they do not issue in regular behavior within a situation type. Indeed, if one examines behavior within situation types, the behavior of the sages may appear to be highly irregular because it is situation specific. But at this stage of moral development, Xunzi would counter that to seek regular behavior within a situation type is to use the wrong observational class. The mistake that Doris and Harman make is to assume that robust character traits that are themselves stable and consistent must yield stable and consistent behaviors. On a dispositionalist or an intellectualist account, it is a mistake to assume that the postulated psychological entity and consequent behavior must have the same properties.59 Because of this mistaken assumption, Doris and Harman assume that regular behavior within situation types provides observational evidence for stable psychological traits. Using this observational class, the flexible and situation-specific behavior of the sages looks irregular, and so ren and yi would appear to lack stability.
Going beyond this general point, the flexible behavior of the sages will have a high degree of regularity partly because it is specific to situations that are themselves highly similar. Although each situation is unique, situations share features and have family resemblances. As Hutton notes, the sage distinguishes “differences in situations that one would otherwise not consider important or pay attention to” (2001, p. 259). Insofar as ren and yi are stable and consistent as part of a process of deliberation, sages will be attuned to the specificity of situations, and their behaviors will co-vary in reliable and predictable ways accordingly. Because Doris and Harman conceive of stability as regular behavior within situational types (refer back to their initial argument against globalism), they take irregular behavior within situational types to be evidence against the stability of virtue. However, flexibility within situational types results from the stability of virtues such as ren and yi in a process of deliberation: stability is here redefined as reliable co-variance or flexibility. As a consequence, the type of evidence used to make inferences about the stability of virtues alters from a consideration of regular behavior to flexible behavior within a situational type. In looking for stability as regular behavior within situational types, Doris and Harman are, quite simply, looking for the wrong evidence to tell against robust virtues such as ren and yi.61
openness typically is not unlimited: e.g. to characterize someone as open to a range of possibilities is not to say that there is a real chance that within the next year she or he will torture a small child. Even in the case of someone who very noticeably is not always “the same”, there can be some imaginable forms of behaviour that we can be highly confident in not expecting. (2001, p. 249)
But Doris may charge that the consideration of virtues as constant within a process of deliberation and the consequent redefinition of stability is a retreat to an intellectualist interpretation of virtue. He argues that “an attractive account of virtue should be concerned not only with cognitive and affective patterns, but also with patterns of overt behavior; the ethical quality of a life is determined by actions as well as psychological states” (Doris 1998, p. 510). However, every virtue theorist holds that virtue should be concerned with patterns of overt behavior. Indeed, this is why the gentleman engages in a deliberative process to reflect on ritual and its appropriate application, and the sage perfects deliberative situational embeddedness. The issue here is not whether robust character traits are dispositional or intellectual, but whether robust character traits must yield behavioral regularity. To assume that virtues must yield regular rather than flexible or reliably co-variant behavior begs the question about which behavioral properties license the attribution of stable and consistent traits. Further, the charge that such behavior is irregular rather than appropriately situation specific itself assumes a normative standard.62 While it is true that the ethical quality of a life is determined by actions as well as psychological states, it appears that the means by which Doris wants to evaluate the ethical quality of life is decidedly bourgeois: behavioral regularity (stage one development). Xunzi’s account is illuminating because it adopts a different set of normative assumptions. For Xunzi, flexible behavior in response to situational variation is the apex of moral development and the goal of moral education, and is not telling evidence against—but convincing evidence for—his account of moral development.
7 The Dissolution of the Challenges
Xunzi’s embeddedness model of moral development not only serves as a critical lens to analyze problematic assumptions, but also dissolves the situationist empirical and theoretical challenges. The empirical challenge targets the causal efficacy and existence of virtues as robust, or global, character traits by offering situationism as a competing explanatory hypothesis. Xunzi’s account of moral development is particularly interesting because it demonstrates that situationism is not a separate empirical hypothesis; situational features collude with a variety of psychological entities to yield behavior throughout his spectrum of stages.
At the first stage of moral development, ritual simultaneously structures situations and instills type-level models for behavior and thought which are stable, but not consistent. As noted previously, Xunzi does not posit these psychological type-level models as virtues, but nonetheless, they are causally efficacious and inform one’s emotional responses, moods, interpretations, expectations, and actions. Insofar as the spirit of the situationist challenge questions the causal efficacy of psychological traits, one might think that situationism offers a competing explanatory account to Xunzi’s stable type-level psychological models.
However, these are only explanatory competitors if one assumes a false dichotomy of causal relations. The insight that Xunzi offers is that behavioral regularity63 is explained by both the regularity within situations and the regularity afforded by stable psychological type-level models.64 Behavioral regularity is not merely overdetermined; rather, situational regularity and type-level psychological models are interactive and mutually reinforcing. Situationism serves as a competing hypothesis if one assumes that type-level psychological models have no developmental or social role of enforcement in structuring situations. This assumption is its own reductio. As Xunzi makes clear, neither situational regularity nor type-level psychological models are explanatory by themselves, and adequate predictions and explanations of behavior need to attend to this interactive and joint cause.
Although situationism does not fare well in the above comparison, perhaps it can stand up to Xunzi’s account of ren and yi as robust virtues. Because virtues, as robust character traits, are thought to yield regular behavior within and across situation types, situationists take behavioral irregularity to provide empirical support for the power of situational influence against robust character traits. A multiplicity of research in social psychology shows that individual, specific situational components or factors affect one’s behavior. For example, in research on helping behavior and mood effects, subjects do not uniformly help in situations where others need assistance. The difference or variation in behavior between those who do and those who do not help is some individual situational factor such as previously finding a dime in a phone booth.65 Despite whatever claims virtue ethicists make about the robustness of virtue or character traits, situationists point to these individual situational factors as the real driving force behind our behavior: positive moods increase the probability of helping behaviors, and positive moods are induced by small factors like finding a dime. So setting aside normative judgments about “irregularity” for the moment, the phenomenon to be explained here is situation specific behavior. And one might think that, here at last, situationism offers a viable competing explanatory hypothesis.
Unfortunately, situationism finds itself caught between two horns of a dilemma: either it is not a separate explanatory hypothesis from a Xunzian account and hence cannot offer different empirical predictions, or it is forced into making some unpalatable assumptions in order to offer some different empirical predictions. Regarding the first horn, situationism is not a separate explanatory hypothesis from a Xunzian account: each account predicts the impact of significant situational features, the impact of seemingly insignificant features, the variation of behavior within a situational type, etc. Although Doris and Harman make much of the power of seemingly insignificant situational features, Xunzi was more than aware of the fact that individual situational features have powerful mood effects: this is one reason for the power of ritual and the point of harnessing it for use within a program of moral education.66 Further, as should be clear from previous sections, Xunzi’s account goes far beyond consideration of simple mood effects to the influence that individual situational factors have on eliciting various emotional responses (e.g., disgust induced from the smell of a corpse), influencing and altering the salience of particular situational features (e.g., the family’s grief over one’s own), etc. Clearly Xunzi did not have god-like knowledge about the causal efficacy of every factor in every situation, so social psychologists have an important role to play in future research to uncover details about the causal efficacy of factors we previously thought were insignificant. But uncovering and cataloging such details (few, if any of which would surprise Xunzi) within social psychology is a far cry from Doris and Harman’s attempt to offer a different explanatory hypothesis about the causes of our behavior with different empirical predictions.
Given the comprehensiveness and complexity of a Xunzian account, in order to offer different empirical predictions about our behavior and its causes, one would have to make some assumptions about global character traits that are unpalatable and unwarranted; namely, that virtues or global character traits rigidly determine behavior irrespective of situational influence or causal efficacy.67 Kamtekar agrees that this is a peculiar conception of global character traits: “something that will, if present, manifest itself in characteristic behavior, and will do so no matter what else there is in the situation for the person to respond to” (2004, p. 474). Xunzi offers a complementary insight here: situationism’s quest for a competing explanatory hypothesis would also require making some equally unpalatable and unwarranted assumptions about situations; namely, that situations are perpetually “brute givens” that rigidly determine behavior in isolation from psychological influence (salience, interpretations, etc.). If one assumes that situational features and global character traits are causally isolated, then one can offer situationism as a competing explanatory hypothesis and predict situation specific behavior.
However, this competition is illusory. The insight Xunzi offers is that situational features and robust character traits are interactive rather than perpetually opposed forces of behavior; he rejects the assumptions that (1) robust character traits determine behavior irrespective of minute situational features as complementary causal factors and that (2) situations are brute givens. Global character traits such as ren and yi are fully compatible with, and indeed, enable nuanced attention to and deliberation about situational features which yield behaviors that are highly flexible and specific to the token situation. Further, the causal effect of individual situational factors and their insignificance or lack of salience can themselves be altered over time. At later stages of moral development, situations themselves are shaped and interpreted via the virtues of ren and yi; ren and yi guide what one attends to and what one deliberates about within a situation.68 They are virtues cultivated over time that cultivate salience within a situation, and hence, are partly responsible for the causal effect of various situational features via its being interpreted as the kind of situation it is, needing to be responded to in the way that it does.69 Even though the nature of the collusion between situational features or factors and psychological entities changes over the course of moral development, the point remains the same: the situationist empirical challenge is dissolved completely.
Of course, the theoretical challenge collapses along with the empirical. Doris claims that situationism provides a better normative ground for ethical behavior than virtue ethics. He argues that “situationism suggests a certain redirection….rather than striving to develop characters that will determine our behavior in ways significantly independent of circumstance, we should invest more of our energies in attending to the features of our environment that impact behavioral outcomes” (1998, p. 515). Doris worries that focusing on robust character traits, or traits that are “substantially resistant to contrary situational pressures” may not only yield inflexible behavior, but also yield more unethical behavior through the overestimation of the strength of one’s character (1998, p. 506). Rather than relying on the strength of our characters, we should attend to situational features. He argues that “our duties may be surprisingly complex, involving…a sort of ‘cognitive responsibility’ to attend, in our deliberations, to the determinative features of situations” (1998, p. 517). He holds that situationism provides a better theoretical grounding for ethical norms because it does not promote the development of global character traits and it focuses attention on morally relevant features of situations.
While Xunzi can appreciate this new found responsibility, situationism does not offer a substantive theoretical alternative for ethical norms. By now, it should be clear that Xunzi would be mystified by the claim that virtue ethics strives to develop characters that “determine our behavior in ways significantly independent of circumstance”: such a claim borders on the absurd. And he would agree that we have a “cognitive responsibility” to attend to and deliberate about the features of situations. Indeed, this is the normative ground for his ethic and the apex of moral development.
Further, Xunzi’s virtue ethic provides a richer account of moral development, moral education, and moral psychology. The three stages of moral development offer normative criteria for development, descriptive accounts of psychological entities and their collusion with situational features, and a descriptive and normative account of flexible behavior in response to situational variation. Doris and Harman simply suggest learning about one’s situational influences in order to avoid them; a paltry proposal for moral education and development.70 Xunzi’s program of moral education, spanning years and yielding several stages of development by inculcating a multiplicity of enforcements, seeks to develop deliberative situational embeddedness. And while Doris and Harman treat situational influence as a “one size fits all” specter, Xunzi’s account of moral development and program of moral education suggest that there are different levels and kinds of situational influence that are appropriate; indeed, the degree of influence for situational factors, one’s cognitive awareness of them, and the role they play in yielding one’s behavior varies greatly among each of the stages. Consequently, the standard(s) of appropriate influence of situational factors must be indexed to a specific level of moral development. All told, Xunzi’s virtue ethic provides a very rich theoretical basis for empirical differences in situational influence that may well tailor our approach to moral education.71 Situationism does not provide a unique ground for ethical norms, and does not offer a substantive theoretical alternative for moral education or development. There simply is no comparison; the theoretical challenge withers.
8 Moving Forward
I had three projects in this paper: first, to present and argue for the “embeddedness model” of moral development; second, to use Xunzi’s account as a lens to critique key ontological and normative assumptions about moral psychology and moral development within the arguments posed by situationists; third, to show how Xunzi’s virtue ethic uniquely responds to the situationist empirical and theoretical challenges and offers a new approach to conceive of virtue ethics. Although others have focused on the conception of virtue assumed by the situationists, my focus has been on moral development. Taking the notion of moral development seriously requires an examination of the ontological claims about virtue, robust character traits, and psychological models at various levels of development, which affects observational classes, evidential status, and inferential warrant. Notice that for Xunzi, we can be realists about virtue only at the third stage of moral development, and only for the virtues of ren and yi. We can be instrumentalists about virtues at earlier levels of moral development, but it would be a gross mistake to think that a normative and descriptive account of moral development postulates each descriptive or evaluative label or each step within the development as an ontological entity or trait of an individual.72 Taking the notion of moral development seriously also requires an examination of the normative claims about the nature of development and the goal of moral education. If Xunzi is right that the purpose of moral education is to develop stable and consistent character traits that lead to increased response to and embedded deliberation in situational features, then one may agree with Kristjansson that “because of the inherent particularity of morality, the fact that people’s moral decisions always turn out to be situation specific may not be such a bad thing after all” (2008, p. 65). On Xunzi’s virtue ethic, this is exactly the right answer.
What is so unique about Xunzi’s account is that it undercuts the threats posed by situationism: situationism cannot offer a different explanatory hypothesis with different empirical predictions without making problematic ontological and normative assumptions about the causal role of situational features and psychological entities at various stages of moral development. Whether such arguments aid other accounts of virtue ethics depends on a comparison with Confucian accounts that is beyond the scope of this paper.73 However, my arguments cannot aid any account of virtue ethics that relies on the aforementioned problematic assumptions. What is so promising about Xunzi’s virtue ethic is that it provides a normative ground for attending to and deliberating about situational features without making the problematic ontological and normative assumptions that plague situationism (and indeed which, situationists charge, plague other accounts of virtue ethics).
By reconceiving the normative ground of ethics and the nature of moral development, Xunzi presents a way to move forward; he offers a conception of virtue ethics that dissolves the debate between Western virtue ethicists and situationists and suggests a new purpose and focus for situationist research. Although modern Western philosophers and psychologists are often unaware of the significance of situational features, situationist research can correct our mistaken assumptions about significance. Situationist research has the potential to answer—what I take to be—more interesting questions about why and how situational features have such powerful effects, essentially detailing why a program of moral education such as Xunzi’s could be so effective.74 Perhaps we can put situationist research in the service of helping us all on our path to sagehood.
I would like to thank audience members for a lively discussion at the 2009 American Philosophical Association Pacific Division meeting as well as Steve Angle, Gabriel Palmer-Fernandez, and Bruce Waller for their insightful comments on earlier drafts of this paper. I would also like to thank two anonymous reviewers for their suggestions which greatly improved the paper. Research for this paper was generously supported through a National Endowment for the Humanities Seminar in 2008.
For example, see Darley and Batson’s (1973) study, in which situational features such as “degree of hurry” influenced the helping behavior of subjects, or Isen and Levin’s (1972) infamous study in which the contingent consequence of finding a dime created positive mood effects and increased the rate of helping behaviors.
Ivanhoe (2000) argues that Xunzi holds a “re-formation” account of self cultivation or moral development because of his view of human nature and the necessity of moral training to overcome one’s desires. Stalnaker (2006) holds a similar account. I think they overstate the case, but this disagreement over theoretical interpretations is far beyond the scope of this present essay.
As a consequence, Aristotelians can avail themselves of these arguments as long as they avoid these problematic assumptions.
Although Harman and Doris assume that these are competing hypotheses, there are reasons to doubt this claim, which I address later in the paper.
Doris and Harman part company here. Harman explicitly rejects the explanatory value of narrowly individuated or local character traits. He states that “character traits are broad based dispositions that help to explain what they are dispositions to do. Narrow dispositions do not count” (1999, p. 318). He holds that character traits must offer “common explanations” across situations (p. 318). Clearly, cross-situational consistency is an important explanatory value for Harman.
Others have challenged Doris’s distinction between local and global traits. For example, Slingerland rejects “the implicit assumption that we know what we are talking about when we contrast ‘local’ with ‘global’ traits: that is, that there is a clear, principled distinction between the two. This is not at all the case. When offered as an analytic dichotomy, the ‘local’ versus ‘global’ distinction is simply not tenable, because any truly ‘local’ trait would not be a ‘trait’ at all, but merely a single occurrence: John performed behavior X in situation Y at this particular time and place. The sorts of local traits that Doris thinks worthy of our attention count as ‘traits’ because they are already abstract to various degrees” (2011, 399).
Although Doris’s 1998 and 2002 texts make more circumspect claims, in an article with Stich he appears to agree with Harman’s stronger conclusions: “the Aristotelian conception of traits as robust dispositions…is radically empirically undersupported…[so that] programmes of moral education aimed at inculcating virtues may very well be futile” (Doris and Stitch, 2005, p. 119–120, my emphasis).
Xunzi’s birth and death dates are unknown, and these dates are mere guidelines for comparison. For example, it is known that Xunzi held an administrative position in the state of Chu until his ruler was assassinated in 238 BCE, and that he then retired. However, scholars have no actual date for his death.
For more detailed information on Xunzi and the Classic period of Confucianism, please see Ivanhoe’s (2000) introduction and chapter on Xunzi, as well as the introductions in Knoblock (1999) and Watson (2003). For an interesting interpretation of Xunzi and a comparison with Augustine, see Stalnaker (2006).
Xunzi argues that ritual is “the utmost in patterning….Those under Heaven who follow it will have good order. Those who do not follow it will have chaos. Those who follow it will have safety. Those who do not follow it will be endangered. Those who follow it will be preserved. Those who do not follow it will perish” (Hutton Ch. 19:145–150). See also Ivanhoe and Van Norden (2005, p. 276); Knoblock (1999, p. 611); Watson (2003, p. 98).
Ritual also serves to constrain or limit the expression of emotion. Xunzi explains that the standard practice of limiting mourning to a 3-year period is to set “a limit for the utmost hurt….After the 25 months of the 3-year mourning period, the worry and hurt are not yet done, and the feelings of longing and remembrance are not yet forgotten. Nevertheless, ritual breaks off the mourning at this time….in order that there may be a proper stopping point for sending off the dead and proper regulation for resuming one’s normal life” (Hutton Ch. 19:465–475). See also Knoblock (1999, p. 637); (Watson, p. 109).
Tu notes that “in the Confucian context it is inconceivable that one can become truly human without going through the process of ‘ritualization,’ which in this particular connection means humanization” (1972, p. 198).
Lai (2006) argues that there are three stages of self-cultivation within the Analects. She argues that “the first is the novice’s stage during which li [rituals] are essential in inculcating correct forms of behavior. At this stage, adherence to the dictates of li introduces the learner to the appropriate proprieties in different contexts. The second stage is an experimental one during which the learner extracts principles from these behavioral forms through constant practice. The emphasis at this stage is on the learner testing out his application of moral principles….The third phase is marked by the deliberations of the mature, cultivated person, who has a good grasp of the principles and ideals encoded and realized in meaningful social interaction. At this stage, li have a different significance…they do not function as instruments of rote learning but rather are channels for meaningful self-expression” (p. 69). Although I agree with Lai’s emphasis on the changing role of li (see Tu 1972 for a similar account), we disagree on the nature of Confucian moral development. Lai takes moral development to be the increasing understanding and application of moral principles found within rituals, but I argue that the very fact that rituals are type-level models means that they do not contain extractable principles applicable to token situations. Although it is an interesting question as to the extent of contrast between the Analects (authored by Confucius) and the Xunzi (authored by Xunzi), such a comparison beyond the scope of this essay.
Xunzi is clear that humans all have the same nature, but that differences in moral development depend on one’s willingness to engage in moral self-cultivation: “Someone says: sageliness is achieved through accumulation, but why is it that not all can accumulate thus? I say, they can do it, but they cannot be made to do it. Thus, the petty man can become a gentleman, but is not willing to become a gentleman” (Hutton Ch. 23:290). See also Ivanhoe and Van Norden (2005, p. 305); Knoblock (1999, p. 767); Watson (2003, p. 171).
Hutton argues persuasively that Confucian non-liberal political philosophy may explain the emphasis on character and global character traits. He notes that “one can see how acknowledging situationist concerns might actually drive one to emphasize the importance of robust, virtuous character even more, rather than less, because it may be that only if some people really do have robust character [i.e., the sages as exemplars and teachers] can society turn out well” (2006, p. 51). Hutton is right that Confucian political philosophy attends to situationist concerns, but the emphasis derives from the model of moral development assumed and not the political ideals. Although it is beyond the scope of this paper, I would counter that the embeddedness model of moral development I argue for here, as a psychological model, explains both the emphasis on ritual for moral education and the ideals within Confucian political philosophy.
Xunzi expounds on such descriptive and normative labels at length. For example: “to endorse what is right and condemn what is wrong is called ‘wisdom.’ To endorse what is wrong and condemn what is right is called ‘stupidity.’….To injure a good person is called ‘villany.’ To call the right as right and wrong as wrong is called ‘righteousness.’….To conceal one’s actions is called ‘deceptiveness’” (Hutton Ch. 2:55–65). See also Ivanhoe and Van Norden (2005, p. 262); Knoblock (1999, p. 29); Watson (2003, p. 27).
Xunzi notes that “sounds and music enter people deeply and transform people quickly” and has a lengthy discourse on the character of sound: how it affects individuals’ emotional states, moods, and behaviors, and how it may be used for political purposes to ensure social order. For example, he notes that “if music is balanced and peaceful, then the people will be harmonious and not degenerate. If music is solemn and majestic, then the people will be uniformly ordered and not cause chaos….If music is dissolute and dangerous, then the people will be degenerate, arrogant, vulgar, and base” (Hutton Ch. 20:60). See also Knoblock (p. 655); Watson (2003, p. 117).
Xunzi also notes the subtle effect others have on our attitudes and behaviors: “If you obtain a worthy friend to befriend, then what you see will be conduct that is loyal, trustworthy, respectful, and deferential. Then you will make daily progress toward ren and yi and you will not even realize it….Now if you live alongside people who are not good, then what you hear will be trickery, deception, dishonesty, and fraud. What you will see will be conduct that is dirty, arrogant, perverse, deviant, and greedy….A saying goes, ‘If you do not know your son, observe his friends….’ Everything depends on what you rub up against!” (Hutton Ch. 23:390). See also Knoblock (1999, p. 773); Watson (2003, p. 174). Similarly, Merritt (2000) notes the “sustaining social contribution” of situations to moral behavior, and Sarkissian (2010) also highlights the interconnectedness of social behavior through the influence of our behavior on others and vice versa.
One might think that this is simply a redescription of Doris’s notion of local character traits. Recall that local character traits such as “battlefield physical courage” are productive of behavior in very narrowly specified conditions, and simply are not causally operative during storms on the high seas, while being tortured, or upon stepping into a den of snakes. I argue (see below in the main text) that traditional Western virtues such as “honesty” are descriptive and normative behavioral labels, so I have no problem agreeing that we can parse them as finely as we want based on whatever behavioral patterns we identify. The level of descriptive grain we prefer is governed by our pragmatic interests in prediction and explanation. (See Jacobson 2005 for a related argument about parsing virtues.) However, my hunch is that Doris is after a much more fine grained and limited dispositional entity than Xunzi promotes, and I address the causal role of psychological entities and their conception extensively later. For an immediate difference, note that Doris is positing narrowly specified traits, while Xunzi endorses type-level comprehensive models as implicit, abstract, emotive, interpretive, and procedural knowledge structures (which may support but are not merely dispositions, nor dispositions of a simple sort). The only thing they seem to have in common is that they are each indexed to situational types, and hence, given Doris’s definition, stable. Using this simplistic comparison, one may still attempt to argue that Xunzi is merely endorsing something akin to Doris’s local character traits. But even granting this for the sake of the argument would not deflate Xunzi’s account: Doris endorses the ethical development of local character traits while Xunzi holds that type-level psychological models are “training tools” that are transcended in moral development.
These are behavioral labels insofar as one only has observational evidence of behavior. Unfortunately, we do not have access to the mental states of others; I do not know when someone feels sorrow internally unless I can observe some external or behavioral manifestation of it.
Indeed, one might wonder what purpose attributing a global character trait of “honesty” would serve in this case. Clearly I might describe her as “honest” because she routinely returns money, but this is an accurate description as well as a commendation of her behavior. Insofar as she exemplifies a type-level pattern or model that I can identify via a label, at the least, I can make fairly accurate predictions on the basis of her past behavior, and at most, I can be an instrumentalist about such type-level psychological models. But for Xunzi, attributing a psychological entity to her is less explanatory than one might assume: first, the type-level psychological models are only stable (hence, they would not yield viable predictions across situation types anyway), and second, as is detailed in the main text below, the observed regularity in her behavior is a consequence of the joint regularity between type-level psychological models and situations (hence, there is little temptation to attribute psychological entities—with a strong form of realism no less—if what we observe is a consequence of a causal nexus).
Merritt (2000) argues for the sustaining social contribution to character (SSC) against strong forms of the motivational self-sufficiency of character (MSC). Doris claims that “the notion of socially sustained virtue seems to be predicated on the realization that robust traits are deeply problematic” (2002, p. 91). However, as Merritt notes, the degree of social contribution versus internal sufficiency in virtue ethics is itself a normative issue. I argue further that it is deeply problematic to assume a sharp divide between situational and psychological causal relations, quite apart from the normative assumptions one makes. I address this point at length later in the paper.
Merritt (2000) argues that character can be sustained through social contribution and relations with others (e.g., the death of one’s drinking partner can minimize one’s disposition to drink), and Sarkissian (2010) argues that we influence situations through our effects on others (e.g., social signaling, tone, mannerisms, etc.). My point here is not merely pragmatic, but ontological.
Note that the regularity of situations (as complex event types) is not to be confused with the recurrent appearance of individual situational features, factors, or components across situational types. I address this point at length later in the paper.
Studying music is clearly an important part of Xunzi’s program of moral education; music complements ritual in eliciting moods and emotions (e.g., think of the contrasts between classical, techno, or elevator music), focusing attention via suggestive interpretive frameworks (e.g., hearing a funeral dirge while viewing caskets of fallen soldiers versus hearing patriotic music), and providing models for proper thought and action (e.g., thinking of the families of the fallen soldiers and their loss rather than ruminating on the nature of patriotism). However, music is secondary to ritual: “nothing is more direct than following ritual, nothing is more important than having a good teacher, and nothing works with greater spirit-like efficacy than to like it with single-minded devotion” (Hutton Ch. 2:85). See also Ivanhoe and Van Norden (2005, p. X); Knoblock (1999, p. 31); Watson, p. 28).
The Romanized word “yi” used here is a different Chinese word than “yi” of “appropriateness” or “righteousness” discussed in the context of ren and yi as developing attitudes and virtues.
As Jacobson notes, “skill development primarily involves learning how to apply these maxims to particular cases and, eventually, knowing when to ignore them” (2005, p. 309).
Although Doris and Harman urge attention to situational features, this largely appears to be gathering knowledge of features that instigate one’s moral failings and avoiding those particular things. Harman offers some specific advice: “if you are trying not to give into temptation to drink alcohol, to smoke, or to eat caloric food, the best advice is….Don’t go to places where people drink! Do not carry cigarettes or a lighter and avoid people who smoke! Stay out of the kitchen!” (2003, p. 91). Doris and Harman urge attention to particular situational features, but neither urge the development of situational perceptiveness akin to moral perception as urged by Cua or Xunzi.
Kamtekar agrees on this point: “the difference between the situationist concern with behavioral consistency and the virtue ethics concern with consistent success raises the question: relative to what standard of consistency ought people to be judged consistent or not? If one’s purpose is to evaluate folk psychology, then it is reasonable to take ordinary people’s expectations as the standard. But if one’s purpose is to evaluate virtue ethics, then the standard will have to be different and to take account of the fact that we are thinking, goal-oriented creatures” (2004, p. 484).
Miller makes a similar point from the perspective of Aristotelian accounts of virtue. He argues that “the results that Harman takes to be a reductio of character-based explanations, may actually turn out to be precisely what one should expect on a sufficiently nuanced understanding of virtue ethics” (2003, p. 370).
Achieving moral perfection requires great effort. Xunzi explains that “if the people on the streets were to submit themselves to study and practice learning, if they were to concentrate their hearts and make single-minded their intentions, if they were to ponder, query, and thoroughly investigate—then if they add to this days upon days and connect to this long period of time, if they accumulate goodness without stopping, then they will break through to spirit-like powers and understanding….Thus, becoming a sage is something that people achieve through accumulation” (Hutton Ch. 23:280). See also Ivanhoe and Van Norden (2005, p. 305); Knoblock (1999, p. 765); Watson (2003, p. 171).
Xunzi explains that “when one can deliberate and be firm, and adds to this fondness for it, then this is to be a sage” (Hutton Ch. 19:170). See also Ivanhoe and Van Norden (2005, p. 277); Knoblock (1999, p. 615); Watson (2003, p. 99).
Xunzi notes that “in every selection that people make, the object of their desire does not come pure. In what they reject, their object of dislike does not go away pure. Thus for every action people make, they must come prepared with a balance….The Way is the correct balance from ancient times to the present” (Hutton Ch. 22:350). See also Knoblock (1999, p. 735); Watson (2003, p. 157).
Bias and over-focus distort one’s deliberation and yield inaccurate moral judgments. Xunzi explains that “almost always, the problem for people is that they become fixated on one twist and are deluded about the greater order of things….In whatever respect the myriad things are different, they can become objects of fixation to the exclusion of each other” (Hutton Ch. 21:1, 30). See also Ivanhoe and Van Norden (2005, p. 286, 287); Knoblock (1999, p. 671); Watson (2003, p. 125).
Although it is beyond the scope of this paper, I take it that this is the insight behind Hursthouse (1991) discussion in reconceiving the abortion debate as a focus not on biological facts or moral concepts. She argues that there are more facts that need to be considered, and that once one does, this raises consideration of still more facts (see p. 229 in particular). I take this to be an example of deliberative situational embeddedness that attends to the minute and unique features of the situation that distinguish it from other situations and which matter—such that if these additional details were ignored, one would be in danger of making the wrong judgment. One might also see other positions in the debate as suffering from fixated thinking: an over-focus on biological facts or concepts such as personhood yields inaccurate moral judgments.
McDowell notes that a “kind person has a reliable sensitivity to a certain sort of requirement that situations impose on behavior. The deliverances of a reliable sensitivity are cases of knowledge; and there are idioms according to which the sensitivity itself can appropriately be described as knowledge: a kind person knows what it is like to be confronted with a requirement of kindness. The sensitivity is, we might say, a sort of perceptual capacity (1998, p. 51).
Indeed, Doris proposes local character traits as an alternative to global traits because he assumes there must be a property correspondence between behavior and psychological traits. However, there is no motivation or reason to posit local character traits or fragmented character in the absence of this assumption. Exposing the assumptions that gird the empirical challenge to global character traits destroys any support for local character traits as a theoretical alternative.
Merritt (2000) argues that the sustaining social contribution to character (SSC) and the motivational self-sufficiency of character (MSC) are competing normative conceptions of virtue. Note that Xunzi emphasizes different norms at different stages of moral development: he endorses the sustaining social contribution to character (in the sense that Merritt presents it) strongly for those at the first stage of moral development, with a gradual transition for those at the second stage, to the strong endorsement of the motivational self-sufficiency of character for those at the third stage of moral development. Merritt suggests that our interests in stability might favor SSC as a normative conception of virtue. However, our preferences may depend more on which virtues one considers: ren and yi would offer what Merritt seems most interested in (and indeed why she favors a Humean account), yet they are a conception of virtue which is strongly motivationally self-sufficient.
The point here is not about how to characterize behaviors (e.g., what counts as “honesty” within a situational type), but is rather about how to characterize the concept of stability as a property attributed to a virtue and, hence, the nature of the evidence used to make inferences about the robustness of a virtue.
See Weilenberg (2006) for some interesting differences in Socratic, Kantian, and Aristotelian assumptions about consistency.
Understood as whatever particular actions we choose to identify via descriptive and normative labels such as “honesty, “civility,” “respectfulness,” “generosity,” etc.
There is an important distinction to keep in mind here between a situation as a complex event, and individual factors or components that may be salient, causally effective, etc. For example, the fragrant smell of baking bread (see Doris, pp. 30–31) or the stirring rhythms of patriotic music are highly causally effective factors to alter moods. A factor may be present and reliably causally effective across many situational types, but its presence does not make them all instances of the same “situation.” My point here is not about whether, or which, individual factors are reliably causally effective, but rather, about the regularity within situations as types: the regularity of a collection of essential factors or components, their combination or structure, the order or sequence of the components, etc. in virtue of which the situation is an instance of a specified type. For example, refer back to the case of attending a lecture.
See the infamous “dime study” done by Isen and Levin (1972). As noted in the introduction, virtue ethicists have taken aim at such studies, challenging the design, results, assumptions, etc. For the sake of the discussion here, I will assume the correctness of such research.
Such features can be harnessed within a political system as well. As noted previously, Xunzi details the effect of music and sound on individuals, and exhorts political leaders to attend to such features.
See Merritt (2000) for an argument that Aristotelian accounts of virtue may rely on this problematic assumption. While there are Aristotelian scholars who do seem to make this assumption, I doubt it is inherent to an Aristotelian account of virtue ethics. I leave this defense and debate to Aristotelian scholars.
Given Xunzi’s account of moral development, it may be the case that the sages will be less susceptible to the distorting effects of particular situational factors and their corresponding induction of emotional responses, moods, etc. While Xunzi’s program of moral education is designed to forestall some effects and to cultivate others through the power of ritual, there is no reason to expect sages to be completely immune from the effects of all situational factors. First, one should be careful not to read a dichotomy between situational and psychological factors back into Xunzi. Second, Confucian moral development is a life-long process, even for sages.
In an excellent article, Slingerland (2011) offers an account of Confucian moral education and how such a program of explicit and lifelong environmental manipulation and character training inculcates stable and consistent virtues that can meet the situationist “high-bar” argument. He correctly focuses on the role of perpetual (life-long) situational support through “immersion in carefully designed cultural forms” (p. 414) (i.e., ritual), but then mischaracterizes (at times) the nature of the developing virtues for the gentleman and the sage (p. 413). The consequence of Confucian moral education is not independence from situational forces simpliciter, but an alteration in the effect of specific and particular situational factors. Despite a few small disagreements, our positions are complementary. The operation of ren and yi within deliberative situational embeddedness and the regularity within situations yields behavior that very highly reliably co-varies with situational features, thereby offering additional theoretical resources to explain why Confucians may meet the “high-bar” argument (reconceived as reliable flexibility rather than mere regularity).
One might wonder how avoiding situational influences suffices for moral development. As Wielenberg points out, “there are definite limits to what Doris’s situation-management strategy can accomplish. We cannot always control the kinds of situations in which we find ourselves” (2006, p. 489). Sarkissian makes a similar point by noting that first, “in order to avoid a certain type of situation, one needs be aware of its eliciting a particular pattern of behavior” and second, “certain relationships or situations, even if known to elicit undesirable behavior, may nonetheless be practically unavoidable” (2010, p. 5). Beyond these pragmatic concerns for one’s own personal practice, Doris and Harman’s tactic of avoidance seems completely unrealistic on a broad scale for social policy or enacting a program of moral education.
Clearly an ancient Chinese program of moral education cannot meld seamlessly with modern educational policy, but there are rich theoretical resources and empirical insights to glean from Xunzi’s account of moral development and program of moral education.
One might interpret Xunzi as holding additional virtues beyond ren and yi. While I do not claim that ren and yi are the only virtues, I have only argued for a realist interpretation of ren and yi. Addressing the status of other possible virtues (e.g., filial piety) and at the various stages of moral development is far beyond the scope of this paper. A reviewer helpfully noted that filial piety may qualify as a virtue on my account. Indeed, it may, but in the absence of extended discussion and textual analysis, I take no position. However, given the sloppiness of the debate between virtue ethicists and situationists and varied conceptions of virtue, I predict that, in general, what may appear to be virtues are more likely to be (1) the descriptive and normative labels found at the first stage of moral development, (2) instrumental characterizations at the first and second stages of moral development, or (3) a mischaracterization of the changing function of rituals within moral education throughout the stages of moral development. Clearly, much more work needs to be done.
As Kristjansson (2008) notes, “some moral particularists have been eager to enlist Aristotle as their ally. The trouble for them is that in Aristotle’s view, phronesis adjudicates moral conflicts, and phronesis relies not only upon situational appreciation but also upon general moral truths. Thus, when Aristotle uses particular examples, he does not abandon generalizations and tell us to attend only to the particularities of the described situation; rather he describes the generalisations we should seek” (p. 65.) In contrast, Hutton (2002) notes that this type of deliberation “involves carefully weighing and mutually adjusting the reasons at work in a given situation within the broader framework of the end constituted by the Dao, which, like the role of eudaimonia in Aristotelian practical reasoning, is an overall conception of how to live upon which deliberation focuses” (p. 372). If Kristjansson is right, then perhaps Aristotle is not proposing a model of moral development similar to Xunzi’s, in which case, Aristotelian accounts of virtue ethics may pose a target for some situationist challenges. However, Hutton argues quite persuasively for a high degree of similarity between Xunzi’s and Aristotle’s accounts of virtue. Not being an Aristotelian, I will leave this issue to those who are.
One may worry that learning which features of situations are significant would reduce their effectiveness in just the same way as learning what panic attacks are and how they are caused tends to reduce the severity of their effects. However, if Xunzi is right that particular situational features induce emotion and are important for moral education, some may be cognitively impenetrable: a straight stick inserted into water looks bent, and it will continue to look bent despite my knowledge that the stick is straight. The effect of moral education through ritual may be similarly cognitively impenetrable, regardless of one’s knowledge of the causal efficacy or the significance of situational features. However, whether there are such cognitively impenetrable situational features or factors is an empirical question, and it is an empirical question that can only be answered through extensive long-term analysis throughout phases of moral education and stages of moral development. Discovering such information would be fascinating and could inform moral education in interesting ways by suggesting alternative strategies for addressing and responding to such situational features in moral education.