Ernest Sosa Encountering Chinese Philosophy: A Cross-Cultural Approach to Virtue Epistemology 9781350265806, 9781350265776

Citation preview

1

Ernest Sosa Encounters Chinese Philosophy An Introduction Yong Huang

1 Introduction Virtue epistemology can be seen as a parallel to virtue ethics in a number of aspects. First, both can be traced (at least) to Aristotle. Aristotle regards rationality as something distinctively human, and rationality for him can be divided into the theoretical and the practical. As virtue is merely excellence of rational activities of the soul, there are intellectual virtues and practical (moral) virtues. While virtue ethics is concerned with the latter, virtue epistemology is concerned with the former. Second, virtue ethics emphasizes not just the right action but the right action done through the characters of the moral agent. Similarly, virtue epistemology emphasizes not just the true belief but the true belief acquired through the character or competence of the epistemic agent: “Our virtue epistemology and virtue ethics focus rather on the agent and cognizer. When the agent’s actions are said to be right and the cognizer’s beliefs knowledge, we speak implicitly of the virtues, practical or intellectual, seated in that subject” (Sosa 2009: 189). Third, there is an impressive revival of virtue ethics in contemporary philosophy as a rival to deontology and consequentialism. Similarly, virtue epistemology has also emerged as a viable alternative to naturalized epistemology, with the former being normative and the latter descriptive. Fourth, contemporary virtue ethics has become pluralized, partially because different virtue ethicists today are appealing to different philosophers, such as Aristotle, Stoics, Hume, Nietzsche, Dewey, among others, for their inspirations. Similarly, contemporary virtue epistemology has also become pluralized, with different virtue epistemologists drawing on different sources such as Plato, Aristotle, Aquinas, Descartes, Hume, Kierkegaard, Nietzsche, Pierce, among others, in developing their ideas. However, there is a clear disanalogy between virtue ethics and virtue epistemology. As virtue ethics prospers in contemporary Western philosophy, scholars doing Chinese philosophy have quickly produced an impressive number of publications on virtue ethics in Chinese traditions in the form of journal articles, book chapters, and single-authored monographs. In contrast, virtue epistemology has not attracted its deserved attention from these scholars, with only a small handful of articles exploring the virtue epistemological 1

2

Ernest Sosa Encountering Chinese Philosophy

potentials in Chinese tradition so far published. In this context, it is significant that, in this volume, eleven experts of Chinese philosophy critically and constructively take the work of Ernest Sosa, the pioneer of contemporary virtue epistemology, and engage it in fruitful dialogues with classical Chinese philosophers, especially Confucius, Xunzi and Wang Yangming (Confucianism), Zhuangzi (Daoism), and Linji (Chan Buddhism).

2 Sosa’s virtue epistemology at a glance It is now accepted that there are two main schools of contemporary virtue epistemology, responsibilism and reliablism, and Sosa belongs to the latter. Sosa himself provides the following characterization of these two versions of virtue epistemology: One of these finds in epistemology important correlates of Aristotle’s moral virtues. Such responsibilist character epistemology builds its account of epistemic normativity on the subject’s responsible manifestation of epistemic character. The other form of virtue epistemology cleaves closer to Aristotelian intellectual virtues while recognizing a broader set of competences still restricted to basic faculties of perception, introspection, and the like. Sosa 2017: 140

As a reliabilist, Sosa argues that the faculty virtues or intellectual competences are constitutive of knowledge. While he does not exclude trait virtues in the formation of knowledge, he regards them as merely auxiliary, as we will see in Section 4 below. In several of his more recent writings (Sosa 2019, Sosa 2021, as well as his responses to contributors included in this volume), Sosa also highlights his reliabilist virtue epistemology as a telic virtue epistemology, according to which “the epistemic domain is one where we perform alethically, aiming at getting it right, whether through judgment (intentional and even conscious) or through functional perception or belief, where the aim would be teleological rather than intentional” (Sosa 2019: 15). There are at least three salient features in his characterization of his virtue epistemology as a telic one. First, just as virtue ethics is often contrasted with deontology, Sosa emphasizes that “we have a telic normativity in contrast with the deontic normativity of norms, obligations, permissions, and so on” (Sosa 2019: 15). Second, the goal that telic virtue epistemology emphasizes is intrinsic to it: true belief, and thus excludes external values from our assessment of any epistemic attempt. Sosa uses hunting in the form of archery as an analogy to explain this. A telic assessment of the archer’s attempt is to see whether she hits the target, while whether the archer’s shot “may bring food to the hunter’s starving family, or may constitute a horrible murder” is irrelevant (Sosa 2019: 16). Similarly, telic assessment of an epistemic attempt is to see whether it leads to true belief, which is its intrinsic goal, and bracket any consequentialist assessment of the utility or disutility of such a true belief. Thus, Sosa emphasizes that “the ‘virtue’ of our telic virtue epistemology is aim-relative (thus its ‘telic’ character). Telic assessment does not import any assessment of the aim. Telic excellence allows awful aims, as in a ‘perfect’ murder” (Sosa 2021: ch. 2, n. 11). Third, we have been saying that the intrinsic

Introduction

3

goal of epistemic attempt is true belief. To have reached such a goal indicates a successful attempt, which of course is better than unsuccessful attempt. However, Sosa emphasizes that “an attempt is a better attempt, it is better as an attempt, if competent than if incompetent; and it is better to succeed through competence—aptly—than through sheer luck” (Sosa 2019: 15). So a telic assessment of an epistemic attempt is not just to see whether it is successful but how reliably successful it is. This last feature involves Sosa’s distinction between animal knowledge and reflective knowledge, his conceptions of triple-A conception of knowledge and triple-S conception of competence. It is important to keep in mind that, for Sosa, animal knowledge is not (non-human) animals’ knowledge. He is talking about human knowledge, which has two varieties, animal knowledge and reflective knowledge. According to him, “animal knowledge does not require that the knower have an epistemic perspective on his belief, a perspective from which he endorses the source of that belief, from which he can see that source as reliably truth conducive. Reflective knowledge does by contrast require such a perspective” (Sosa 2009: 135). For example, when you see your hand when you are awake, the knowledge you have about your hand is animal knowledge if you don’t know you are awake. Reflective knowledge about the hand, in the contrast, requires not only that you have seen it but also that you know that you are not dreaming, among others (Sosa 2009: 21). As human knowledge, animal knowledge or unreflective knowledge, just like reflective knowledge, is epistemically related to reasoning, although in two different ways: “animal” knowledge “is concerned with the acquisition and sustenance of apt, reliable belief,” whereas reflective knowledge “requires the belief to be placed also in a perspective within which it may be seen as apt” (Sosa 2009: 75). In Sosa’s virtue epistemology, knowledge has an AAA structure: accuracy, adroitness, and aptness. As an analogy, Sosa uses his favorite archer’s shot to explain it. When assessing it, we first need to see whether it hits the target, which is what accuracy requires. Second, if it hits the target, we need to see whether this shot manifests the skill of the archer, which is what adroitness requires. Third, if it manifests the skill of the archer, we need to see whether the shot’s being accurate is because it is adroit, which is what aptness requires. While the first two points may be clear, the last point is not so. To explain this, Sosa illustrates: Take a shot that in normal conditions would have hit the bull’s-eye. The wind may be abnormally strong, and just strong enough to divert the arrow so that, in conditions thereafter normal, it would miss the target altogether. However, shifting winds may next guide it gently to the bull’s-eye after all. The shot is then accurate and adroit, but not accurate because adroit (not sufficiently). So it is not apt, and not creditable to the archer. Sosa 2007: 22

Similarly, a belief can be knowledge if it also fits this triple-A structure. First it has to be accurate, i.e. true; second, it must be adroit, i.e. it manifests epistemic virtue or competence; and third, it must be apt, i.e. it is true because it manifests the competence (see Sosa 2007: 23).

4

Ernest Sosa Encountering Chinese Philosophy

With this AAA structure, Sosa explains the animal knowledge and reflective knowledge in this way: animal knowledge is essentially apt belief without being required to be defensibly apt belief. In contrast, reflective knowledge is not only apt belief but is also defensibly apt belief, “i.e., apt belief that the subject aptly believes to be apt, and whose aptness the subject can therefore defend against relevant skeptical doubts” (Sosa 2007: 24). Sosa also distinguishes animal knowledge as of the first-order aptness, and reflective knowledge as of the full aptness. To understand this distinction, Sosa introduces the notion of meta-aptness or second-order aptness. A belief of the firstorder aptness is a true belief whose being true is due to the competence. A belief of the meta-aptness is a belief that is formed (instead of being withheld) after and because risks in forming this belief are competently assessed. Here, not only may a belief of the first-order aptness not be a belief of the meta-aptness, but a belief of the meta-aptness may not be a belief of the first-order aptness, as a belief that is formed after and because risks in forming (instead of withholding) the belief are competently assessed may still not be true and is thus not first-orderly apt. A reflective belief is a belief not merely of both the first-order aptness and meta-aptness but is a fully apt belief in the sense that its first-order aptness is due to or manifests its meta-aptness (Sosa 2011: 7–9). Closely related to this AAA structure is the SSS constitution of competence; so much so that Sosa often uses them together, regarding his virtue epistemology as an AAA/SSS account of knowledge. To illustrate, Sosa uses as an analogy a driver’s competence, which has a triple-S (seat, shape, and situation) constitution: (a) the innermost driving competence that is seated in one’s brain, nervous system, and body, which one retains even while asleep or drunk; (b) a fuller inner competence, which requires also that one be in proper shape, that is, awake, sober, alert, and so on; and (c) complete competence or ability to drive well and safely (on a given road or in a certain area), which requires also that one be well situated, with appropriate road conditions pertaining to the surface, the lighting, etc. The complete competence is thus an SSS (or an SeShSi) competence. Sosa 2017: 191–192

Similarly, we may say, for example, that a perceiving (epistemic) agent’s competence also has a triple-S constitution: (a) the innermost perceiving competence, say vision, that is seated in the perceiving agent; (b) a fuller inner competence, which, in addition, requires one to be in proper shape (for example, having not taken some medicine that may cause one to see blue things red and red things blue); and (c) complete competence to perceive well, which in further addition, requires one be well situated (for example, the object to be perceived is not too far away in the distance). Sosa emphasizes the importance of the SSS constitution of epistemic competence because it is crucial to meta-aptness and thus full aptness, discussed above. The function of meta-aptness is to assess the risk factors in forming or withholding a belief. In such an assessment, one should not only check whether one has a good innermost competence but also whether one is in good epistemic shape and in a good epistemic situation. So even if one has good vision (seat), but is not in good epistemic shape and not in a good epistemic situation, one may decide to suspend one’s judgement, instead of affirming or denying

Introduction

5

it. It is in this sense that Sosa connects the SSS constitution with full aptness: We need to consider factors that might go wrong under the SSS description: “whether we have the skill, while in proper shape and well enough situated . . . . This speaks in favor of the pertinence of full aptness in general, and of its epistemic importance in particular” (Sosa 2015: 111). Moreover, Sosa also uses his view of competence with the triple-S constitution to respond to situationist criticisms of virtue epistemology, similar to the situationist criticism of virtue ethics. According to this criticism, “we are surprisingly likely to be less reliable than we had thought and that we are made less reliable by influences astoundingly irrelevant to the truth of the beliefs that they nevertheless do influence. So the bottom line here is that we are less reliable epistemically than we had implicitly supposed all along” (Sosa 2017: 203). More specifically to competence, this criticism points out that “in a certain situation X our supposed competence to ø regularly fails us. And it is inferred that we do not really have any such competence. What we have is rather something like this: we have a ‘competence to ø when not in X’ ” (Sosa 2017: 204). In Sosa’s view, his triple-S conception of competence can easily respond to such a criticism, because this conception already has the situation built into it. In other words, competence is situation-relative. He uses a good three-point shooter in a basketball game as an analogy. This player has the competence of making a three-point shoot. However, clearly this competence is situation-relative: he cannot reliably make the three-point shoot in the center or at the other side of the court. Similarly, a person who has good vision may not be able to see things clearly in the dark. So, Sosa points out that “what is required for a competence to ø is only a disposition based on a certain pre-selected range of shape/situation combinations,” which may have already excluded the particular sort of situation that situationist has in mind (Sosa 2017: 205).

3 Sosa’s first encounter with Chinese philosophy Before touching on the main themes of this encounter between Sosa and Chinese philosophers, I would like to point out that this is not the first time that Sosa has experienced such an encounter. Indeed, among the small handful of papers published so far that explore the virtue epistemological potentials of Chinese philosophy is one by Sosa himself, “Confucius on Knowledge.” This is a paper he initially read at the panel “Virtue Turn: Constructive Engagement between East and West,” with Michael Slote, Eric Hutton, Michael Mi, and myself as co-panelists, held at the 2013 APA Eastern Division in Baltimore. It was later published in the journal I edit myself, Dao: A Journal of Comparative Philosophy in 2015. The paper not only provides an illuminating interpretation of Confucius’s view on knowledge, at least one aspect of it, but also introduces a fundamental distinction between two types of knowledge, animal knowledge and reflective knowledge, key to Sosa’s own virtue epistemology, which underlies many of the discussions in the chapters included in this volume as well as in Sosa’s responses to them. Sosa’s paper focuses on one single passage in the Analects, passage 2.17.1 After comparing and contrasting three different English translations of this passage, by

6

Ernest Sosa Encountering Chinese Philosophy

James Legge, Brooks and Brooks, and Edward Slingerland, respectively, Sosa renders it thus: “When you know, to know (recognize) that you know; and when you don’t know, to know (recognize) that you do not know; that is knowledge (wisdom).” Sosa provides what he regards as a banal interpretation and a more insightful interpretation of this passage. According to the former, “When you know that you know, you thereby know; and when you know that you do not know, you (also) thereby know”; and according to the latter “knowing that you know when you know, and knowing that you do not know when you do not know; that is real knowledge (a knowledge closer to wisdom)” (Sosa 2015a: 326). The question then, Sosa asks, is: which is the real knowledge? Indeed, in the Chinese commentary history over 2,000 years, this is also the question that has intrigued almost all commentators. So before we examine Sosa’s answer, it is helpful to provide a brief overview of how classical Chinese commentators have tried to understand it. Here I limit myself to the four most representative interpretations. According to the first, this simply means that the person knows that he knows something and that he doesn’t know something. For example, the Song dynasty scholar Xing Bing states that “to regard what one knows as known and to regard what one doesn’t know as not known is genuine knowledge. In contrast, to conceal what one knows as unknown and to fancy what one does not know as known is not knowledge” (Xing 1990: 17).2 According to the second interpretation, it refers to the way to know in the sense that only when one knows that there are things that one does not know yet can one be motivated to know them. For example, Cheng Yi states that “if one is ashamed of being known as not knowing something and thus does not ask people about it, then he will never know it; in contrast, to acknowledge that one does not know something and thus seek to know it, then he will know it. This is why Confucius says that ‘that is knowledge’ ” (Cheng and Cheng 1989: 1134).3 According to the third interpretation, it means the knowledge that there are things that can be known and there are also things that cannot be known. Xie Liangzuo, one of Cheng Yi’s students, for example, states that, on the one hand, “one shouldn’t not try to know what is knowable,” and, on the other hand, “one should not try to know what cannot be known” (see Zhu 2002: 86).4 According to the fourth interpretation, which connects this Analects passage with a similar but much more extended passage in the Xunzi, the point of this passage is to be honest: if you know something, say that you know it; and if you don’t know something, then say that you don’t know it; to be so is to have wisdom. This could make sense as the student that Confucius talks to, Zilu, is known for his tendency to pretend to know things that he doesn’t know. Included in the Xunzi passage is the following saying, attributed to Confucius, which makes a contrast between inferior persons and superior persons and draws an analogy between knowledge and ability: Those who are overconfident in saying are flashy, and those who are overconfident in doing are showy. A person pretending to have knowledge and ability is an inferior person. In contrast, superior persons say that they know when they know and say that they have abilities when they have them. This is key to saying. When they have abilities, they say they have abilities, and when they don’t have abilities

Introduction

7

they say they don’t have abilities. This is key to action. To hold the key to speech is knowledge/wisdom, and to hold the key to action is benevolence. Xunzi 29.6

Sosa, as a contemporary philosopher in the analytic tradition and not a scholar of Chinese thoughts, is of course not familiar with these various interpretations in the Chinese commentary tradition. However, on the one hand, Sosa believes that “there is nothing at all momentous in the knowledge of one’s own epistemic states, such as this is something one does know or something one does not know” (Sosa 2015a: 327). Of course, it is possible that Confucius does not mean to say anything momentous, but if he does, or if we think he does, then we have to exclude almost all of the first three interpretations above, since they are about one’s own epistemic states. On the other hand, Sosa states, “[n]or does it help much to focus on the potential ambiguities in zhi and to suggest that the last occurrence should be read outright as ‘wisdom’ rather than ‘knowledge’ ” (ibid.). If so, possibly the fourth interpretation above will also have to be excluded. So, if we believe that Confucius says something momentous and helpful, Sosa suggests that “we need to have a fresh look at the Confucian saying, in search of a reading on which the saying will contain some insight” (Sosa 2015a: 327). Sosa does acknowledge that it is only one good way, not the only way, to seek insight from this Confucian saying, although he also does not say that this must be the insight that Confucius meant to convey in the saying. To do so, Sosa’s focus is “on the importance of metacognition, and of reflective as opposed to merely animal knowledge” (Sosa 2015a: 327). The distinction between animal knowledge and reflective knowledge is an important one that Sosa makes within human knowledge, and so it is good to keep in mind that Sosa’s animal knowledge is not knowledge that can also be possessed by non-human animals, as I have pointed out earlier. According to Sosa’s AAA model, a belief can be regarded as animal knowledge only if it is true (accurate), it manifests the belief holder’s competence (adroit), and its being true is because of the manifested competence (apt). Animal knowledge is the ground level knowledge. In contrast, “reflective knowledge, while building on animal knowledge, goes beyond it” (Sosa 2009: 75). Crucial to reflective knowledge is the risk assessment on the basis of which one affirms the animal knowledge, denies it, or suspends one’s judgement about it. One way to explain the difference is the example used by Descartes and Moore. Being awake, we form a true belief that there is a fire in front of us, we exercise our competence of perceiving the fire, and our belief is true because of our exercising our perceptive competence. This is animal knowledge.5 Then, on the meta-level, we decide whether to endorse this animal knowledge by making a risk assessment: even though we are actually awake, do we really know that we are awake instead of dreaming? If we know, also on the AAA model, that we are indeed awake, and then endorse the ground-level animal knowledge that there is a fire in front of us, then this animal knowledge that is thus endorsed on the meta-level becomes reflective knowledge. With this distinction between animal knowledge and reflective knowledge, Sosa provides the following interpretation of the Analects passage in question:

8

Ernest Sosa Encountering Chinese Philosophy When you have first-order (animal) knowledge, to know (recognize) that you do know; and when you do not have such knowledge, to know (recognize) that you do not know; “that” is (reflective) knowledge (where the reference is to the firstorder knowledge, which rises to the better, higher level of reflective knowledge, and does so through the proper, “recognitional” second-order endorsement).

In other words, according to Sosa’s interpretation, Confucius here is first concerned with the ground level animal knowledge that one either has or does not have, and then with a meta-level knowledge that makes a judgement about the animal knowledge after the risk assessment, which is reflective knowledge. Sosa’s interpretation of this Analects passage is indeed novel, momentous, and helpful, as he claims. Interpreted this way, what Confucius has in mind is not merely a platitude of being honest, a psychological state about whether one has a particular knowledge, an awareness of the existence of both things knowable and things unknowable, or quantitative accumulation of knowledge (acknowledging one’s lack of certain knowledge is the beginning of the search for it). Rather, Confucius is concerned about the quality of the knowledge: how to move from animal knowledge to reflective knowledge. One may of course contend whether this interpretation, however novel and insightful, is what Confucius really intends, but conceivably one can make such a contention with any of the four main interpretations of the same passages by the Chinese commentators briefly mentioned above.6

4 The divide between responsibilism and reliabilism in virtue epistemology As mentioned above, there are two main schools of virtue epistemology, responsibilism and reliabilism, and Sosa is considered to be the most important advocate of the latter. Understandably, a significant number of contributors to this volume critically engage Sosa’s reliabilist virtue epistemology. By drawing on various Chinese philosophical ideas, they either argue that the distinction between these two cannot be adequately made or that trait virtues can be constitutive of knowledge and faculty virtue may be not so constitutive. Kim-chong Chong, in Chapter 2, by drawing on pre-Qin Chinese philosophers Xunzi and Zhuangzi, challenges Ernest Sosa’s reliabilist virtue epistemological claim that such virtues as open-mindedness and intellectual courage are not knowledge constitutive but are at most auxiliary to knowledge-constitutive competences (manifesting accuracy, adroitness, and aptness). Chong argues that Xunzi, by emphasizing veridical perception and reasoning about and examining things under the appropriate condition, is consistent with Sosa discounting certain character traits as necessarily helpful to achieving knowledge.7 In contrast, for Zhuangzi, motivational attitude and dispositional characters, such as humble acceptance of one’s limitation, open-mindedness, and willingness to learn from discussion, play a role in cognition. However, Chong is not content to making this modest claim that, while Xunzi, like

Introduction

9

Sosa, is a reliabilist virtue epistemologist, Zhuangzi, unlike Sosa, is a responsibilist virtue epistemologist. Rather he aims to show that, in an important sense, both Xunzi and Zhuang pose a challenge to Sosa. In Chong’s view, the reason that Sosa holds exclusively the reliabilist view of knowledge is that he assumes that there is a pure epistemic understanding of knowledge that can be described in terms of necessary and sufficient conditions. Chong argues, however, that knowledge always takes places in contexts, and in different contexts what constitutes knowledge is also different. Xunzi’s account of knowledge is presented in the context of his defense of Confucian order, and thus describes cognitive, instrumental thinking and non-distorted perception as constituting knowledge (reliabilism); Zhuangzi’s account of knowledge, however, is presented in the context of his criticism of Confucian narrowness and monism, and thus accepting such limitation and being open-minded become virtues that constitute knowledge (responsibilism). Chris Fraser, in Chapter 3, also draws on Xunzi in his dialogue with Sosa. On the one hand, he argues that Sosa’s conception of full aptness echoes and further elucidates Xunzi’s discussion of epistemic pitfalls. More specifically, Fraser notices Xunzi’s warning against forming a belief or making a judgment when the epistemic agent has the illusion, intoxication, unreliable means, and incompetent testimony, and he argues that this “converges with Sosa’s view that one dimension of an epistemic agent’s competence is competence in assessing the reliability of one’s own cognitive operations in various contexts and in guiding one’s epistemic attitudes accordingly.” On the other hand, however, Fraser argues that “Xunzi’s treatment of the epistemic agent’s awareness of and commitment to norms of judgment helps to enrich Sosa’s view of epistemic agency,” in two important respects. First, Fraser observes, in his distinction between functional belief, which is not intentional, but formed automatically in a way similar to how the heart circulates the blood, and judgmental belief, which is deliberative, Sosa is biased toward the latter, with the view that the former lacks full agency. In contrast, Fraser argues that, for Xunzi, “an aspect of epistemic competence is that the agent actively manages the operations of the heart/mind so as to perform well epistemically. The agent seeks to improve both explicit judgment . . . and immediate, implicit distinction-drawing.” Indeed, for Xunzi, like expert performers in many other fields, the expert epistemic agents, due to their expertise, often render explicit, deliberative judgment unnecessary.8 The second is also related to Sosa’s reliabilist virtue epistemology vs. the responsibilist one. Fraser argues that, in order to aim for truth, the agent has to approve it, and to approve it is to endorse certain values or norms, which is “characteristic of responsibilist virtues such as epistemic conscientiousness, diligence, or perseverance.” So a responsibilist commitment to such values is a necessary condition for competence in applying criteria that underwrite reliabilist virtues. Gregor Paul’s chapter is also focused on the divide between responsibilism and reliabilism (Chapter 4). It argues that Sosa’s virtue epistemology, which, as a version of reliabilism, is based on faculty virtues such as perception, memory, and inference, is too narrow. Thus, it should be enriched by the addition of trait virtues, such as intellectual courage, open-mindedness, love for truth, etc. emphasized by responsibilist virtue epistemologists, in order to have a more comprehensive virtue epistemology, since these two types of virtues are “compatible with each other, even fruitfully

10

Ernest Sosa Encountering Chinese Philosophy

complementing each other.” In itself, I think Sosa has no disagreement with this, as Sosa, while emphasizing faculty virtues, also thinks that these virtues are compatible with, and can be supplemented by, trait virtues. By appealing to Xunzi, whom he characterizes as a responsibilist reliabilist virtue epistemologist, however, it seems that Paul wants to emphasize three unique features of his comprehensive virtue theory of knowledge. First, disagreeing with Sosa, Paul argues that not all faculty virtues are constitutive of knowledge and not all trait virtues are merely conducive to knowledge; second, not a separate means of gaining and justifying, each virtue, whether as a faculty or as a trait, is an integral part of a whole of interrelated and interdependent powers/ virtues that comprise a reliable source of knowledge; and third, (at least some) moral virtues are also epistemic ones, which he supports by citing Xunzi’s saying that “benevolence (ren), rightness (yi), and (morally) virtuous conduct are reliable means to safety,” arguing that safety here also requires exercising correct perception, inference, and consistent reflection and thus includes epistemic safety. Tao Jiang’s chapter (Chapter 5) starts with a general characterization of the debate between reliabilism and responsibilism in contemporary virtue epistemology, with the former focusing on epistemic competence in perception, memory, inductive and deductive reasoning, which are regarded as faculty virtues, and the latter on curiosity, open-mindedness, and intellectual courage in epistemic pursuit, which are regarded as trait virtues. In distinguishing between these two types of virtues, Jiang seems to have adopted Sosa’s view that the faculty virtues are constitutive of knowledge, while the trait virtues are merely auxiliary in epistemic pursuit. Then he turns to Chan Buddhism, with a focus on the Linji school, arguing that the central virtue in Buddhism, detachment, is one of the important epistemic virtues that have largely remained outside the purview of both schools of contemporary virtue epistemology. More importantly, Jiang argues that this virtue of detachment, in the context of Buddhism, is both responsibilist and reliabilist. It is responsibilist since detachment is clearly a trait virtue that has to be cultivated and vigorously trained in Buddhist practices. It is reliabilist, however, because it is constitutive of the Buddhist knowledge, wisdom, and not merely auxiliary in our pursuit of this knowledge. This is because the opposite of knowledge or wisdom is ignorance, and in Buddhism ignorance is caused by “the pernicious and ubiquitous attachment that permeates all of our cognitive activities,” including those aiming to solve the Gettier-type problems, the single most important motivation of contemporary philosophy to turn to virtue epistemology. So in order to overcome ignorance and acquire knowledge/wisdom, detachment is necessary. I assume that Jiang’s point can also be made in a different way: since detachment is a trait virtue and is constitutive of knowledge, Sosa is wrong to claim that trait virtues are merely auxiliary and not constitutive of knowledge. Xiang Huang’s chapter (Chapter 6) compares Sosa’s virtue epistemology, which is concerned with the epistemic performance, with Wang Yangming’s meta-ethics, which is concerned with moral performance. On the one hand, he uses Sosa’s AAA model of epistemic performance to explain Wang Yangming’s unity of knowledge and action as moral performance. On the other hand, he uses Wang’s unity of knowledge and action to reconcile the conflict between Sosa’s reliabilism and responsibilism defended by Jason Baehr, among others. The key in Huang’s interpretation is Wang Yangming’s

Introduction

11

analogy of the innate knowledge of the good (liangzhi) as root and concrete moral practice as the branches and leaves, with the latter naturally growing from the former. On this analogy, Huang claims that the root of virtue epistemology is the AAA structure, and its branches and leaves are different intellectual virtues functioning as norms that guide concrete epistemic practices. Through illuminating examples, Huang shows that in some domains of epistemic practices, competence virtues, highlighted in reliabilism, are indeed constitutive of knowledge, while character virtues, emphasized in responsibilism, are merely auxiliary; in some others, however, the opposite is true; and in still some others, some combination of character virtues and competence virtues are constitutive of knowledge, while some other character virtues and competence virtues are merely auxiliary. He thus concludes that, in this contextualist understanding, following Wang Yangming’s analogy of root and branches/leaves in his explanation of virtue and virtuous practice, “no intellectual virtue can always be knowledge constitutive in all epistemic practices, even though in a determinate epistemic practice, we can always determine some virtues are constitutive and some are auxiliary.” As many attempts have been made in the above chapters to challenge Sosa’s view that only faculty virtues are constitutive of knowledge and trait virtues are merely auxiliary to knowledge by drawing on Chinese philosophical resources, a large part of Sosa’s response is focused on this issue. His main point is that nothing in Chinese virtue theory could favor responsibilist over reliabilist virtue epistemology, which he supports by clarifying his reliabilist version of virtue epistemology. First, Sosa argues that much dissatisfaction with reliabilism originates from a simplified and misconceived understanding of it as the simplest perceptual knowledge, as it focuses on the so-called faculty virtues, such as perception, memory, inference, etc., and talks about animal knowledge. Sosa argues that reliability required by virtue reliabilism for the epistemic standing of a belief is a necessary condition. It is never supposed that all knowledge is to be found at the level of animal knowledge. Nowhere in the writings that develop virtue reliabilism is there any restriction to such basic faculties or to the representation that we humans share with lower brutes so as to give us our animal knowledge.

Sosa here makes two related points. On the one hand, the animal knowledge that humans have is not the animal knowledge that beasts can have. What he means by animal knowledge is “the knowledge that a knower would have even unaided by any reflective higher-order perspective,” but Sosa points out that this does not mean that human knowledge on this level is purely animal, as the higher order of endorsement is already involved, at least implicitly, in such knowledge. On the other hand, there is a need to go beyond animal knowledge to reflective knowledge. So while the human animal knowledge is already different from animal animal knowledge, virtue reliabilism also recognizes a higher human judgment which involves higher faculties. Second, to respond to the criticism that his reliabilist virtue epistemology overlooks character traits of special interest to virtue responsibilism, Sosa emphasizes that the epistemology is not a branch of ethics, i.e. intellectual ethics, which, according to Sosa,

12

Ernest Sosa Encountering Chinese Philosophy

is on a par with business ethics and medical ethics, ethics restricted in intellectual performance. The virtues that responsibilists highlight, in Sosa’s view, in so far as they are supposed “to bear on the personal worth of the agent largely because of their distinctive motivation and respect for truth,” do not belong to theory of knowledge but to intellectual ethics. He often uses the archer’s shot as an analogy of knowledge. The goal of the shot is to hit the target, “independent of the moral concerns such as whether the act is a murder.” One example he uses as an analogy is the tennis assessment of a brilliant ace, “despite the cruelty you might find in that ace as a concluding demolition of a much weaker opponent who had won not a single point in that match,” as “cruelty is just irrelevant assessment of that blazing ace as a great serve.” A commentator of this performance who acclaims that it is a fine shot is making an apt and perhaps fully apt belief claim, even though it may be cruel, inappropriate, and deplorable to the losing player and his/her family and close friends. Sosa makes it clear that this does not mean that intellectual ethics is unimportant or “those who aspire to know must disregard the moral or other practical or aesthetic values that may be relevant to inquiry or to any intellectual action,” but only that it is not part of the theory of knowledge and that “in assessing whether a belief amounts to knowledge, we must focus on the gnoseological competences aimed at reliable enough attainment of truth and aptness.” Third, Sosa claims that some character traits can be both moral/prudential and gnoseological. We have seen above that moral and prudential character traits belong to intellectual ethics, but gnoseological character traits belong to theory of knowledge. However, Sosa claims that even gnoseological “character traits of responsibilists have limited significance for the history of epistemology, with its central focus on issues of philosophical skepticism.” By the limited significance, Sosa means two things. On the one hand, such traits as open-mindedness and intellectual courage are not constitutive of but merely auxiliary to knowledge, in the sense that they put us “in a position to know the answers to questions into which we may inquire.” Elsewhere Sosa uses as an analogy finding out what is in a mysteriously shut box. In order to do so, one first needs to have the skills and competences to open the lid of the box. Such skills and competences are thus necessary for you to see what is inside the box, but these are not skills and competences whose exercises constitute your knowledge of what is inside the box, but are merely ones that put you in a position to exercise your perception, which is constitutive of your knowledge of what is inside the box (Sosa 2017: 143). In Sosa’s view, open-mindedness and intellectual courage, even when epistemically considered, are merely virtues auxiliary to knowledge in the sense that they put us in a position to know or to exercise our competence in such a way. In this sense, detachment in Buddhism, which Jiang claims is a character virtue constitutive of knowledge, is no different for Sosa. Detachment assumes that we originally have the competence to get things right, which is obstructed by attachment. The detachment unblocks our original competence to get things right, but it is the exercise of the original competence, and not its unblocking, that is constitutive of knowledge. With detachment, the original competence becomes unblocked and thus can be exercised, but if the original competence unblocked by detachment is not exercised, there will still be no knowledge.

Introduction

13

On the other hand, Sosa acknowledges that some of the character traits can be constitutive of knowledge, but even in such cases they play this role “not in the way of responsibilist virtue epistemology,” but in the way of a reliabilist one, in the sense that they enhance the reliabilility of one’s judgment or belief. They are constitutive of knowledge because they are virtues directly related to faculty virtues. Still using the example of the mysterious box: suppose that the lid is finally open, and a person is now in a position to see what is inside with his faculty of perception. However, if the person is careless or inattentive in perceiving it, he may not be able to have the apt knowledge of what is inside. In this sense, proper care and attentiveness can be manifestations, in a particular case, of stable character traits of an epistemic agent. As such, they will help constitute the complete intellectual competences or virtues that the agent exercises and manifests in particular judgments, and in the correctness of these judgments. So these traits will be implicated in competences whose manifestations might constitute the agent’s knowledge. Sosa 2015b: 45

The example he uses in his responses here to illustrate the same point is the character trait of conscientiousness. When it “avoids any epistemic negligence or recklessness in arriving at one’s judgment, which results in rash judgement omitting sufficient inquiry, conscientiousness is constitutive of knowledge in the sense that it makes it more reliable.”

5 Other issues of Sosa’s virtue epistemology In addition to the divide between reliabilism and responsibilism, contributors to this volume also engage a number of other aspects of Sosa’s virtue epistemology from one or another Chinese philosophical perspective.

5.1 Testimony Winnie Sung’s chapter (Chapter 7) focuses on Sosa’s view of testimonial knowledge in contrast to knowledge from instruments. On the issue of testimonial knowledge, Sung identifies the two extreme views: one being a reductivist view that testimonial justification is reducible to reliability of instruments, and the other being a nonreductivist view that testimonial justification is unique. Then she characterizes Sosa’s view as something that falls in the middle. The purpose of her chapter is to support Sosa’a apparently contradictory view by drawing on the pre-Qin Confucian philosopher Xunzi’s understanding of testimony, which leaves open both the reductivist and nonreductivist views of testimonial knowledge. The key text Sung uses from The Xunzi is a paragraph in which Xunzi seems to suggest that the heart/mind of the hearer always makes a decision to accept or reject the testimony presented by the speaker, which is preceded by the hearer’s registering a piece of information as testimony. In order for the hearer to register the received information as a piece of testimony to a

14

Ernest Sosa Encountering Chinese Philosophy

proposition p, Sung argues that the hearer believes that giver (1) makes an utterance U, (2) means by U that p, and (3) believes what she means by U. Once the hearer registers a piece of information as a testimony to p, he will have to decide whether to accept p, and if he decides to accept p, it can happen in one of the two ways. He can accept it simply because there is a testimony to it given by the speaker, where testimony itself is a sufficient warrant irreducible to anything else; he can also accept p because he accepts the testimony to p after deliberation of the testimony and deems it acceptable. In the former, we have the non-reductivist account of testimonial knowledge, while in the latter we have the reductivist one. In his response, Sosa accepts Sung’s characterization of his view on testimonial knowledge as something between reductivism and non-reductivism, and he aims to further develop his view in his previous publications that Sung’s chapter relies on. The key to this new development is Sosa’s distinction between mere uptake as if p and “gathering” that p. The distinction is apparently similar to Sung’s distinction between the two ways of accepting p after registering some information as testimony to p, discussed toward the end of the last paragraph. However, for Sosa, the former is nonfactive, while the latter is factive, as the gathering is more than just an accidental acquisition of the fact, but to gain possession of it through one’s own doing, which can be aided by other entities, such as a testifier who properly reports it. Here Sosa emphasizes not only the fact but also the testifier’s proper reporting it instead of malfunctioning or straightforward lying, as in the latter case the testifier’s reporting a fact is mere luck. In addition, Sosa stresses that the receiver’s “getting it right that p must itself manifest pertinent competence” on his part. In one sense, Sosa argues that testimony is not different from perception and instruments, as in all of the three sources of knowledge there is a graduation from mere seemings as if p to gathering that p. In another sense, however, testimony is a different instrument as a source of knowledge: we learn to trust instruments by relying on other sources such as perception and testimony, but the testimonial source of knowledge has the “sort of default standing found also in our perception.”

5.2 The animal knowledge/reflective knowledge distinction Leo Cheung’s chapter (Chapter 8) starts with an attempt to see Xunzi’s conception of knowledge of Dao from the perspective of Sosa’s distinction between animal knowledge and reflective knowledge and yet ends with an attempt to see Sosa’s distinction between these two types of knowledge from the perspective of Xunzi’s knowledge of Dao. While Sosa’s view on the distinction between these two types of knowledge changes over time, his mature view, Cheung claims, is that an apt belief is an instance of animal knowledge, while an apt belief that the subject aptly believes to be apt is an instance of reflective knowledge, and so reflective knowledge is animal knowledge but not vice versa. Seen from this perspective, Cheung claims that knowledge of Dao, according to Xunzi, is not animal knowledge, since it is purely internalist, acquired by the heart/ mind’s being empty, unified, and quiet (xu yi er jing 㲋а㘼䶌), and in this sense purely formal; and for this reason it is not reflective knowledge either, since reflective knowledge is animal knowledge, although going beyond it. However, Cheung argues

Introduction

15

that Sosa’s definition of reflective knowledge as an apt belief that the subject aptly believes to be apt is unreasonably restrictive from the perspective of Xunzi’s conception of knowledge of Dao. Cheung argues that a belief that the subject aptly believes to be apt does not have to be apt. So he suggests that Sosa revise his definition so that “reflective knowledge is a true [instead of apt] belief that the subject aptly believes to be true, if the content of the subject’s belief is formal, or apt, if otherwise.” Defined this way, knowledge of Dao, though not animal knowledge, is an instance of reflective knowledge. In his response, Sosa is not convinced that his definition of reflective knowledge should be weakened so that reflective knowledge is an apt belief that one’s first-order belief is merely true and not apt. Sosa’s worry is that, in Cheung’s proposed definition, the reflective knowledge, the second-order apt belief, collapses into animal knowledge, the first-order apt belief. If one’s first-order belief is merely true, and not apt, reflective knowledge is merely the apt belief that the first-order belief is true, which is equivalent to the first-order belief that is apt. I think the disagreement between Cheung and Sosa is whether there is an exception to Sosa’s strong definition of reflective knowledge when the first-order belief is merely formal, such as knowledge of Dao in Xunzi’s conception.

5.3 The Gettier problem As one of the primary motivations of contemporary virtue epistemologists, including Sosa, is to provide some anti-luck mechanism to address the so-called Gettier problem, Yingjin Xu, in Chapter 9, aims to argue that the Confucian idea of rectification of names can deal with the Gettier problem better than Sosa’s virtue epistemology does. In Xu’s view, contemporary epistemology in the analytic tradition includes both Gettier who raises the famous problem and many others who attempt to solve the problem. Thus, a belief ’s being deduced from another in a valid way is viewed as evidence for the first one’s being justified. In contrast, in Confucian tradition, the emphasis has been on the intra-belief relationship, the relationship between the subject-term and the predicate-term of an integrated belief. According to the theory of rectification of names, a belief, “A is B” for example, is justified, or a name, “A” for example, is rectified, if A is consistent with the norms embedded in B. For example, in the first case that Gettier discusses, the belief “the man who will get the job has ten coins in his pocket” is considered to be justified as it is deduced from the justified beliefs “Jones is the man who will get the job, and Jones has ten coins in his pocket.” On the Confucian model of rectification of name, instead of one belief deduced from another, Xu claims, we have an integrated belief, “Jones, who is supposed to be the job-winner, has ten coins in his pocket.” In order for the belief to be justified, the name “Jones” has to be rectified under the name of “job-winner”; in other words, Jones has to be instantiating norms imbedded in the job winner. In this way, he claims that the above integrated belief is not justified. In his response, Sosa points out that there are difficulties in (his understanding of) Xu’s Confucian solution to the Gettier problem. To explain, he uses Xu’s treatment of the second case of the Gettier problem as an example. In this second case, the belief

16

Ernest Sosa Encountering Chinese Philosophy

that “either Jones owns a Ford, or Brown is in Boston” is considered to be justified as it is validly derived from the justified belief “Jones owns a Ford.” Xu argues that, on the Confucian model of rectification of names, the belief is not justified, since neither of the two disjuncts is justified: “Brown is in Boston” is unjustified in Gettier’s design, and “Jones owns a Ford” is also not justified because the name “Jones” cannot be rectified in light of the norm embedded in the “Ford owner.” While I’m still not convinced by Xu’s argument that there cannot be justified but untrue belief except on some very complicated issues in sciences, Sosa’s difficulty is in accepting Xu’s claim that a disjunction is justified if and only if at least one of the disjuncts is justified. For example, even if neither the belief p nor the belief not-p is justified, I can still have the justified belief “either p or not-p.”

5.4 Skepticism Yiu-ming Fung’s chapter (Chapter 10) focuses on the issue of skepticism about our knowledge of the external world, claiming that Sosa is not entirely successful in arguing against it and the Chinese Daoist philosopher Zhuangzi can help him do a better job. Fung focuses on Sosa’s argument against skepticism in his Reflective Knowledge: Apt Belief and Reflective Knowledge. In that book, Sosa takes as example of an adult victimized by Descartes’ Demon or brain enslaved in Putnam’s vat. Such a person may be indistinguishable from the best of us in terms of the comprehensiveness and coherence of beliefs and experiences. However, Sosa argues that this victim’s beliefs, even when right about environing objects and events, i.e. even when they are true, are still short of knowledge, because his mental coherence might be detached from his environing world, which deprives him of reliable access to truth and leaves him short of cognitive or intellectual competence. Fung argues that there are two problems with Sosa’s argument. On the one hand, Fung argues that the victim is not detached from the external world, although his way of getting connected to the external world is different from ours. On the other hand, Sosa’s own connection to the external world is not sufficient, because the relatum on the side of the external world is not identified in Sosa’s theory. Then he presents Zhuangzi’s view as an alternative. As Fung understands it, the key feature of this alternative is that there is no entity in reality that can play the role of the base on which to make any epistemological relation, whether correspondence, representation, stimulation, or causation. Reality and what is experienced for Zhuangzi are two different states of the same world: the natural state before epistemic invasion and the artificial state full of epistemic coloring. In his response, Sosa mainly focuses on what Fung considers to be the first problem in his (Sosa’s) argument against skepticism using the example of “brain in vat”: Fung’s argument that the envatted victim may still be connected with, and not detached from, the external world. Sosa acknowledges that this is logically or even metaphysically possible. But instead of indulging himself in what he considers to be such wild science fiction, Sosa says that he opts for a less radical notion of detachment in the sense of “not normally attached, in accordance with human normalcy.” In this sense, the envatted victim may indeed be detached from his environing world.

Introduction

17

5.5 Knowing-to and besire One of the two Platonic questions that Sosa aims to answer is about the nature of knowledge, with the other being about the value of knowledge. Sosa states that knowledge “amounts to belief that is apt, belief that is an apt epistemic performance, one that manifests the relevant competence of the believer in attaining the truth.” After quoting this, Hansen (Chapter 11) claims that the classical Chinese view on the nature of knowledge is different from Plato’s in that, while the latter focuses on knowing-that (believing), the former is on what Gilbert Ryle calls knowing-how (performing). Although Sosa regards knowing as a kind of performing, Hansen complains that “Sosa’s formulation still ties us into the intellectual Greek Rationalist conceptual neighborhood of knowing as believing”; in other words, in Sosa, the epistemic performance is nothing but the intellectual performance of believing, while in classic Chinese philosophy, the epistemic performance is knowing-how, which includes knowing of, knowing-to, and knowing how to, among others, involving performance that is not purely intellectual or, using Ryle’s term, is anti-intellectualist. For example, knowing (how) to ride a chariot is a non-intellectual, or anti-intellectualist but practical, performance of knowing. It is in this sense that he argues against what he considers to be Sosa’s attempt to translate knowledge constructions like knowing-how (to) into knowing-that constructions using the word “ought.” Not denying that there is also knowing-that among classical Chinese philosophy and regarding knowing-to, instead of a subset of knowing-how, as a third type of knowing, along with knowing-that and knowing-how, Yong Huang’s chapter (Chapter 12) focuses on knowing-to, especially as it is developed by the neo-Confucian philosopher Wang Yang-ming. Huang argues that the notion of knowing-to provides an answer to the third Platonic question, the question about the impact of knowledge, the one that Sosa’s epistemology does not deal with, along with the other two Platonic questions, the questions about the nature and value of knowledge respectively, that Sosa aims to answer. By the third Platonic question, Huang means the one raised in Plato’s dialogue, Protagoras: whether it is possible for one to know something normative without acting according to such knowledge. Huang claims that Sosa allows a person knowing full well even if the person does not act according to such knowledge, which is unintelligible from Wang Yangming’s point of view: how can a person, for example, who knows that he ought to love his parents but never loves his parents, be regarded as knowing full well that he ought to love his parents? In Wang’s view, there has never been one who knows and yet does not act; to know without acting is really without knowing. So a person who knows something normative is always inclined to act according to this norm. Huang characterizes the type of knowing that Wang has in mind here as knowing-to. Knowing to love one’s parents, for example, is not merely knowing, however aptly in Sosa’s sense, that one ought to love one’s parents, but is also the knowing that motivates the knower to actually love their parents. Since knowing-to is thus not merely a belief (that one ought to love one’s parents) but also a desire (to love one’s parents), Huang also characterizes such a knowing as besire, not the mental states of belief and that of desire mixed together, but one single mental state which is both belief-like and desire-like. In this sense, Huang claims that either the person who

18

Ernest Sosa Encountering Chinese Philosophy

knows full well in Sosa’s sense is not really knowing full well or there can be someone who knows better or more fully well than the person who knows full well in Sosa’s sense. In his response, Sosa acknowledges that it would speak better of people who don’t act despite their belief if their will aligned better with their belief, but he says that it is not so obvious that this is because of some failure on the side of their belief. The question is, Sosa realizes, that what we are encountering here is not merely a mental state of belief but a single mental state, which, in addition to being belief-like, is also desire-like, in which one cannot believe something without at the same time desiring it and vice versa. To respond, Sosa starts with the desire-like aspect of besire, which is to aim to do the thing that is then to be done. He asks the question: suppose you do do what is to be done (the question to be asked is, though, whether you, who knows full well what is to be done, will do it), how do we assess your action? Sosa says that the adroitness or competence of the action (to do what is then to be done) largely and even wholly resides in one’s deliberation on what is to be done. The deliberation is intended to yield a judgment on whether you are to do what is to be done. So the adroitness of your action is derived from the adroitness of your judgment. The question seems to remain: will the judgment, because of its adroitness, lead to the corresponding action, the action you judged to be done, whether the action itself is done adroitly or not?

6 Conclusion In his response to critical chapters included in this volume, engaging his virtue epistemology from Chinese philosophical perspectives, Sosa states that I am delighted to have learned how deeply ingrained in Chinese philosophy is the interest in theorizing about such phenomena. Even if the greats of the Chinese traditions do not agree on every point with the contemporary virtue-theoretic approaches, just as they do not agree on every point with each other, and just as there is similar diversity among contemporary theorists, it is still gratifying indeed to learn of the deep similarities of theoretical aims along with parallel ways and means.

He further states that, to respond to such critical chapters, “I am led to some new extensions of the virtue theoretic account, and I am especially grateful for the probing and comparisons that have let to those extensions.” However, this encounter between Ernest Sosa and Chinese philosophy is certainly a two-way bridge. It also provides a good opportunity for scholars in Chinese philosophy to read the texts by a leading virtue epistemologist in Western philosophy carefully and, with this acquired new perspective, to read ancient Chinese philosophical texts they are otherwise familiar with anew. As one of these scholars in Chinese philosophy, I have to confess that I have learned a great deal about the fascinating issues of virtue epistemology and intrigue solutions that Sosa presents in his writings, even though in my own contribution to

Introduction

19

this volume I am trying to provide a Confucian amendment to his version of virtue epistemology. This is a living testimony that Chinese philosophy and Western philosophy can be better studied when they are brought together, and people studying them are interested in listening to and even arguing with each other.

Notes 1 2

3

4

5

6

7

Here is the Chinese text of the passage: ⭡ʽ䃘ྣ⸕ѻѾ˛⸕ѻ⛪⸕ѻˈн⸕⛪н ⸕ˈᱟ⸕ҏ. There is also the view that the last sentence in the Analects passage, “that is knowledge,” refers only to the second part of the preceding passage: when you don’t know something, acknowledge that you don’t know it. For example, the Qing dynasty scholar Mao Qiling states that “not knowing is indeed not knowing, but knowing one’s not knowing is knowing” (Mao 2009: Juan22). This interpretation was later accepted by Zhu Xi in his profoundly influential Collected Commentaries on the Analects 䄆䃎䳶⌘, and has since become the standard interpretation. This interpretation has an earlier origin. In the Guoliang Commentary on the Spring and Autumn ᱕⿻䉧ằۣ, Year 3 of Duke Yin 䳡‫ޜ‬, there is a saying that “to know that something is unknowable is knowledge/wisdom.” When quoting this, Li Dingzuo ᵾ唾⾊ in Tang Dynasting, in his Commentary on the Analects, adds that “to not seek to know what is unknowable is knowledge/wisdom” (Li 1989: 280). Sosa states that, “Although animal knowledge that we see a fire requires only that we be awake, and not that we know we are awake, reflective knowledge still does require knowledge that we are awake, not just dreaming” (Sosa 2009: 21). One question we can raise concerns the reference of “that”, which is reflective knowledge. In Sosa’s interpretation, it appears to refer to both (a) the knowledge or (recognition) that one has animal knowledge when one does and (b) the knowledge (recognition) that one doesn’t have such knowledge when one doesn’t. Now case (a) makes perfect sense, since reflective knowledge is the endorsement of animal knowledge, but it is not clear why case (b) is also reflective knowledge. As Sosa states, reflective knowledge includes, though going beyond, animal knowledge, but in case (b) there is no animal knowledge, even animal knowledge for one to deny, and so how can one have reflective knowledge that is not based on animal knowledge? Indeed, in this sense, case (b) is even different from the two meta-attitudes toward animal knowledge in addition to the attitude of endorsement: denial and suspension. Just like endorsement, these two attitudes are also based on animal knowledge. In the example of the animal knowledge that there is fire in front of me, I can endorse it so that I can have reflective knowledge that there is fire in front of me if I know, aptly, that I’m awake; I can reject it if I know, aptly, that I’m dreaming; and I can suspend my judgment if I don’t know, aptly, whether I’m awake or dreaming. Although in neither of the latter two cases there is reflective knowledge, there is still animal knowledge. However, in case (b) there is even no animal knowledge, and one, after reflection, only affirms that one has no animal knowledge. As we will see later in his response, Sosa may agree that certain trait virtues are necessarily helpful to achieving knowledge but still deny that they are constitutive of knowledge.

20 8

Ernest Sosa Encountering Chinese Philosophy In his response to this part, Sosa states the agency is also involved in functional representation, “even if this is the agency of a subconscious functional teleology.” Another way to see this issue, it seems to me, is that these sage-like agents can render explicit, deliberative judgment unnecessary precisely because, due to their life-long cultivation, they have internalized such judgment so that their performance appears to be functional and not deliberate.

References Cheng, Hao, and Cheng Yi. 1989. Collected Works of the Two Chengs Ҽ〻䳶. Beijing: Zhonghua Shuju. Guliang Commentary on the Spring and Autumn ᱕⿻䉧ằۣ. 1997. Shenyang: Liaoning Jiaoyu Chubanshe. Li, Dingzuo ᵾ唾⾊. 1989. Collected Interpretations of the Book of Chang ઘ᱃䳶䀓. Shanghai: Shanghai Guji Chubanshe. Mao, Qiling ∋ཷ喑. 2009. Corrections of Mistakes on the Four Books ഋᴨ᭩䥟. Beijing: Airusheng Digital Technology Research Center. Sosa, Ernest. 2007. A Virtue Epistemology: Apt Belief and Reflective Knowledge Vol. 1. Oxford and New York: Oxford University Press. Sosa, Ernest. 2009. Reflective Knowledge: Apt Belief and Reflective Knowledge Volume 2. Oxford and New York: Oxford University Press. Sosa, Ernest. 2015. “Confucius on Knowledge.” Dao 14: 325–330. Sosa, Ernest. 2015b. Judgment and Agency. Oxford and New York: Oxford University Press. Sosa, Ernest. 2017. Epistemology. Princeton: Princeton University Press. Sosa, Ernest. 2019. “Telic Virtue Epistemology.” In Heather Battaly (ed.), The Routledge Handbook of Virtue Epistemology. New York and London: Routledge. Sosa, Ernest. 2021. Epistemic Explanations: A Theory of Telic Normativity, and What It Explains. Oxford: Oxford University Press Xing, Bing ࡁᱪ. 1990. A Sub-commentary on He Yan’s օ᱿ Commentary on the Analects 䄆䃎䁫⮿. In Commentary and Subcommentary on the Analects and Commentary and Subcommentary on the Book of Filial Piety, Shanghai Guji Chubanshe. Zhu, Xi. 2002. Essential Interpretations of the Analects and the Mencius 䄆ᆏ㋮㗙. In The Complete Work of Master Zhu, volume 7. Shanghai: Shanghai Guji Chubanshe.

2

Xunzi, Zhuangzi, and Virtue Epistemology Kim-chong Chong

1 Introduction Sosa makes a central analogy between knowledge and archery. For the archer to be credited with an accurate shot, it must manifest “a competence seated in the agent,” in which case it has “the property of being apt” (Sosa 2011: 10). Similarly, “Knowledgeable belief aims at truth, and is accurate or correct if true. It has accordingly the induced aim of attaining that objective. Such belief aims therefore not just at accuracy (truth), but also at aptness (knowledge)” (Sosa 2011: 61). The archery analogy is applied differently in Confucius’s Analects (Behuniak 2010). For Confucius, knowledge has an aim, namely, to attain ren or humaneness. Achieving this depends on the agent’s stance or orientation. However, the target is not external to the agent. “Is ren really far away? No sooner do I desire it than it is here” (Analects 7.30). In other words, both the stance and aim reside in the agent. This gives an autonomous ethical sense in which there is “a competence seated in the agent.” Sosa would say this is not what is meant by knowledge in the Platonic tradition and from which his own work derives (Sosa 2011: 1). What is at issue is cognitive competence, not ethical. This relates also to the dispute in virtue epistemology between reliabilists and responsibilists. As a reliabilist Sosa argues that it is cognitive competence and apt performance which constitute knowledge. Virtues such as open-mindedness and intellectual courage may be a means to knowledge but not knowledge-constitutive. He says: It is such knowledge-constitutive competences that are of main interest to a Competence Virtue Epistemology aiming to explain human knowledge. Other epistemically important traits—such as open-mindedness, intellectual courage, persistence, and even single-minded obsessiveness—are indeed of interest to a broader epistemology. They are of course worthy of serious study. But they are not in the charmed inner circle for traditional epistemology. They are only “auxiliary” intellectual virtues, by contrast with the “constitutive” intellectual virtues of central interest to virtue reliabilism.” Sosa 2015: 42

21

22

Ernest Sosa Encountering Chinese Philosophy

In this chapter, I shall take up this issue of what constitutes knowledge in a somewhat roundabout way. I shall first describe the accounts of two Chinese philosophers, Xunzi and Zhuangzi, on perspective and knowledge. As we shall see, Xunzi can perhaps be said to have been concerned with knowledge in the sense that Sosa is concerned with. This would dispel the tendency to think that, following Confucius, others in the tradition necessarily treated knowledge in the ethical sense. Stephen Owen, for instance, has said that “The question that Confucius initiates is one of ‘recognizing’ the good in a particular case, rather than ‘knowing’ what the Good is. Chinese literary thought began its development around this question of knowledge, a special kind of knowing as in ‘knowing a person’ or ‘knowing the conditions of the age’ ” (Owen 1992: 20). This claim that Confucius “initiates” the question of ‘knowing’ in the Chinese literary tradition (including philosophy) as the special case of knowing a person may suggest that, thereafter, others too treated knowledge in the same way. Owen is not alone here. In discussing the term zhi (ᘇ), which means (giving oneself) an aim or direction, James Behuniak says with reference to the archery analogy that For Mencius, zhi is the all-encompassing ethical term. It simultaneously denotes the “position” from which aims are formulated and one’s ability to “discharge” them with conviction.1 As Mark Lewis explains, the single term zhi denotes the entire “moral conception of personality” for the early Confucians; it represents, in his words, the whole “thrust of a person’s being.” Behuniak 2010: 589

But Xunzi is an early Confucian who should not be lumped together with his predecessors, Confucius and Mencius, in how he handles the conception of knowledge. Traditionally, the disagreement between Xunzi and Mencius has been regarded as an ethical one. Xunzi is said to hold that human nature is bad, contradicting Mencius. But this is a mistaken reading (Chong 2003 and 2008). For Xunzi, human nature is neither good nor bad, in any essentialist sense. Instead, given certain human dispositions and limited resources, there is a tendency for disorder. Under these circumstances, through their cognitive and instrumental capacities, the sages were (cumulatively) able to discover and implement principles of ethical, social, and political order. The issue for Xunzi is cognitive and instrumental competence, not ethical. Xunzi holds that there are veridical perspectives when we exercise perceptual competence in the right external conditions. Zhuangzi, on the other hand, holds that perspectives are limited and calls for open-mindedness. In a sense, Zhuangzi’s account complements Xunzi’s. Following Sosa, it could be said that open-mindedness is a means toward or enables knowledge but is not knowledge-constitutive. However, Xunzi and Zhuangzi were not engaged with epistemology per se. Their discussions of knowledge took place within a social and political context. Seen in terms of the context, their positions were not complementary but opposed. When Xunzi argues that there are veridical perspectives he has in mind the discovery of social and political principles, which in the final analysis are Confucian principles. He was defending the Confucian social and political order. And Zhuangzi’s call for

Xunzi, Zhuangzi, and Virtue Epistemology

23

open-mindedness opposes these principles, which he sees as contrived and overly restrictive. If we ignore the context then we shall miss the perspectives in which they stand. In the conclusion I shall contend that within particular contexts the belief in open-mindedness and its corresponding attitude can be said to be knowledgeconstitutive. I shall conclude with a suggestion that what constitutes knowledge should be more broadly construed, taking into account the different contexts of knowledge. This would include the context in which Confucius talks of knowledge as ethical competence.

2 Perspective Let us first briefly characterize two conceptions of the term “perspective.” Consider the following story. There was a man who lost an ax and suspected his neighbor’s son. He observed that his walk, his expression, and his words were those of a man who would steal an ax. Every action and attitude of the youth betokened someone who would steal an ax. Later, when the man was clearing a ditch, he found his ax. When, on another day, he again saw his neighbor’s son, there was nothing in his actions or attitudes that resembled those of one who would steal an ax. It was not his neighbor’s son who had changed, but the man himself who had changed. This change had no other cause than that he had been prejudiced (you suo you ye ᴹᡰቔҏ). Knoblock and Riegel 2000: 286–287

Two conceptions of what a perspective is may be discerned. First, the object takes on an aspect according to how it is perceived and this in turn is a function of beliefs about it. The man’s belief that the youth stole his ax makes him a thief and this comes through his every action and expression. Under this conception, one’s belief could be a mistake. Realizing this, the perception that goes along with the belief also undergoes a change. Thus, a perspective is corrigible since it is a function of one’s beliefs. The second conception arises if we highlight what is stated at the end of the story, namely, that the object of perception can be different though physically the same. In other words, it need have no determinate description. It is one’s perspective that determines what the object is. Instead of “prejudiced,” there is another reading of the concluding phrase you suo you ye (ᴹᡰቔҏ). Following D. C. Lau, we may read this as the focus on something such that it is “confined” (Lau 1991: 16–18). Here, being confined means that there is some limit to one’s perspective. Let us refer to the first conception of perspective as “veridical.” Insofar as one’s perspective is based on a belief, it could be mistaken. This means that there is a correct or veridical perspective. And call the second “limited” since one’s perspective determines the nature of the object and is confined in that way. Under the veridical conception, the aim of the agent is to make correct judgments and decisions. To succeed, it is necessary to acquire accurate information. An implication is that the agent needs to ensure accuracy of beliefs about such information. In contrast, under

24

Ernest Sosa Encountering Chinese Philosophy

the limited conception, seeing something as something is a function of custom, upbringing, character, and so on.2 We now turn to the two texts, Xunzi and Zhuangzi,3 to see how the two conceptions are developed. A particular term, xin (ᗳ), will figure in the two accounts. This has been translated as either “heart” or “mind.” For the early Chinese philosophers generally, the xin has both cognitive and affective capacities. Thus, I shall translate xin as “heart-mind.” It should also be noted that, in the following discussion, “affective” is used in a broad sense, to cover not just feelings and emotions, but also desires, aversions, dispositions, and motivational attitudes.

3 Xunzi Xunzi (fl. 298–238 BC) can be said to have responded to Zhuangzi (bet. 399 and 295 BC). However, I discuss him first because some of Zhuangzi’s stories and metaphors can be understood by posing them against Xunzi’s more straightforward statements. In chapter 21 of the Xunzi, “Dispelling Blindness,” the word bi (㭭) describes a common fault, namely, to be “blinded” (alternatively, “blinkered” or “beclouded”) by a single aspect of a situation and thus fail to see it comprehensively. This might involve intellectual faults such as holding on to an extreme position or insisting on a particular doctrine. Alternatively, it could involve more personal faults such as partiality toward one’s own views and not being able to accept the praise of other views, or indulging private desires (such as for sex, power) at the expense of proper judgment on public issues. Xunzi does not differentiate between intellectual and other personal faults. In either case, what happens is that the cognitive capacities of the heart-mind fail to function properly. At one level, the heart-mind is described as authoritative and autonomous in commanding bodily movement and making judgments according to rational principles. Xunzi is impressed by its endless capacity for fresh input, its ability to make distinctions and to concentrate on specific purposes. At another level, however, it can be “unsettled” (bu ding нᇊ) (Li 1994: 495; Knoblock 1994: 21.8). An analogy is made with a pan of water. If the pan is upright and still, sediment will sink and the water will be clear, reflecting objects clearly. Similarly, the heart-mind’s ability to make distinctions can be affected if it is unsettled (Knoblock 1994: 21.7b). From the examples provided, it is clear that this occurs when the senses are affected by external conditions or impaired in some way. In other words, the heart-mind can function clearly only insofar as the senses are in good working order under appropriate conditions. Thus, poor lighting conditions affect vision and may lead one to identify objects wrongly. Alternatively, the senses may not be in good working order when a person is drunk (“A drunk may jump across a ditch a hundred paces wide, thinking it a drain half a pace wide, or may stoop down to go out the city gate, thinking it a small doorway”) or when a sense organ is interfered with in some way (for example, pressing against the eyeballs makes an object appear double) (Knoblock 1994: 21.8). This makes it seem as if the heart-mind is passive and functions clearly only if the senses are not adversely affected, just as water can reflect clearly without the interference

Xunzi, Zhuangzi, and Virtue Epistemology

25

of sediment. Although not mentioned directly, however, it is implicit that it has rational and logical capacities. Xunzi says: “As a general rule, when examining things about which there are doubts, if the heart-mind is not inwardly settled, then external things will not be clear. If my reflections are not clear, then I will never be able to settle what is so of a thing and what is not so of it” (Knoblock 1994: 21.8). For instance, it can adjust for appearances due to distance and height. Oxen appearing the size of sheep from a mountain top will not be mistaken for sheep; nor will trees on the mountain be taken for chopsticks when observed from below. Xunzi mentions a psychological state such as fright, citing someone who took his own shadow to be a ghost. Those who think there are ghosts are startled or confused. They “take what does not exist for what does and what does exist for what does not, and they settle the matter on the basis of their own experience” (Knoblock 1994: 21.8). In other words, subjective experience is inadequate for verification. Xunzi is arguing that veridical perspectives exist. These are cognized by a heart-mind that is thinking clearly, aided by senses in good working order under appropriate external conditions, and open to verification. We should mention one aspect of Xunzi’s position, namely, that an austere lifestyle is not required for the heart-mind to function clearly. Xunzi says that certain people who practice austerity lack “subtlety.” He mentions someone who lives in isolation because “if the desires of his eyes and ear were stimulated, then his thoughts would be shattered.” Against this, “The sage follows his desires and fulfills his emotions, but having regulated them, he accords with principles of order. Truly what need has he for strength of will, for endurance, or for keeping guard against unsteadiness?” (Knoblock 1994: 21.7d).

4 Zhuangzi Zhuangzi is pessimistic about the authority of the cognitive over the affective. This is apparent when he refers to the interminable debates of scholars who are described as fighting with their heart-minds. The doctrine that each puts forward is meshed with the scholar’s own partialities, impulses, worries, hopes, fears, and so on. Against Zhuangzi, it might be objected that partiality can be discounted. But he seems to anticipate this. Describing the various organs of the body, Zhuangzi asks: Which of them would I take as being intimately close? Would you be pleased with all of them? Or would there be a partiality here? Or would I just treat them all as my servants? Would these servants not adequately govern one another? Or would they take turns to be ruler and servant? Would there be a true ruler existing among them? Chen 1999: 52, my translation. Compare Watson 1968: 38

These remarks are intelligible if we pose them against Xunzi’s statement that the heartmind is “the lord of the body” and is “its own authority” such that it will accept or reject something according to whether it is right or wrong (Knoblock 1994: 105). Contrary to this, Zhuangzi is denying the sole authority and autonomy of the cognitive

26

Ernest Sosa Encountering Chinese Philosophy

capacity. Referring to the disputing scholars, Zhuangzi says that they have a “preestablished heart-mind” (cheng xin ᡀᗳ) (Chen 1999: 56, my translation).4 In other words, they already have established views. But what follows from this? Does it mean, for instance, that having a pre-established view precludes one from being open to another perspective and to realize that there are different possibilities? Consider the following story. Zhuangzi’s friend, Hui Shi, complains of a gourd which was too huge and unwieldy to serve as a water container. Zhuangzi tells him of a man who had bought the formula to a salve from silk bleachers. For generations, it had been used to protect the bleachers’ hands. This man sold the formula to troops of a certain state. With the aid of the salve, they won a naval battle in winter and the man was handsomely rewarded with a fiefdom. Zhuangzi suggests to Hui Shi: instead of being bothered by the gourd’s failure as a container, why doesn’t he float with it on the rivers and lakes? He concludes, “It would seem that you have a heart-mind that is clogged by weeds!” (Chen 1999: 30–31). Both the bleachers and Hui Shi are limited in their own ways. The former had used the salve for generations as a protectant with little profit. The latter is only able to think of the gourd’s utility as a container. The suggestion to float with it on the rivers and lakes indicates going beyond customary limits of thought. Thus, to answer the question raised above, it does not follow that having a pre-established view prevents one from being led to see that one’s view is limited and the possibility of changing that view as a result. But it might be argued that having a pre-established heart-mind means that affective interests are in operation in an unconscious way, such that it might seem arbitrary to deem the cognitive to be in control. However, it is unclear that Zhuangzi never allows for this. Stories of craftsmen and other characters with amazing skills suggest that the heart-mind can be focused on a cognitive and practical task. For instance, a cicada catcher describes training to the extent that he is “aware of nothing but cicada wings.” The fictitious character of Confucius comments that “He keeps his will undivided and concentrates his spirit” (Watson 1968: 200). In another case, an archer bets against opponents in a contest and becomes a nervous wreck when the bet is on precious gold (Watson 1968: 201). This might be thought to show that the cognitive has no control. However, the story ends by saying that one can succeed by ignoring “outside considerations.” Unlike Xunzi, Zhuangzi posits a variety of perspectives without asserting that any of them is the true or correct perspective. This is aided by the use of stories where an assertion or attitude is brought up short by showing that there exist other possibilities, as in the story of Hui Shi and the gourd. Other stories open our eyes to infinite possibilities. For example, “Lord of the River” takes pride in the vastness of his river during the autumn floods. Journeying eastward, he reaches the North Sea and realizes there is a vastness beyond. He meets “Ruo of the North Sea,” who tells him: There is no end to the weighing of things, no stop to time, no constancy to the division of lots, no fixed rule to beginning and end. Therefore great wisdom observes both far and near, and for that reason recognizes small without considering it paltry, recognizes large without considering it unwieldy, for it knows

Xunzi, Zhuangzi, and Virtue Epistemology

27

that there is no end to the weighing of things . . . . Looking at it this way, how do we know that the tip of a hair can be singled out as the measure of the smallest thing possible? Or how do we know that heaven and earth can fully encompass the dimensions of the largest thing possible? Watson 1968: 78

Zhuangzi’s conception of perspective as “limited” can be described in terms of being confined within a certain horizon. Under this conception, there is still room for the notion of a mistaken belief. We may be mistaken about something because there are limits to our perspectives on things. However, this is not because the horizon has an absolute outer limit which we have failed to capture. Instead, the metaphor of a horizon intimates not just a limit but at the same time an ever-broadening expanse and what lies beyond. Zhuangzi’s stories also intimate that a perspective involves a joint cognitive/affective orientation of an agent. Consider Zhuangzi’s own experience. Poaching for game in a park and aiming at a huge magpie, he observes it is about to pounce on a praying mantis, itself on the verge of snatching a cicada. Each is oblivious to what is behind it. Zhuangzi shudders and says to himself, “Ah!—things do nothing but make trouble for each other—one creature calling down disaster on another!” He turns to leave and suddenly finds the park keeper going for him. Zhuangzi is deeply shaken by the whole episode (Watson 1968: 218–219). David Nivison has given the following explanation: What he observed showed him that existence is a net of mutual trouble-making in which all creatures are caught, by their attachments; and his own experience of being rudely surprised by the game-keeper showed him that the philosopher’s conceit that he can distance himself by simply observing is itself a mode of involvement; you are in the world at all times yourself, and you can’t get out of it . . . . It was his philosophy of withdrawal itself that caused him to be so fascinated . . . that he forgot his own safety. That philosophy at that instant had “selfdestructed.” Withdrawal itself is an entanglement. Nivison 2000: 179–180

This change of perspective is a cognitive/affective experience. There is no “fact” of the matter which Zhuangzi cognizes that is independent of his personal orientation or reorientation. His earlier belief that he could withdraw from worldly entanglements is now seen to be mistaken. However, this “mistake” is not a fact that stands independently of his new personal orientation. We necessarily see something from a perspective. For jointly cognitive and affective reasons, too, our horizons may change and we may be led to see new features of situations and things. Another example shows how someone is disposed to see things in a certain way. On a stroll above the River Hao, Hui Shi challenges Zhuangzi’s assertion that the fish in the river are happy (Watson 1968: 188–189). After a brief exchange about whether and how one can know what another creature or person feels, Zhuangzi says, “I knew it from above the Hao.” It has been observed that what Zhuangzi is doing is getting Hui

28

Ernest Sosa Encountering Chinese Philosophy

Shi to re-orient himself to a perspective in which “they have been strolling on the bridge above the Hao river and sharing all along the experience of witnessing the fish swimming easily and smoothly.” And perhaps Zhuangzi is saying, “Do not exert yourself unnecessarily; let go of your personal interests and fears; as we witness the fish swimming easily and smoothly, we should be in the same state, be happy, and move easily and smoothly” (Teng 2006: 45). This is supported by Roger Ames’s observation about how “knowledge” applies here: “one’s posture or perspective is . . . integral to and constitutive of what is known, and contributes immediately to the quality of the experience” (Ames 1998: 220).

5 Comparisons with Sosa Although Zhuangzi was pessimistic about the cognitive being in charge, this does not seem to be ruled out, as seen in the stories about characters who have mastered certain crafts and skills (even though some would argue that they demonstrate “knowing how” instead of “knowing that”).5 And despite the strategy of using stories without truthaffirmation, Lord of the River’s admission of limitedness, for instance, is an admission of a cognitive error since he realizes that there is more to the universe than just his river. Similarly, Hui Shi is made to realize that there is more to a gourd than just being a container. These stories are in fact parables. They have a lesson, namely, to keep an open mind and not be shackled by customary practice and doctrine. As Ruo of the North Sea says: You can’t discuss the ocean with a well frog—he’s limited by the space he lives in. You can’t discuss ice with a summer insect—he’s bound to a single season. You can’t discuss the Way with a cramped scholar—he’s shackled by his doctrines. Now you have come out beyond your banks and borders and have seen the great sea— so you realize your own pettiness. From now on it will be possible to talk to you about the Great Principle. Watson 1968: 175–176

“Cramped scholar” is a translation of qu shi ᴢ༛. The word qu here means “corner” and a more literal translation is “scholar in a corner.” In his discussion of bi, being blinded or blinkered, Xunzi mentions bi yu yi qu 㭭ᯬаᴢ (Li 1994: 472). This means to be blinkered by a corner, or one aspect, of a situation. (The “Great Principle” refers to the “Way” or the dao, but more about this later). Given these similarities, it might be argued that Xunzi’s and Zhuangzi’s epistemic positions are not necessarily opposed and could be complementary. For instance, we should be careful not to overstate Zhuangzi’s position that, as Roger Ames puts it, “one’s posture or perspective is . . . integral to and constitutive of what is known” (see above). If this statement is taken generally without reference to particular examples, it would be tantamount to epistemic relativism, a position commonly attributed to Zhuangzi. In the example which gives rise to this statement, however, Zhuangzi is alluding to Hui Shi’s dispositional tendency to (what he sees as)

Xunzi, Zhuangzi, and Virtue Epistemology

29

tiresome logical questioning and telling him, instead, to relax, soak in the scene, and, in this sense, “know” that the fish are happy. What is true is that, for Zhuangzi, motivational attitude and dispositional character play a role in cognition, in the following way. For example, although Lord of the River’s realization of his limitedness may be an admission of a cognitive error, this is not just the discovery of a new or hidden fact. Instead, it is a re-orientation which makes it possible for Ruo of the North Sea to begin to have a discussion with him. This new orientation involves humble acceptance of his limitations, open-mindedness, and a willingness to learn from discussion. In an encyclopedic entry on “virtue epistemology,” John Turri and Ernest Sosa describe “virtue responsibilists” as follows: Responsibilists define epistemic virtues as praiseworthy and refined character traits with a distinctive motivational profile, such as conscientiousness and openmindedness, which underwrite robust, broad-based dispositions to inquire well across a wide range of circumstances. These epistemic traits share the same profile as the ethical traits featured in the Aristotelian tradition of virtue ethics, such as generosity, justice and compassion. Turri and Sosa, 2013: 1041

In the context of virtue epistemology, the joint operation of the cognitive and the affective could mean that open-mindedness and other character traits are epistemic virtues which have a necessary role to play in the pursuit of knowledge. Turri and Sosa (2013), however, refer us to the work of John Doris, who has argued against the reliability, or even the existence of character traits (Doris 2002). This is controversial. But more to the point, Sosa has argued that “we can best understand the responsibilist, character-based intellectual virtues highlighted by responsibilists as auxiliary to the virtues that are a special case of reliable-competence intellectual virtue” (Sosa 2015: 36). Furthermore, although virtues such as open-mindedness or intellectual courage may help one to reach the truth, they cannot be said to constitute knowledge (Sosa 2015: 40–41). The “virtue reliabilist” answer to what counts as an epistemic virtue is that it “includes the subject’s truth-conducively reliable doxastic dispositions” (Turri and Sosa, 2013: 1041). We have seen how Xunzi emphasizes veridical perception and reasoning about and examining things under the appropriate external conditions. And in accordance with this, some remarks of his cited earlier are consistent with discounting the role of certain character traits as necessarily helpful to achieving knowledge: “The sage follows his desires and fulfills his emotions, but having regulated them, he accords with principles of order. Truly what need has he for strength of will, for endurance, or for keeping guard against unsteadiness?” (see above). Sosa has argued that although, say, a healthy and disciplined regimen may help in the search for truth, sometimes it could be the other way around: “It might be that someone’s obsessive pursuit of truth, even at the cost of malnourishment and depression, puts them in a position to attain truths that are denied to their healthy rivals” (Sosa 2015: 41).

30

Ernest Sosa Encountering Chinese Philosophy

Xunzi’s description of how the senses could be distorted and how the agent could be misled by external conditions are akin to some of Sosa’s examples, such as when lighting conditions affect perception of the color of an object (Sosa 2011: 28). Sosa also distinguishes between first-order animal knowledge and reflective knowledge. “The latter, more demanding, level requires that the subject also believe aptly that his firstorder belief is apt, i.e., is one that manifests his competence” (Sosa 2011: 92). Xunzi stresses something which has an affinity to this and which he refers to as comprehensive thinking.6 We will see what he means by this below. Xunzi talks of the xin more as the “mind” than “heart-mind.” Not only is it “lord of the body” (Knoblock 1994: 21.6a) that issues commands, it also has the ability to critically examine things. This examination brings out the “patterns” or “principles” (li ⨶) of things. He says: “By laying out the warp and woof of Heaven and Earth, he (the sage) tailors the functions of the myriad things. By regulating and distinguishing according to the Great Ordering Principle, he encompasses everything in space and time” (Knoblock 1994: 21.5e). The “Great Ordering Principle” (da li བྷ⨶) refers to the fact that, overall, the universe is structured by principles that may be discovered through critical examination. For Xunzi, this constitutes the Way, or the dao, of the universe. He also lays out the following general principle about the relation between the faculty of knowing and what is known, at the same time stressing the need for a boundary (having parameters) in the search for knowledge: As a general principle, the faculty of knowing belongs to the inborn nature of man.7 That things are knowable is part of the natural principle of order of things. Men use their innate faculty of knowing, which allow things to be known. But if no boundary to the search is fixed, then even to the end of your life you will be incapable of knowing everything. Although you may make countless attempts to master the natural principles of order, in the end your effort will be insufficient to encompass the complete cycle of the transformation of the myriad things, and you and the fool will be as one. Knoblock 1994: 21.9

Being a sage involves focused thought and investigation so as to draw connections between things and to unify them in a principled way. In this regard, the sages were experts who articulated an ethical and social system: One who misses a single shot out of 100 does not deserve to be called an expert archer. One who travels a journey of 1,000 li, but does not take the last half-step does not deserve to be called a carriage driver. One who does not fully grasp the appropriate connection between modes of behavior and the various categories of things and who does not see the oneness between the requirements of the principle of humanity and the moral obligations that inhere in it does not deserve to be called an expert in learning. The truly learned are those who keep this unity. Those who leave with one principle and return with another are men of the streets and alleys. They are experts in a few things, but inexpert in many . . .. Be complete and whole in it, and then you will be truly learned. Knoblock 1994: 1.13

Xunzi, Zhuangzi, and Virtue Epistemology

31

Xunzi does not specify the conditions under which the archer missed that one shot out of a hundred. The judgment that he/she is not an expert seems overly stringent. But the point is that, for the expert, ability or competence is not a random affair. Instead, it is the result of learning and acting on the basis of a unified and interactive system of principles. The possession of comprehensive learning is not a particular skill but an ability to oversee things at a reflective level. Thus, someone with a particular skill may not necessarily be appointed “director” of the relevant trade or profession, for instance. On the other hand, another person without the first-order skill could be commissioned as director. This is similar to Sosa’s example of the statistician-coach-observer who manifests full aptness, in a way which the basketball player does not (Sosa 2015: 71). He will treat things in all their combinations as things. The noble person is one with the Way and uses it to further his testing of things. If he is at one with the Way, then he will be right; and if he uses it to further his testing of things, then he will be discerning. If using the right frame of mind and proceeding with discernment he deploys things in their proper positions, the myriad things will perform their natural functions. Knoblock 1994: 21.6b

Comprehensive thinking allows the sage to learn the rationale of the ritual principles so that “they will come to dwell within him,” to the extent that he would be averse to whatever is contrary to them and delight in whatever is in accordance with them (Knoblock 1994: 1.14). Contra Mencius, this is not due to an innate ethical capacity (Chong 2003). Instead, the ability or competence to discover and to implement ethical and social norms involves the comprehensive exercise of cognitive and instrumental capacities.

6 Conclusion In describing Xunzi’s and Zhuangzi’s discussions of knowledge, I have had Sosa’s account of virtue epistemology in mind and, as such, my description of how they are situated in relation to one another could be said to have been guided and structured by his account. Accordingly, the central conclusion seems to be that despite Zhuangzi’s “limited” conception of perspective, the epistemic virtue of open-mindedness in his account complements Xunzi’s “veridical” conception. Zhuangzi emphasizes character traits such as humbleness, open-mindedness, and willingness toward discussion, in the pursuit of knowledge. This complements Xunzi, who describes cognitive, instrumental thinking and non-distorted perception under appropriate external conditions as constituting knowledge. This parallels Sosa’s view that virtues such as open-mindedness and intellectual courage may be a means toward or enable knowledge but not knowledge-constitutive. However, Xunzi’s and Zhuangzi’s discussions of knowledge were embedded within a certain social and political context. Briefly, the context of Zhuangzi’s concept of

32

Ernest Sosa Encountering Chinese Philosophy

knowledge as encompassing an ever-expanding horizon of perspectives lay in his questioning the ethical, social, and political distinctions of his time, especially those endorsed by the pervasive Confucian ritual doctrines and practices which he thought were artificially imposed and overly restrictive. He opposed the Confucian justification of these doctrines and practices as endorsed by a moral tian or “Heaven” with his alternative conception of tian as “nature” or the spontaneous dao of the natural world with its wondrous diversity (Chong 2016). Xunzi agreed with Zhuangzi that tian is “nature” and not a moral “Heaven.” But for him, the dao is the “Great Ordering Principle.” In other words, the universe consists of patterns or principles which are open to discovery. The universe can be cut up at its joints, as it were, and there are patterns of social order which can be discerned and managed for the effective governance of human affairs. Although their accounts of knowledge may be said to be complementary from the view of Sosa’s virtue epistemology, the situation is more complex when we take into account the social and political context. For Zhuangzi, being open-minded is closely integrated with his criticism of Confucian monistic principles as contravening the diversity of nature, including human diversity. In response, Xunzi defended the Confucian order through his account of cognitive ability and of how nature contains principles that may be discovered and implemented for the sake of social order. Instead of being complementary, Xunzi and Zhuangzi stand on different ground. In at least one instance, Zhuangzi parodies the sense of knowledge as based on perception and intellect, calling for a “forgetting” instead: “I smash up my limbs and body, drive out perception and intellect, cast off form, do away with understanding, and make myself identical with the Great Thoroughfare. This is what I mean by sitting down and forgetting everything” (Watson 1968: 90).8 From the longer passage in which this occurs, it is clear that he has in mind the “forgetting” of Confucian values through casting aside the cognitive functions. Elsewhere, Zhuangzi says, “A true person comes before true knowledge” (Chen 1999: 178, my translation).9 In other words, certain attitudes and states of mind of the true person entitle him or her to be said to possess “true knowledge.” But how does this mention of context relate to Sosa’s project of virtue epistemology? I have in mind the broader context of Western epistemology. Here, I have to confess to being largely ignorant of the literature and debates and can only state my admittedly crude impression not of Sosa’s project per se, but of the style of Western epistemology in general since the latter part of the twentieth century. It seems to me that it has taken the concept of knowledge in a purportedly pure epistemic sense, to seek the necessary and sufficient conditions of knowledge, independently of context. At least this is the impression I get from the Gettier counterexamples to the traditional justified true belief model, the paradoxes that they have generated, and the seemingly endless attempts that have been made to resolve them. It is to Sosa’s credit that he attempts to go beyond these through his account of cognitive competence which discounts the role of luck in epistemic performance. His account, however, itself faces counterexamples such as the “fake barn” case: someone accurately perceives the only real barn in an area full of barn facades and thereby manifests his competence as a perceiver. His belief is apt and qualifies as knowledge. But this is counterintuitive since the case shows that the person does not have knowledge. Sosa’s own response has been “to explain away

Xunzi, Zhuangzi, and Virtue Epistemology

33

the counterintuitiveness of this result by emphasizing the lack of a further epistemically valuable state, which he calls ‘reflective knowledge’ ” (Ichikawa and Steup 2018). The examples (Gettier cases, fake barns, and so on) in these discussions are highly contrived. They have been generated for the sake of refuting a certain account and (to a bewildered outsider like myself) demonstrate the ingenuity of their inventors more than anything else. Thus, the style of discussion has been: a definition of knowledge is stated in terms of its necessary and sufficient conditions; doubts are raised about the adequacy of these conditions and counterexamples are given; this leads to responses that try to resolve the paradoxes generated by these counterexamples; others in turn question whether these responses are adequate; new theories are provided that try to go beyond these examples; these theories are in turn faced with further counterexamples; and so on. But it might be argued that Sosa’s virtue epistemological account does have a context, namely, that of apt cognitive performance. He provides the example of finding out the contents of a closed box, for instance. In doing this, certain competences may be exercised, perhaps even including character traits such as persistence and resourcefulness and lead reliably to finding out what is in the box (finding the truth about a state of affairs). Sosa argues that, nevertheless, “the exercise of such intellectual virtues need not and normally will not constitute knowledge, not even when that exercise does indirectly lead us to the truth.” On the other hand, when we manage to open the box, “we may immediately know” what is in the box, with a perceptual belief . . . which manifests certain cognitive competences for gaining visual experience and belief. Perhaps this complex, knowledge-constitutive competence first leads to things seeming perceptually a certain way, and eventually to the belief that things are indeed that way, absent contrary indications. A belief manifesting such a competence, and crucially, one whose correctness manifests such a competence, does constitute knowledge, at a minimum animal knowledge, perhaps even full-fledged knowledge (including a reflective component). Sosa 2015: 42

In talking about knowledge in terms of manifesting cognitive competence, Sosa has in mind activities such as: accurate perception, finding something, solving a puzzle, ascertaining that something is or is not the case, discovery, finding evidence, and so on. But although these are no doubt important activities of what makes up knowledge, they are not the only ones. I would still emphasize the point I have made earlier. That is, what constitutes knowledge and competence cannot be specified independently of a fuller account of context. Here, we should take a look again at what Zhuangzi says in one of the examples discussed. Lord of the River comes to see and to believe that there are wider vistas. This is not just the exercise of cognitive competence as Sosa has described it. As I have argued, this is not to be described as merely the admission of a cognitive error and the discovery of a new or hitherto hidden fact. Instead, there is a new orientation that involves a humble acceptance of personal limitations, open-mindedness, and a willingness to learn from discussion. Sosa would claim that these are just knowledge

34

Ernest Sosa Encountering Chinese Philosophy

enabling. But Lord of the River’s new orientation means that he possesses knowledge in terms of the virtues described. In this regard, there is a good sense in which these virtues are knowledge-constitutive, especially within the context of (what is seen as) Confucian narrowness and Confucian monism. And this alerts us also to the fact that when Xunzi talks of veridical perception and comprehensive thinking as constituting knowledge, he has a conception of social order in Confucian terms. In other words, Xunzi’s view of what constitutes knowledge is already loaded in Confucian terms. Zhuangzi would describe this as having a “preestablished heart-mind” (see the discussion of Zhuangzi above). In reply, it might be argued that, unfortunately, Xunzi’s view of veridical perception is corrupted by prior assumptions about the rightness of the Confucian order, but he is on the right track in his account of knowledge which (as we have shown) broadly parallels Sosa’s own. But this in turn raises an interesting question: Can there be a pure and “uncorrupted” account of knowledge which stands independently of context and the assumptions of that context? I would reply negatively, given the contexts of Xunzi’s and Zhuangzi’s accounts. In the example of finding out what is in a box, Sosa mentions certain virtues or traits. In remarks that I have already cited in the beginning of this chapter, he admits that these are “epistemically important traits,” which “are indeed of interest to a broader epistemology.” In addition, “They are of course worthy of serious study.” Nevertheless, Sosa claims that they are not knowledge-constitutive and do not belong to the “charmed inner circle” of traditional epistemology. In a handout responding to papers read at the conference on “Sosa Encountering Chinese Philosophy,” he coined the term “gnoseology” to delineate the area of this traditional epistemology “with its special focus on issues of philosophical skepticism.” I am reminded of the following remarks of Wittgenstein: “When philosophers use a word—‘knowledge’, ‘being’, ‘object’, ‘I’, ‘proposition’, ‘name’—and try to grasp the essence of the thing, one must always ask oneself: is the word ever actually used in this way in the language-game which is its original home?” (Philosophical Investigations 116). These remarks would remind us that there are different contexts of knowledge and accordingly different ways in which what is “knowledge-constitutive” may be construed. We began with a mention of how Confucius views knowledge in terms of ethical competence. I shall conclude with an example from Confucius. His disciple Yan Yuan had this to say of what he learned from his teaching: The more I look up at it the higher it appears. The more I bore into it the harder it becomes. I see it before me. Suddenly it is behind me. The Master is good at leading one on step by step. He broadens me with culture and brings me back to essentials by means of the rites. I cannot give up even if I wanted to, but, having done all I can, it seems to rise sheer above me and I have no way of going after it, however much I may want to. Analects 9.11

There are a few things to note about this passage. First, Yan Yuan’s statement refers to the essentials of the rites which cover a broad spectrum of rules governing individual ethical and social behavior. There is a close link here between knowledge and behavior.

Xunzi, Zhuangzi, and Virtue Epistemology

35

In this context, knowledge is not simply intellectual knowledge which is divorced from character and attitude. For right from the beginning of a person’s life, learning and knowledge are intimately linked with self-cultivation and ethical norms. This is affirmed by a second feature of the passage, namely, the reference to being broadened by culture. In this regard, the rites are closely integrated with music and literature. Thus, ritual training also involves study of music and literature with the aim of instilling sensitivity to the various nuances of individual and social behavior and the making of a balanced personality. Referring to a passage from the Odes (Shi or Shijing 䂙㏃)10 for example, Confucius refers to “joy without wantonness, and sorrow without self-injury” (Analects 3.20). This means too that the rites are not purely formalistic. Their ideal performances are imbued with the attitudes and values of aesthetic sensitivity, genuineness of feeling, grace, sincerity, earnestness, and so on. At the same time, one should guard against arrogance, presumptuousness, hubris, and mere display. Third, there is a reference to knowledge and learning as a process with no end-point. In other words, a cultured humanistic education gives one a sense of depth and subtlety of human problems in different dimensions. In this context, the competence that is exhibited or manifested in finding out or discovering pieces of information does not in itself prove, culminate in, or constitute knowledge. Instead, it is the whole orientation of the person in terms of the virtues described that matters and, as such, constitutes knowledge. It is this orientation—involving the belief and attitude that learning is a task with no end-point—that Yan Yuan embodies through his remarks and which shows his knowledge. What constitutes knowledge in this context cannot be stipulated independently of the attitude shown. Western epistemology, generally speaking, has assumed that there is a pure epistemic understanding of knowledge that can be described in terms of necessary and sufficient conditions. This has led, in Sosa’s case, to the insistence that cognitive competence (manifesting accuracy, adroitness and aptness) constitutes knowledge. There is no doubt that there are important contexts in which the manifestation of cognitive competence as Sosa has so intricately described constitutes knowledge. But these are not the only contexts of knowledge. I have tried to show that there are different contexts of knowledge and accordingly different ways in which what constitutes knowledge may be construed.

Notes 1

2

Mencius makes the following analogy: “Humaneness is like archery: an archer makes sure his stance is correct before letting fly the arrow, and if he fails to hit the mark, he does not hold it against his victor. He simply seeks the cause within himself ” (Mencius 2A:7). I have changed D. C. Lau’s “Benevolence” for ren to “Humaneness.” In a discussion of “seeing as,” Wittgenstein refers to a drawn triangle and considers that we could react to it as a number of things: “Could I say what a picture must be like to produce this effect? No. There are, for example, styles of painting which do not convey anything to me in this immediate way, but do to other people. I think custom and upbringing have a hand in this” (Philosophical Investigations 201).

36

Ernest Sosa Encountering Chinese Philosophy

3

Parts of the Zhuangzi may have been written by different people and it is not a thoroughly consistent text. However, for the sake of discussion and in comparison with Xunzi, I shall take the text as a whole and ascribe the views therein to Zhuangzi. 4 Watson translates this as “the mind given him” (Watson 1968: 38). 5 Following Sosa, it might also be said that these craftsmen possess intuitive, first order animal skills. Xunzi’s sages, on the other hand, possess comprehensive knowledge with reflective skills at the meta-level (Sosa 2015: ch. 3). 6 “Comprehensive” is Knoblock’s translation of the word jin ⴑ (Knoblock 1994: 21.9). Sometimes, another word quan ‫ ޘ‬or “complete” is used instead (Knoblock 1994: 1.13, 1.14). 7 Xunzi is not stressing that the faculty of knowing is the defining innate nature of humans. Instead, there is a faculty of knowing, ascribable to the nature of humans that enables the search for knowledge of the patterns of things. This is clear from the next two sentences. 8 “Great Thoroughfare” refers to the dao or the way of the universe. 9 Watson translates: “There must first be a True Man before there can be true knowledge” (Watson 1968: 77). For fuller discussion see Chong (2016): 68. 10 See Lau (1992) “Introduction” for an account of the role of the Odes in Confucius’s moral philosophy.

References Ames, Roger. 1998. “Knowing in the Zhuangzi: From Here, on the Bridge, Over the River Hao.” In Roger T. Ames (ed.), Wandering at Ease in the Zhuangzi. Albany: State University of New York Press. Behuniak Jr., James. 2010. “Hitting the Mark: Archery and Ethics in Early Confucianism.” Journal of Chinese Philosophy 37: 588–604. Chen, Gu-ying. 1999. Zhuangzi jin zhu jin yi 㦺ᆀӺ䁫Ӻ䆟. Taipei: Taiwan Shangwu. Chong, Kim-chong. 2003. “Xunzi’s Systematic Critique of Mencius.” Philosophy East and West 53: 215–233. Chong, Kim-chong. 2008. “Xunzi and the Essentialist Mode of Thinking on Human Nature.” Journal of Chinese Philosophy 35: 63–78. Chong, Kim-chong. 2016. Zhuangzi’s Critique of the Confucians: Blinded by the Human. Albany: State University of New York Press. Doris, John M. 2002. Lack of Character: Personality and Moral Behavior. Cambridge: Cambridge University Press. Ichikawa, Jonathan Jenkins and Steup, Matthias. 2018. “The Analysis of Knowledge.” In Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2018 edn), https://plato.stanford.edu/archives/sum2018/entries/knowledge-analysis/. Knoblock, John. Trans. 1994. Xunzi—A Translation and Study of the Complete Works. Stanford: Stanford University Press. Knoblock, John and Riegel, Jeffrey. Trans. 2000. The Annals of Lü Buwei. Stanford: Stanford University Press. Lau, D. C. 1991. “On the Expression Zai You ൘ᇕ.” In Henry Rosemont, Jr. (ed.), Chinese Texts and Philosophical Contexts. Illinois: Open Court. Lau, D. C. 1992. Trans. Confucius: The Analects. Hong Kong: The Chinese University Press. Lau, D. C. 2003. Trans. Mencius. Hong Kong: The Chinese University Press.

Xunzi, Zhuangzi, and Virtue Epistemology

37

Li, Disheng. 1994. Xunzi Jishi 㥰ᆀ䳶䟻. Taipei: Xuesheng Shuju. Nivison, David. 2000. “Xunzi and Zhuangzi.” In T. C. Kline and P. J. Ivanhoe (eds), Virtue, Nature, and Moral Agency in the Xunzi. Indianapolis: Hackett Publishing Company. Owen, Stephen. 1992. Readings in Chinese Literary Thought. Cambridge, Massachusetts: Council on East Asian Studies, Harvard University. Sosa, Ernest. 2011. Knowing Full Well. Princeton: Princeton University Press. Sosa, Ernest. 2015. Judgment and Agency. New York: Oxford University Press. Teng, Norman. 2006. “The Relatively Happy Fish Revisited.” Asian Philosophy 16: 39–47. Turri, John and Sosa, Ernest. 2013. “Virtue Epistemology.” In Byron Kaldis (ed.), Encyclopedia of Philosophy and the Social Sciences. London: Sage. Watson, Burton. 1968. Trans. The Complete Works of Chuang Tzu. New York: Columbia University Press. Wittgenstein, Ludwig. 1968. Philosophical Investigations. Translated by G. E. M. Anscombe. Oxford: Basil Blackwell.

38

3

Epistemic Competence and Agency in Sosa and Xúnzˇı Chris Fraser

1 Introduction Knowledge is an achievement manifesting a type of competence, akin in important respects to a skill. Accordingly, epistemic judgment is an exercise of agency. Ernest Sosa’s work has elaborated these and related insights into a meticulous, persuasive version of a virtue epistemology. Given the framing assumptions of mid-twentieth century Anglo-American epistemology, developing a competence-centered explanation of judgment, knowledge, and justification required brilliant critical and creative thought. So it is intriguing and perhaps instructive to consider how some of Sosa’s views relate to the outlook of early Chinese thinkers, for whom the idea of knowledge as a competent performance required no argument, being implicitly taken as an obvious, shared starting point. Here I will focus on Xúnzˇı 㥰ᆀ, whose epistemological concerns in some respects dovetail with, and in others complement, Sosa’s.1 I will draw on concepts from Sosa’s framework to elucidate features of Xúnzˇı ’s epistemology and in turn suggest how Xúnzˇı ’s theoretical orientation might cast light on Sosa’s project. In particular, I will suggest that Sosa’s conception of full aptness helps to elucidate the significance of Xúnzˇı ’s discussion of epistemic pitfalls, while Xúnzˇı ’s treatment of the epistemic agent’s awareness of and commitment to norms of judgment helps to enrich Sosa’s view of epistemic agency. This comparative exploration has interesting implications for the long-standing divide between virtue reliabilism and virtue responsibilism in contemporary discussions of virtue epistemology.2 A common way of characterizing the divide is that the reliabilist approach identifies intellectual virtues with truth-conducive faculties such as perception and memory, whereas the responsibilist approach identifies them with character traits such as epistemic conscientiousness, perseverance, and openmindedness. Reliabilists focus more on the project of analyzing or explaining knowledge in terms of epistemic virtues, while responsibilists broaden the focus to explore what we might think of as the epistemic good life. On a Xunzian conception of epistemic agency, the ability to apply truth-conducive faculties in the “full” manner Sosa associates with human knowledge, rather than mere “animal” knowledge, is 39

40

Ernest Sosa Encountering Chinese Philosophy

grounded in normative commitments characteristic of responsibilist virtues such as intellectual conscientiousness, diligence, or perseverance. Xúnzˇı ’s approach thus suggests that reliabilist and responsibilist virtues may be constitutively intertwined.

2 Sosa on epistemic competence and agency Judgment and knowledge are special cases of intentional action, suggests Sosa, and hence their nature can be clarified through a normative structure that applies generally to performances of all kinds. Epistemology is in effect a special case in which we apply this normative structure to a domain of epistemic performance. Sosa proposes a framework in which any attempt to carry out a performance to attain some aim can be informatively evaluated along three dimensions. We can examine whether the attempt is successful, competent, and apt. A performance is successful iff it achieves its aim. It is competent iff it issues from the agent’s competence in that type of activity. It is apt iff it is successful because competent. Consider Sosa’s canonical illustration of an archer attempting to hit a target. The archer’s shot is successful iff it hits the target. It is competent iff performed in a way that usually would be successful.3 It is apt iff its success is due to the competence with which the shot was made—and not, for example, a lucky accident in which a sudden gust blows the arrow off course only for it to ricochet against a wall and hit the target anyway. Sosa distinguishes between animal, reflective, and full aptness, allowing us to recognize different dimensions and levels of competence. Animal aptness refers to the first-order aptness manifested in performances that are successful because competent, as when a shot hits the target because of the archer’s skill. Reflective aptness refers a further, reflective dimension of competence extending beyond animal aptness. Normally a competent archer not only attempts shots but while doing so attempts to assess the chance of making an apt shot given the conditions—the distance and size of the target, the light, the wind, whether the archer is fresh or fatigued, and so on. In making such assessments, the archer aims at an apt second-order awareness of the likelihood of achieving a first-order, animally apt shot.4 A reflectively apt shot, then, is one that is animally apt and is attempted while the archer is aptly aware that the shot will be animally apt. Reflective aptness requires only the conjunction of first-order, animal aptness and apt second-order risk assessment. Full aptness is achieved through a causal connection between these. A performance is fully apt iff “it is guided to aptness through the agent’s reflectively apt risk assessment” that the performance would indeed be apt (Sosa 2015: 69). Sosa proposes that full aptness is a normatively desirable status for performances in general (2015: 85). Sosa’s notion of full aptness is in effect an attempt to explain how full-fledged competence requires not only that one’s first-order performances be reliably successful but that this success reflect a second-order grasp of the extent of one’s own first-order reliability. This higher-order understanding works to increase reliability by reducing or eliminating failed attempts. It also removes the element of luck from what Sosa calls a

Epistemic Competence and Agency in Sosa and Xúnzıˇ

41

“creditable” performance. For if our first-order performances are not guided by an apt second-order understanding of our own first-order competence, we could all too easily make inapt attempts (Sosa 2015: 72). In that case, even when our attempts do succeed, they might as easily not have, and so our successes are partly a matter of luck. This framework for evaluating performances applies as well to judgment and belief, for these can be regarded as epistemic performances that constitutively aim at truth (Sosa 2016: 11). Judgment is an exercise of epistemic agency in which we voluntarily attempt to represent aptly how things are. To judge that p is to affirm that p in an endeavor to affirm aptly that p (Sosa 2015: 80). So judgment is fully apt when it produces an apt representation that is such partly because it was guided by the agent’s apt self-assessment of competence. Human knowledge (as contrasted with “animal” knowledge) is then fully apt belief or, as Sosa also calls it, “knowing full well” (2015: 85). Unlike judgments, beliefs are often not the result of voluntary attempts to represent how things are. Many beliefs seem to be states that simply occur in us as a result of the functioning of perceptual systems, for example. My belief that it is daytime right now is an attitude I just have, by virtue of being awake, not the result of an intentional performance. Acknowledging this point, Sosa suggests that we can explain such epistemic attitudes also as the result of performances, but implicit performances that are not intentional or voluntary. He draws an analogy to the functional teleology of an organ such as the heart. The heart’s pumping can be regarded as a performance aimed at circulating blood (Sosa 2016: 5). Like intentional attempts, such non-intentional, functional aimings can succeed or fail. So functional capacities such as perception can be described as engaged in implicit performances that have a constitutive aim, such as true representation (Sosa 2015: 19, 51, and 87). Perception provides us with “seemings,” representational states that function as “attractions” to represent that such and such (Sosa 2015: 93). These seemings occur below the level of agentive control but nevertheless aim at correct representation (Sosa 2015: 93). We can think of them as inclinations to believe. When the inclinations are relatively strong, they amount to implicit, functional belief (Sosa 2015: 51), a distinct type of belief from the explicit, intentional, or judgmental beliefs that result from judgment.5 When the strength of these inclinations passes a certain threshold, the agent acquires the disposition to affirm a judgment and thus comes to hold a judgmental belief (Sosa 2015: 92). Like judgmental beliefs, functional beliefs can be assessed as to their success, competence, and aptness.6 However, they are not subject to the same sort of deontic evaluation as judgmental beliefs concerning what one may or ought to believe, as such evaluation pertains only to intentional endeavors (Sosa 2015: 193). This is because functional beliefs are not “endeavors,” which derive from freely determined choices and judgments (Sosa 2015: 192). Sosa’s account of the relation between functional and judgmental beliefs raises several issues that call into question aspects of his explanation of epistemic agency. The crux of agency for Sosa seems to lie in the capacity to freely determine one’s aims or choices (2015: 192). It thus contrasts with events we suffer passively or movements we make purely by reflex. He recognizes two sorts of agency: agency simpliciter, in which we freely choose our endeavors, and an intermediate form of

42

Ernest Sosa Encountering Chinese Philosophy

agency in which we do not freely determine our performances, which issue from implicit, passive functioning, but the outcome of these performances is nevertheless rationally derived from freely chosen endeavors and so also subject to a kind of rational evaluation (2015: 193–194).7 This characterization of the two varieties of agency makes it seem that epistemic agency for Sosa lies mainly in arriving at beliefs by explicit, voluntary affirmation through reflectively self-conscious judgment. He seems to see the crux of epistemic agency as our “free choice to judge affirmatively” on some question “provided we are aiming to get it right” (2015: 210). If this is indeed his stance, I suggest that this model is too intellectualist. Judgments that we make an explicit, free choice to affirm seem rare in practice and are probably implicated in only a small portion of our epistemic attitudes.8 Hence they do not seem especially central to understanding epistemic agency. Sosa recognizes that exercises of epistemic agency can be subconscious (2016: 10), and that performances can be apt, and agentive, when they spring from implicit, automatic reactions or involuntary operations that agents have trained themselves to perform, as when a tennis player automatically volleys a ball or we automatically remember a phone number we worked to memorize (Sosa 2015: 94–95). Beyond this, I suggest that even when we do explicitly consider evidence and form a belief through self-conscious deliberation, often we do not experience the process of reaching a judgment as free or voluntary.9 To be sure, sometimes we consider the evidence and self-consciously decide to change our belief or withhold judgment. But often, as we ponder the evidence, a judgment just comes to us through implicit functioning, much as the belief that it is day or night right now just comes to us. I suggest that, especially for highly competent agents such as professional athletes and performing artists, the distinction between a voluntary endeavor and an automatic, functional response seems neither as sharp nor as significant in explaining agency as Sosa takes it to be. He contrasts “animal belief,” which he describes as a “constituted by a stored state that guides conduct subconsciously,” with “reflective, judgmental belief,” which he sees as “a disposition to judge affirmatively in answer to a question, in the endeavor to answer correctly, with truth, reliably enough or even aptly” (2015: 209). But both descriptions seem to refer equally to dispositions that guide action in various ways, sometimes self-consciously, sometimes not. Moreover, as we saw, when functionally produced “seemings” cross a certain strength threshold, the dispositions that constitute functional beliefs can instead be considered to constitute judgmental beliefs (Sosa 2015: 92), further blurring the line between the two. Sosa’s approach as developed in Judgment and Agency, then, does not provide a sufficiently informative account of just what agency, and specifically epistemic agency, consists in. Elsewhere he provides hints of the direction in which he might develop such an account. He notes that beliefs sometimes seem not to be voluntary. For example, “I seem not now to be free at will not to believe that I am awake” (Sosa 2016: 10). Nevertheless, he suggests, “what voluntary freedom requires is only the ability to override improper influences” (2016: 10). That is, the sort of freedom required for epistemic agency need not entail that we be able to affirm or disaffirm any belief at all, but that we be free to correct and improve our beliefs “so as to align our action with the requirements of reason” (2016: 10). (Such correction, I suggest, might occur either

Epistemic Competence and Agency in Sosa and Xúnzıˇ

43

through explicit judgment or through guidance by implicit, functional dispositions.) Beyond this, Sosa proposes, epistemic agents are indeed free to believe that p or not insofar as they are free to decide “whether to address the relevant “whether-p” question at all” (2016: 11). Having committed to some end, agents are not free to decide at will what is an effective means to that end. Analogously, having committed to aptly answering whether-p, we are not free to affirm that not-p when faced with strong evidence that p. But we do remain free to adopt or reject our ends, such as the end of answering whether-p. Here Sosa’s approach takes an intriguing turn toward Xúnzˇı ’s, for Xúnzˇı ’s implicit conception of epistemic agency emphasizes self-improvement and presents commitment to an appropriate set of ends or norms as the fundamental attitude of the conscientious epistemic agent.

3 Xúnzˇı Like other early Chinese theorists, Xúnzˇı explains knowledge in terms of competence in discriminating and naming things, specifically for the purpose of guiding action and carrying out the dào 䚃 (way). The capacity for or faculty of knowledge he calls “the knowing” (zhī ⸕) (Hung 1966: 22/5).10 Knowledge is demonstrated through the performance of discriminating correctly, with respect to some “name,” what is shì ᱟ (this, right) from what is fēi 䶎 (not-this, wrong) (2/12), such that the attitudes of “the knowing”“match” (hé ਸ) the distinctions between things (22/5). In Xúnzˇı ’s framework, the functional counterpart to belief is the attitude of deeming something the kind of thing designated by some name. The counterpart to judging is distinguishing or discriminating something as properly taking some name. The use of “names”—mainly general terms referring to kinds—rests on our ability to distinguish shì from fēi, two terms referring what is or is not relevantly similar and thus falls within the extension of the same name. The purpose of names is to distinguish different social ranks and similar from different things, so that intentions can be conveyed and tasks carried out (22/14–15). Names can be used for this purpose because, as creatures of the same kind, with the same sort of constitution, we have sense organs that detect things similarly. This enables us to establish conventions for the use of names by which to discriminate distinct kinds of relevantly similar things (22/16–17). Similarities and differences between features of things are differentiated by means of the sense organs and recognized by the heart (xīn ᗳ, “heart/mind”)—the organ of cognition—through a capacity called “the verifying knowing” (zhēng zhī ᗥ⸕) (22/19). The sense organs “record” the features of things—sounds, shapes, and so on—and the heart “verifies”, or recognizes, them (22/20). To qualify as having perceptual knowledge of something, an agent’s sense organs must “record” it, such that the agent is aware of it, and the heart must “verify” it, such that the agent is able to “explain” (shuō 䃚) it (22/20–21). Presumably, the point is that to count as knowing a thing, the agent must demonstrate competence by applying an appropriate name to it.11 Xúnzˇı and other early Chinese theorists do not specify a justification requirement for knowing. Knowledge is simply success in getting the distinctions right. Lucky

44

Ernest Sosa Encountering Chinese Philosophy

guesses are excluded because knowledge is implicitly associated with systematic competence across a variety of interrelated and contrasting cases. The knowing of a highly competent agent, for instance, is said to connect together a unified system of kind distinctions (8/122) or enable discourse on a myriad cases while demonstrating mastery of such a system of kinds (23/78–79). Arguably, Xúnzˇı ’s conception of knowledge (zhì Ც) here converges with what we might think of as understanding or wisdom. It involves competence not merely in recognizing things but in grasping the “patterns” (lˇı ⨶) by which they are organized, including how they normally function and how they relate to each other. As we will see, Xúnzˇı ’s concept of “patterns” is pivotal to his conception of epistemic excellence. A revealing feature of Xúnzˇı ’s approach to knowledge is just how little he says about it. He takes it as obvious that knowledge is a competence in drawing distinctions between the kinds of similar things that take various names—so much so that the subject merits little discussion. His interest as an epistemologist is mainly not in the theoretical project of explaining knowledge but in the practical project of correcting and improving our distinction-drawing competence. The central topic of his most sustained epistemological discussion—“Resolving Obscuration” 䀓㭭—is not knowledge, per se, but error. In particular, it is how to use our cognitive and reflective capacities to avoid or mitigate what he calls bì 㭭 (obscuring, blinkering), his catch-all term for conditions that can obstruct correct distinction-drawing. Xúnzˇı ’s central claim here is that difficulties in attaining and applying knowledge arise from being fixated on and thus blinkered by only part of the relevant factors or circumstances, such that we are unclear about the broader patterns (lˇı ⨶) of how things work. All troubles that people have are due to their being blinkered by one bend, putting them in the dark as to the greater patterns (lˇı ⨶). (21/1)

The major issues that capture Xúnzˇı ’s attention are how to avoid “blinkering” or falling victim to “obscuration” by managing the operations of the heart (xīn ᗳ)—the seat of cognitive, affective, and conative functioning—and how to seek clarity as to the broader patterns of things by committing to the right dào (way) and thus adopting appropriate norms of judgment.12 Although he does not use this nomenclature, Xúnzˇı ’s discussion of blinkering presents what amounts to a distinctive view of conscientious, competent epistemic agency. He regards epistemic activity as a field of practical skill, which he calls “arts of the heart” or “heart techniques” (xīn shù ᗳ㺃). He discusses the performance of the gentleman or the sage, an epistemic agent who is not merely competent but expert. Such an agent seeks to excel in the “arts of the heart,” much as expert athletes or performing artists do in their endeavors. Here I want to call attention to three prominent features of Xúnzˇı ’s epistemology that reflect his distinctive view of epistemic competence and agency and complement or contrast with Sosa’s approach. The first feature converges with Sosa’s notion of full aptness, and indeed Sosa’s ideas help to elucidate the significance of Xúnzˇı ’s views.

Epistemic Competence and Agency in Sosa and Xúnzıˇ

45

Xúnzˇı contends that, in unclear circumstances, when in observation there are grounds for doubt or one’s heart is unsettled, the competent epistemic agent withholds judgment. Whenever in observing things there is doubt or one’s heart within is not settled, then external things are unclear. Our thinking being unclear, we can’t yet fix “so” or “not-so.” (21/67–68)

His examples of unclear circumstances include perceptual illusion, intoxication, unreliable means, and incompetent testimony (21/68–73). Someone walking in the dark might mistake a horizontal boulder for a crouching tiger or a small tree for a person, for example, because “the darkness obscures their vision.” A drunk will stoop while exiting the city gate, taking it to be a low doorway, because “the alcohol disrupts his spirit” (21/70). We do not judge how attractive we look by our reflection in moving water, because “the water’s position is disturbed” (21/72). Nor do we determine whether there are stars in the sky by asking the blind, because their “functional proficiency is confused” (21/73). Xúnzˇı ’s key claim about these examples is that competent agents are not misled in such dubious circumstances, because they attend to the “greater patterns” and so are not blinkered by the strictly partial resemblance between the objects they are observing and the reference objects they may seem similar to. So from a mountaintop looking down at oxen, they are similar to sheep, but someone seeking sheep does not go down to lead them away; the distance obscures their size. From the foot of a mountain looking up at trees, ten-meter trees are similar to chopsticks, but someone seeking chopsticks does not go up to break them off; the height obscures their length. (21/71–72)

In settling whether things are “so” or not, the competent agent takes the broader context into account and avoids “using the doubtful to resolve the doubtful” (21/74), or relying on unreliable means to settle questions whose answer is unclear. Xúnzˇı ’s treatment of these examples converges with Sosa’s view that one dimension of an epistemic agent’s competence is competence in assessing the reliability of one’s own cognitive operations in various contexts and in guiding one’s epistemic attitudes accordingly.13 When this self-reflexive competence guides the agent to competently adopt correct epistemic attitudes, the result will be an epistemic status corresponding to Sosa’s notion of full aptness. Obviously, Xúnzˇı ’s theoretical apparatus is different from Sosa’s, and he does not employ an explicit conception of aptness, let alone full aptness. He is not proposing an account of knowledge as fully apt belief. Nor does he explicitly conceptualize our grasp of our own reliability, given the conditions, as a second-order risk assessment concerning our own first-order performance. He frames it instead as an awareness of how performance must be adjusted to circumstances, in particular how awareness of the broader context and our own psycho-physiological states can prevent us from being blinkered, and thus misled, by a partial, limited view of our situation.14 Clearly, however, Xúnzˇı does share with Sosa the stance that the

46

Ernest Sosa Encountering Chinese Philosophy

performance of a competent epistemic agent will be guided by a meta-awareness of how reliable the agent’s normal means of generating epistemic attitudes will be in the situation at hand, and this meta-awareness includes a background competence in assessing and coping with difficult or abnormal conditions. Epistemic competence is not simply a matter of distinguishing things correctly in simple cases—just as competence in any field is not demonstrated by handling only trivial cases—because an agent who can manage only simple cases might easily become confused in more complex conditions. To demonstrate genuine competence, the agent must be able to avoid “blinkering” or “obscuration,” even in challenging circumstances, by drawing on a higher-order competence in assessing and avoiding problems. The outcome is an epistemic status corresponding to Sosa’s notion of full aptness. The second feature I want to underscore makes it clear that what Xúnzˇı is describing here is indeed a second-order meta-awareness of one’s own cognitive operations and how they relate to circumstances and broader norms. For Xúnzˇı , a characteristic feature of agents with expert epistemic competence is a second-order concern to improve their competence. Arguably, this stance is a direct consequence of a performance approach to epistemology, which invites parallels between epistemic endeavors and, for example, athletic endeavors. Any athlete is concerned to play well, and agents who aim at success in sport typically devote attention to improving their performance. Analogously, on Xúnzˇı ’s view, an aspect of epistemic competence is that the agent actively manages the operations of the heart so as to perform well epistemically. The agent seeks to improve both explicit judgment (which for Xúnzˇı follows from deliberation, lü` ឞ) and immediate, implicit distinctiondrawing.15 To improve their competence, agents employ second-order knowledge of the difficulties that can arise in epistemic performance. The sage—the expert agent— understands the problems that arise in the “arts of the heart” (21/28). Such an agent prevents error by conscientiously employing the heart to recognize and avoid various sources of bias or misjudgment. This conscientiousness is the basis for the self-reflexive, critical awareness that enables competent agents to avoid epistemic pitfalls. Exemplary agents “take charge of their hearts and manage them carefully” (21/11) so as to avoid blinkering and error. For “if the heart is not employed at it,” we can fail to recognize even the obvious: “though black and white be right in front of us, the eyes fail to see them; though thunder-drums be right beside us, the ears fail to hear them” (21/4–5). To perform properly, the heart must maintain a calm, attentive equilibrium, just as in sports or the performing arts. A key to proper performance is to prevent the heart from losing balance, leading to bias, blinkering, and one-sided, erroneous performance. Thus the human heart can be compared to a pan of water. Place it upright and do not move it, and the sediment settles to the bottom and the clear water rises to the top. Then it is sufficient to see your beard and eyebrows and discern the fine patterns on your face. A breeze passing over it, the sediment moves below and the clear water is disturbed on top, and you cannot get even the general outline right. The heart too is like this. So guide it with proper patterns, cultivate it with clarity, and let nothing bias it. Then it will be sufficient to fix shì-fēi and settle doubts. If

Epistemic Competence and Agency in Sosa and Xúnzıˇ

47

minor things pull it about, then externally one’s uprightness will be altered and internally the heart will be biased, and it will be insufficient to decide even gross patterns. (21/54–58)

As this passage indicates, Xúnzˇı holds that proper “guiding” and “cultivating” in the arts of the heart can train us to avoid error. To discriminate shì-fēi and resolve confusing circumstances reliably, we can train ourselves to maintain an impartial, upright stance and an undisturbed, unbiased heart. For Xúnzˇı , then, fully competent epistemic agency requires that we actively take control of our epistemic performance and train ourselves to improve it. How exactly do we do so? Xúnzˇı holds that differences between things are the basis for drawing distinctions and thus for cognition. At the same time, he explains, differences are a potential source of obscuration, disrupting our ability to “sort” or “grade” (lún ٛ) things properly (21/29). Things can be similar or different in various ways, with respect to various features. One-sidedly or injudiciously attending to any one feature may lead us to overlook or discount others that may also be pertinent. In attending to what is desirable about something, for example, we may overlook what is detestable about it, while in attending to what is beneficial about it, we may overlook what is harmful (3/45–49). More broadly, our recognition of “that one”—one side of a distinction—can interfere with our understanding of “this one”—whatever falls on the other side (21/38). To avoid bias, then, we need neutral, reliable criteria. We need a set of norms that specify the relevant distinctions and guide us in drawing them and thus forming correct epistemic attitudes. The question of criteria takes us to the heart of Xúnzˇı ’s approach to epistemic agency and the third main feature I want to highlight. For Xúnzˇı , the most fundamental aspect of epistemic agency is our higher-order capacity to approve and commit to a system of norms by which we check, correct, and improve our epistemic attitudes. Xúnzˇı contends that, acting from a second-order awareness of the sorts of problems that can arise in the “arts of the heart,” the expert agent guides epistemic attitudes by reference to norms the agent explicitly identifies, endorses, and seeks to maintain (21/29–34). Setting aside potential grounds for bias, the sagely agent “all-inclusively lays out the myriad things and weighs them on the scale” (21/29). The “scale,” says Xúnzˇı , is the dào 䚃 (21/29, 22/74), a normative “way” of conduct. An expert in the dào avoids the biases produced by attending to only one side of a distinction or one part of a scene by approaching things comprehensively, in terms of their relation to everything else, and thus grasping the “greater patterns” (21/51–52). According to Xúnzˇı , since such an agent is focused fully on the dào, rather than on particular, partial interests or features, he is correct; since he uses the dào as a basis for examining things, he is discerning in how he discriminates them into kinds; and thus, since he uses correct intentions to proceed with discerning sorting of kinds, the myriad things find their proper place (21/52–53). Before we can apply the dào in this way, however, our heart must recognize (zhī ⸕, “know”) and “approve” (kě ਟ) it (21/32). “Approval” is a normative attitude; to approve something is to deem it worthwhile and permissible. Elsewhere, in discussing moral

48

Ernest Sosa Encountering Chinese Philosophy

development, Xúnzˇı takes the heart’s approving things to be the central attitude that drives human agency, what most fundamentally controls our conduct (22/55–63). His stance on epistemic agency is thus a special case of his more general view of agency. In both contexts, the pivotal capacity that marks us as agents is the capacity to form and act on normative, action-guiding attitudes of approval or endorsement, whether explicit or implicit. In the case of epistemic agency, the relevant attitude is one of approving the norms by which to evaluate epistemic attitudes as correct or not. These norms then serve as the basis for checking and improving our epistemic competence.

4 Reflections for further inquiry For Xúnzˇı , agency—including epistemic agency—lies most fundamentally in our capacity to commit to a normative dào and apply it to correct and improve both our particular performances and our general competence. Xúnzˇı would thus agree with Sosa that the ability to override improper influences and to decide what “whether-p” questions to address is crucial to epistemic agency. Even more crucial for him, however, is the capacity to recognize, approve, and thereby aim to meet, norms of correctness. Xúnzˇı ’s epistemology explicitly presents this latter capacity as pivotal to expert epistemic competence. Sosa does not, but arguably his framework incorporates a parallel idea. For Sosa, a constitutive aim of epistemic performances is truth, which provides standards of correct performance. Epistemic agency, then, entails a higherorder concern to aim at truth. There are two key concepts here, truth and aiming. Might Xúnzˇı ’s epistemology enrich our understanding of them? For Xúnzˇı , the norms that the epistemic agent approves and commits to are conceptualized not as truth but as dào—a way or path of competent conduct that we aim to follow. In his theoretical scheme, dào includes the norms that determine the correct use of words, thus fixing semantic content and in turn the content of epistemic attitudes. Hence whether some deeming or utterance is true—Xúnzˇı would say whether it is “so” (rán ❦)—is ultimately determined by dào. Truth is a byproduct of and explained by dào—specifically, the dào of distinguishing different kinds of things. The upshot is that epistemic norms ultimately rest on a subset of norms of conduct and indeed on a normative conception of the proper way of life. Xúnzˇı ’s epistemology reminds us of the tacit background assumed by the idea of agentive performances that aim at truth. To aim at something, Xúnzˇı would contend, we must “approve” it. The attitude of approval expresses endorsement of certain values or norms. So an implication of Xúnzˇı ’s stance is that ultimately epistemic agency, or at least competent epistemic agency, is a matter of our holding certain values—of our approving, and hence valuing and committing to, certain norms of performance. Our capacity to be epistemic agents rests on our capacity to value, endorse, and act on norms. Hence to the extent that knowledge requires epistemic competence, we can attain knowledge only through a commitment to certain values. An implication is that a key feature that distinguishes the competent epistemic agent from a mere reliable mechanism—a reliable temperature-reporting device, for instance—is the capacity for

Epistemic Competence and Agency in Sosa and Xúnzıˇ

49

axiological commitment, including the capacity to grasp how a system of norms commits us to various epistemic and inferential relations between our epistemic attitudes and the circumstances we find ourselves in. The sort of normative commitment Xúnzˇı depicts here, I suggest, is characteristic of responsibilist virtues such as epistemic conscientiousness, diligence, or perseverance. On a Xunzian approach, then, a responsibilist commitment to relevant norms seems a necessary condition for competence in applying the criteria that underwrite reliabilist virtues. To possess the sort of systematic competence and reliability Xúnzˇı associates with knowledge, one must be an epistemic agent who manifests at least some degree of responsibilist virtue. To reliably get things right, an agent must be conscientious about getting things right. An implication is that the reliabilist virtues needed for the advanced epistemic competence manifested in, for instance, fully apt belief are partly constituted by and intelligible only against a background of responsibilist virtues. An adequate account of epistemic agency will need to weave together both responsibilist and reliabilist components.

Notes 1 2

3

4 5

6 7

8 9

The “Xún” in Xúnzˇı sounds roughly like the second syllable of the English “friction” pronounced with a questioning tone (“-tion?”). The “zˇı ” sounds like “dz.” Sosa’s own (1991) is widely recognized as a groundbreaking account of a reliabilist approach. Code (1987), Montmarquet (1993), and Zagzebski (1996) are pioneering treatments of a responsibilist approach. Sosa describes competence by saying that “its speed and orientation would in normal conditions take it to the bull’s-eye” (2016: 6). This description allows that a beginner archer with unreliable skills could nevertheless make a competent shot. I suggest that competence is probably better understood as referring to the agent’s reliability in achieving the aim. Sosa (2015: 68). Sosa speaks here of “competent” rather “apt” second-order awareness, but I take his point to extend to aptness. Caution is needed here in handling the analogy to the teleological functioning of the heart. Even if the process of belief formation normally happens subconsciously, the agent can control or at least shape it through explicit training—by learning to apply a new set of concepts, for instance—in ways that we cannot control involuntary physiological functioning such as the pumping of the heart. Since functional beliefs are usually not guided by reflective aptness, however, they are unlikely to attain full aptness (2015: 94). One example is that, having measured the length of a line as ten centimeters, we find ourselves “attracted” to assent to the claim that another line, which looks longer, is more than ten centimeters long (cf. Sosa 2015: 193). As Sosa notes, much of what we do freely and responsibly is not by conscious, deliberative choice or decision (2016: 10). Sosa makes a similar observation regarding beliefs arrived at through subconscious intentional acts (2016: 10), but I suggest the point also holds of many beliefs arrived at through self-conscious deliberation. Perhaps Sosa disagrees, as elsewhere he says that we often “freely conclude deliberation” by deciding “whether to accept that the balance of reasons sufficiently favors either side” (2015: 208). My experience as an epistemic agent is that such decisions do occur but infrequently.

50

Ernest Sosa Encountering Chinese Philosophy

10 References to the Xúnzˇı give chapter and line numbers in Hung (1966). 11 For a more detailed discussion of Xúnzˇı ’s theory of perception, see Fraser (2016). 12 For a detailed interpretation of Xúnzˇı ’s notion of “blinkering” and his associated part-whole conception of knowledge and error, see Fraser (2011: 127–148). 13 By “epistemic attitudes,” I mean beliefs and judgments in Sosa’s framework and deemings and discriminations in Xúnzˇı ’s. 14 As I explain in Fraser (2011), Xúnzˇı here draws on a part-whole theory of error. Errors arise from attending to only part of the relevant circumstances rather than the whole. 15 Unlike Sosa, Xúnzˇı acknowledges purely functional, sub-agentive capacities only at the level of the “differentiating” activity of the sense organs. All cognition, whether implicit and automatic or explicit and deliberative, falls within the scope of the “arts of the heart” and thus epistemic agency. Indeed, some of Xúnzˇı ’s descriptions of expert epistemic agents imply that, as with expert performers in many fields, their expertise can be expected to render explicit, deliberative judgment largely unnecessary (5/60–61). Nearly all of their epistemic attitudes arise automatically, without explicitly self-reflective thought.

References Code, L. 1987. Epistemic Responsibility. Hanover, NH: University Press of New England. Fraser, C. 2011. “Knowledge and Error in Early Chinese Thought.” Dao: A Journal of Comparative Philosophy 10(2): 127–148. Fraser, C. 2016. “Language and Logic in the Xunzi.” In E. Hutton (ed.), Dao Companion to the Philosophy of Xunzi. Dordrecht: Springer, pp. 291–321, http://cjfraser.net/ images//2012/09/Fraser-Xunzi-LanguageLogic-preprint.pdf Hung, W. ed. 1966. A Concordance to Hsun Tzu. Harvard-Yenching Institute Sinological Index Series, supplement no. 22. Cambridge, MA: Harvard University Press. Montmarquet, J. 1993. Epistemic Virtue and Doxastic Responsibility. Lanham, MD: Rowman & Littlefield. Sosa, E. 1991. Knowledge in Perspective. Cambridge: Cambridge University Press. Sosa, E. 2015. Judgment and Agency. Oxford: Oxford University Press. Sosa, E. 2016. “Knowledge as Action.” In C. Mi, M. Slote and E. Sosa (eds), Moral and Intellectual Virtues in Western and Chinese Philosophy. New York: Routledge, pp. 5–15. Zagzebski, L. 1996. Virtues of the Mind. Cambridge: Cambridge University Press.

4

In Favor of a Comprehensive Virtue Epistemology? Gregor Paul1

1 Introduction This chapter deals with Ernest Sosa’s distinctions between epistemic and moral characteristics, and, respectively, epistemology and ethics, critically examining these distinctions from the perspective of Xunzi’s theory of knowledge, according to which knowledge results from and manifests both, epistemic and moral competence. In my discussions, I focus on the question of how knowledge ought to be justified. Since Xunzi deals with this question mainly in his theory of communication, I devote some space to analyzing this theory. When I first came across the label “virtue epistemology,” I immediately thought of the philosophy developed in the Xunzi 㥰ᆀ, the Mencius (Mengzi ᆏᆀ), and the Analects (the Lunyu 䄆䃎).2 In these Chinese classics, epistemology is an integral part of, or at least overlaps with, virtue ethics: it is a normative philosophy that extensively focuses on character traits as sources of and/or catalysts to knowledge. The ethics developed in these three classics combines features of virtue ethics, role ethics, utilitarianism, and even deontological ethics. As a whole, this ethics may be categorized as a virtue ethics that incorporates the other three kinds. Role ethics and utilitarian ethics do not play a dominant role. One could even argue that the Xunzi advances a deontological ethics, since on three occasions (chapter 13.2, and twice at 29.1) the text says that one should follow the dao and not the lord (29.1: Cong dao bu cong jun . . . ᗎ䚃нᗎੋ . . ., Hutton 2014: 325). However, following the dao would entail behaving virtuously. Or one could simply state that following the dao is itself a virtue. Conceiving of the ethics in the three classics as mainly or merely role ethics, as e.g. Roger Ames does, is mistaken. See my Aspects of Confucianism (Paul 1990) in which I criticize Ames’s views on ethics, metaphysics, and rationality in the Lunyu (see also Paul 1991, Paul 1992, and, with particular reference to the Xunzi, Paul 2010). In another article, I again try to show that characterizing classic “Confucianism” (rujia ݂ᇦ) as a philosophy of rational thinking does not mean reading “Western” ideas in “Eastern” thought. Influential Japanese scholars (who knew nothing about “Western” interpretations) even denounced “Confucianism” as being too rational (Paul 2013). In 51

52

Ernest Sosa Encountering Chinese Philosophy

categorizing Xunzi’s philosophy, one misunderstanding should be excluded: in spite of its features as practical philosophy, it does include theoretical parts (see Paul 1990). In what follows, I explicitly thematize only a few questions of how to interpret the book as a whole. Also, I do not expressly deal with Xunzi’s theory of monarchy, hierarchy, and reward and punishment (as means of establishing and upholding social and political order). Nor do I discuss the intolerance Xunzi displays in his rejection of “rival” teachings. Such issues lie beyond the scope of my chapter. But when I then read Sosa, I soon realized a significant difference between his virtue epistemology and the Chinese version(s). Sosa conceives of such sources of knowledge as e.g. perception and reason (including “inferential reason”) not only as faculties, but also as virtues (Sosa 1991: 225–227, 270), whereas, in the mentioned classics, these sources are rather regarded as faculties that should be exercised in a virtuous way. One can of course call such faculties “virtues.” Sosa speaks of these faculties as virtues because of their contribution to something good, namely “our quest for truth” (ibid.). He further justifies his naming by explaining that “there is a . . . sense of ‘virtue’ . . . in which anything with a function—natural or artificial—does have virtues. The eye does, after all, have its virtues, and so does a knife” (1991: 271).3 In “Mind–World Relations,” Sosa mentions that we even give “credit” to (the efficiency of) “a sharp knife” (2017: 133).

2 Sosa’s emphasis on intellectual virtues Although “having virtues” may differ from “being a virtue,” and though virtues are usually regarded as morally significant faculties or character traits (which is also the case in the mentioned Chinese classics), the importance of naming should not be overestimated. One should not quarrel about words. But in Sosa, it is no mere question of naming. As I understand him, by calling perception etc. an “epistemic virtue”4 he intends two things: first, to distinguish the epistemic from the moral, thus, so to say, “saving” epistemology as a discipline of its own (Sosa 2015b: 71, 2017: 144–145, 149– 150), and, second, to argue for a normative epistemology.5 To preclude any misunderstanding: Sosa does also thematize character traits that can be conducive to knowledge (of truth), like e.g. courage, and he has even emphasized the importance of dealing with them. However, whereas he regards perception and reason (e.g. inferring) as constitutive of knowledge, he conceives of character traits as merely auxiliary virtues (Sosa 2015b, 2017, 2019). Again this reflects Sosa’s intention to distinguish between the epistemological and the ethical. Not without justification, of course. As I would put it: without perception and/or inference there could and would be no knowledge at all, and though e.g. courage may sometimes also be constitutive of knowledge (when facing threatening opposition, one may need courage to pursue one’s quest for truth), courage is not always required.6 In other words: the epistemic virtues of perception and/or inferring are necessary conditions of gaining any knowledge at all, whereas character virtues are necessary conditions only under certain circumstances. Again, Sosa, if I understand him correctly, conceives of faculties such as perception (vision), inference (drawing), introspection, memory, and testimony as virtues since

In Favor of a Comprehensive Virtue Epistemology?

53

they are competences conducive to knowledge. Knowledge he defines as “apt” belief, i.e. a true (“adequate”) belief that is true because of its resulting from, and manifesting, the competence (“adroitness”) of the believer (2007: 23–24). Thereby, Sosa distinguishes between different levels of knowledge, namely between unreflective knowledge on the one hand and various higher levels of reflective knowledge on the other (Sosa 1991: 290–291, 2007: 32, 2015a: 63, 65, 73, 2016, and 2017: 100, 153–154). Among the motives that led Sosa to developing his epistemology is (i) his intention to solve problems of traditional theories of knowledge, especially what he regarded as shortcomings and/or inadequacies of foundationalism and coherentism (1991), (ii) the challenges of skepticism illustrated by a believer’s performance in dreams, (iii) performances caused by an evil demon, or by a brain in a vat, (iv) the Gettier problem,7 and (v) the problem of the criterion (e.g. 1991, 2007, 2017).

3 Xunzi’s emphasis on character virtues Like Sosa, Xunzi distinguishes between (i) natural and (ii) acquired sources of knowledge, namely (i) perception and reason on the one hand, and (ii) character traits on the other. The former (i) he calls senses (tianguan ཙᇈ) and heart (xin ᗳ).8 Among the latter (ii), he emphasizes humanity (ren ӱ) and right(eous)ness (yi 㗙) (Xunzi 22, 23). As should be clear from the above, it is only the latter that Xunzi regards as virtues. For, according to Xunzi, the heart can also instigate evil deeds. Sosa’s notion of the senses is similar to Xunzi’s, and Sosa’s notion of reason is a more general notion of the faculty of reasoning and reflecting than Xunzi’s notion of heart/ mind. This corresponds to the respective differences in their virtue epistemologies. Like Sosa, Xunzi also held that there cannot be such a thing as “propositional” pure experience. According to Xunzi, it is due to the (working of the) heart that sense impressions enable knowledge. (Xunzi 22, e.g. Köster 1967: 290, Dubs 1983: 490, Hutton 2014: 238; see also Xunzi 21.1, Hutton 2014: 224). The similarities go even further, for Sosa shares Xunzi’s view that “the heart/mind can also instigate evil deeds.” However, whereas he uses his examples of immoral exercises (e.g. 2015b, 2017) of epistemic competences to illustrate the differences between the epistemic and the moral, Xunzi uses his examples to argue for a notion of epistemic competence that includes—in an inseparable way—abidance by moral standards. The Xunzi attributes to Confucius the statement “The first [of the five worst traits] is having a heart [xin] with penetrating insight and using it for dangerous things” (ᗳ䚄㘼䳚, 28.2, Köster 1967: 366, Hutton 2014: 318; see also Xunzi 31, Hutton 2014: 337, or 27). In stating that an “auxiliary epistemic virtue might be an overall personal vice,” Sosa also wants to clearly distinguish between the epistemic and the moral (Sosa 2017: 144). According to the Xunzi, but also in keeping with the Mencius and the Lunyu, the main goal of human life is following the dao 䚃, i.e. the way, principles, rules, practices etc. of humaneness and right(eous)ness, and, in a wider sense, established civilized convention (li ⿞), always striving after and/or displaying knowledge (zhi ⸕)/wisdom (zhi Ც). This implies becoming and/or being a humane and cultivated person (junzi

54

Ernest Sosa Encountering Chinese Philosophy

ੋᆀ). Realizing this goal requires being a knowledgeable (or even wise) person. Therefore, human beings ought to develop and cultivate respective faculties and virtues. Among the faculties discussed are the faculties of correct sensory perception and correct usage of the heart/mind, especially arguing consistently. Utilizing these faculties, speech should be in accordance with (he ਸ) what, for the sake of brevity, may be called “reality” or “ (sense) objects” (shí ሖ), and free of contradictions. For instance, different things ought to be named/called differently, etc. (Xunzi 22, 23). Also, faculties ought to be employed in a virtuous way. In the Xunzi we, e.g., find the impressive notion that a person who engages in discussion without supporting his propositions by reasons, [likes] quarreling (what a humane person should not do).9 As to character-virtues, besides (i) humaneness/benevolence, and (ii) righteousness/ rightness, (iii) respect for civilized conventions (li ⿞) and (iv) interest in knowledge (zhi) (an eagerness to learn [hao xue ྭᆨ]) are emphasized. Further, honesty, uprightness, trustworthiness (xin ؑ), open-mindedness, modesty (Xunzi 5.5, 22, and 23, Paul 1990: 80–81), and courage (yong ࣷ) (Xunzi 5.5, Hutton 2014: 24–25) are regarded as important.

4 Intellectual virtues (faculties) versus character virtues (character traits)? The question of which relevance should be attributed to intellectual virtues and character virtues, and how the relationship between these two kinds of virtues should be conceived of in virtue epistemology, remains a disputed issue (Sosa 2015b, 2017, 2019, Baehr 2015, Battaly 2016, 2019). In what follows, I pursue this question by taking into account the approach in the mentioned three Chinese classics. Let me note that I have some difficulties in understanding Sosa’s solutions, and especially in attributing them to his theory as a virtue epistemology. As indicated, the question of whether to call only intellectual virtues epistemic ones is probably no problem of mere wording but rather a problem of how to evaluate the epistemic relevance of e.g. character traits. Further, in my opinion, the challenges of skepticism are not really important problems. From commonsensical points of view, they appear farfetched. Although commonsensical convictions can be mistaken or false, they ought to be taken into account in epistemology. Sosa does this of course (e.g. 2017: 218–219), but not as thoroughly as adequate criticism of skepticism would require. I do not deny that some skeptics regarded the question of e.g. undoubtable knowledge as an issue of existential relevance. However, one need not agree with their view, but could rather regard it as over-sophisticated and practically irrelevant. Most skeptics lived a life of “pragmatic self-contradiction.” Also, given the many problems humans have to cope with, the intellectual brilliance in arguing for a skeptic position should perhaps be devoted to solving basic problems of everyday human life. This is also the position taken by Xunzi who, however, even goes as far as maintaining that (in almost every case) skepticism displays egoism and longing for fame. Though, in the history of philosophy in the East and West, this kind of argument has been brought forward repeatedly against subtle and differentiated reflection, especially refined logical

In Favor of a Comprehensive Virtue Epistemology?

55

reflection and so-called sophistic reasoning, it is of course far from being generally true. For instance, Zhuangzi was probably no egoist (just) craving for fame. In Chinese philosophy, Zhuangzi was (one of?) the first who raised the skeptical question of how to distinguish between performances in a dream and in a wake station. So-called Chinese Sophists of his time also called into question the possibility of “real” knowledge. Though Xunzi admits that such attempts could display intellectual brilliance, he dismisses them as farfetched, non-commonsensical, not in accordance with (beneficial) human conventions, and—in my view most important—hindrances to acceptable knowledge and (thus also) hindrances to successful human communication and, as a matter of consequence, socially ideal behavior (Xunzi, e.g., 8, Hutton 2014: 56–57, 22, Dubs: 494–498).10 In other words, the Xunzi attributes the skepticism voiced by contemporary Chinese philosophers to persons who lack epistemic and/or moral virtues. (“We” would, however, not accuse Gettier of lacking virtue.11) As indicated above, according to Xunzi, their display of intellectual brilliance manifests self-love, egoism, and craving for fame, whereas someone who refrains from such misbehavior acts virtuously (ibid.). In short, Xunzi regards the differences between different kinds of knowledge (acquiring, cultivating, justifying, and communicating knowledge) as differences due to differences between different kinds of persons. He expressly distinguishes between the understanding of four kinds of men: the understanding of the “wise” (shengren 㚆Ӫ), the “junzi and man of breeding” (shijunzi ༛ੋᆀ), the “petty man” (xiaoren ሿӪ), and “the menial” (yifu ᖩཛ) (Xunzi 22 and 23). Thus, Xunzi’s notion of knowledge comprises/implies both: epistemic and moral competence. To quote from the Xunzi: “[The junzi] explains with a mind of ren” (Yi ren xin shuo ԕӱᗳ䃚, Xunzi 22.12). The junzi’s actions are not guided by selfishness, but by altruism and empathy (ren), avoiding anything that does not contribute to improving the character and life of other persons (including their knowledge), or that may even confuse them. [The menial’s] talents are varied and many but of no practical use, he is full of subtle distinctions and elegant turns of phrase that serve no practical purpose, he ignores [the differences between] “this” and “not this” [i.e. the rules of consistency and/or right and wrong] . . . this is the understanding of the menial. Xunzi 23.22, see also Dubs 1983: 548, Hutton 2014: 255–256, Paul 1990: 80–81

As to pragmatic self-contradiction, the famous Chinese “sophist” Gongsun Long (third century BC) maintained, “A white horse is no horse.” However, according to an anecdote, when crossing a border with his white horse, he had to pay customs for taking a horse with him.

56

Ernest Sosa Encountering Chinese Philosophy

5 In favor of a comprehensive virtue epistemology? Focusing on the question of how to communicate and justify knowledge Central problems of epistemology can be phrased as follows: 1. 2. 3. 4. 5.

What is truth? What is knowledge (of truth)? How to gain knowledge? How to justify that the knowledge gained is indeed valid knowledge? How to communicate to, and convince, others that something is true?12

Justification can be of different kinds and levels. I mention only two. First, and generally, one can conceive of justified knowledge as knowledge gained in a reliable way, i.e. a way that not easily fails (e.g. Sosa 1991: 145, 293, 2007: 87–88). According to Sosa, this would mean acquiring knowledge through adequately exercising one’s epistemic virtues in “appropriate conditions,” thereby ensuring comprehensive coherence of one’s beliefs (Sosa 1991: 145, 293, 2007: 110–111). This definition or understanding of justification is indeed expressly a concept of virtue epistemology. However, it seems to attribute only minor importance to a second, specific level of justification which is usually required in communication.13 In contrast, the Xunzi devotes much space to this issue, advancing an epistemology that is part of a theory of communication, which asks for (i) expressions that fit reality (Xunzi 22), (ii) (logical) consistency and comprehensive coherence of one’s beliefs and behavior, (iii) correct (established, clear, etc.) language use, (iv) honesty (uprightness, truthfulness, and, if necessary, courageous behavior), and (v) a (nevertheless) becoming (modest, polite, open-minded, refined) way in dealing with one’s communication partners (Xunzi 22, 23).14 Though I cannot discuss the issue of how notions of truth are, or ought to be, understood (i.e. the first of the above questions), I should like to point out that, according to probably every viable epistemology, knowledge (of truth) must be, first, somehow “connected/related” to “reality,” and, second, coherent. This implies that it must take account of experience and fulfil the law of non-contradiction. Philosophers in India and Sino-Asia were also among those who put forward epistemologies according to which a kind of “reality-connection” and logical consistency (and even a kind of overall coherence) are necessary conditions of truth. Xunzi sharply criticized superstition and contradiction (Xunzi 5, 17; 21.13, Hutton 2014: 233).15 In his “The Raft and the Pyramid” (Sosa 1991), Sosa tries to solve what he regards as basic problems, or shortcomings, of (traditional) foundationalist and coherentist epistemologies. To put it roughly, and briefly: presupposing that knowledge is true and justified belief (Sosa 1991: 165), Sosa rightly states that neither foundationalist notions of experience (especially pure experience) nor coherentist notions of mere conceptual consistency can justify a belief as knowledge (Sosa 1991: 169–173). In his attempt to adequately taking into account the salient features of both approaches (as I understand it, to do justice to the requirements of empiricalness and consistency), Sosa developed his theory of justification according to which it is the (human) intellectual faculties, especially the faculties of perception, inference, introspection, and memory (e.g. Sosa 2015b: 66)

In Favor of a Comprehensive Virtue Epistemology?

57

that can reliably generate or preserve knowledge, i.e. produce or preserve (true and) justified belief. As mentioned above, he calls these faculties “intellectual virtues.” On page 10 in his “Introduction” to Knowledge in Perspective, Sosa states “to know [the truth] is to believe through a faculty or intellectual virtue.” He adds (note 13) that his “approach is . . . akin to” a certain kind of reliabilism,” a classification shared by other epistemologists (Baehr n.d., Turri, Alfano, and Greco, n.d.). In his criticism of traditional empiricist epistemologies, Sosa also points out that they neglect faculties such as memory, stating that e.g. knowledge of names is (in many cases?) a function of memory. His virtue reliabilism is often compared to, or confronted with, what is called virtue responsibilism. Whereas the first theories conceive of knowledge as (primarily) a function of “intellectual virtues,” the second theories conceive of knowledge as (primarily) a function of “character virtues.” In what follows, I argue that the so-called “Reliabilist/Responsibilist Divide” is no longer as sharp or deep as may have been thought until around 2015. Sosa himself calls “this orthodox dichotomy . . . deeply misleading” (2015b: 62, 2017: viii), and states (2017: viii) that “a true epistemology will approximately combine all four elements” of the “two camps,” namely “reliability,” “competence,” “responsibility,” and “character.”16 Tying in with these assessments, I try to show that if one (1) argues for a liberal epistemology, as Sosa does in his criticism of traditional theories of knowledge by including as a means of knowledge e.g. memory, and if one (2) thereby also argues for (what in my view could appropriately called) a virtue epistemology, one should perhaps advocate a comprehensive virtue epistemology that does not only combine, but integrate the different approaches, thus aiming at a unifying and unified theory. In so doing, I again bring into play the above indicated arguments put forward by Xunzi whose epistemology does already “combine all four elements.” In his “Two Forms of Virtue Epistemology,” Sosa himself acknowledges that (socalled) virtue responsibilism is an important and interesting field. However, as mentioned above, he maintains that epistemic virtues are competences that constitute knowledge, in contrast to character virtues that are merely conducive or auxiliary to attaining knowledge (though Sosa, by saying this, does not want to devaluate virtuous character traits). Sosa further shows that some character traits, as for instance love of truth, often do not at all contribute to, or at least do not play an important role in, gaining knowledge. Also, the knowledge a murderer acquires and employs, Sosa emphasizes (as indicated above), is certainly no manifestation of a virtuous character, though it remains of course knowledge (Sosa 2015b: 72, 2017: 150–151). Sosa thus insists “that epistemology is not a department of [e.g.] ethics” (Sosa 2017: 149), but a discipline of its own that should be understood as such.“Responsibilists” of course reject calling the truth-conducive competence manifested by a murderer a virtue at all. In contrast to Sosa (e.g. 2015b: 70, 2017: 149, Battaly 2016: 101–106), they insist that only what is good and praiseworthy, is and/or should be called a virtue. In another analysis of murder, Sosa (2017: 114–115) arrives at the conclusion that “there is no objective, intrinsic value in trivial enough truth.” Then, the value attributed to such truth must be something extrinsic (to it), or truth combined with something else. It could be certain consequences, a certain use, conduciveness to human welfare, or simply a combination of (knowledge of) truth and moral responsibility, etc. One answer Sosa gives is that certain knowledge can contribute to “human flourishing”

58

Ernest Sosa Encountering Chinese Philosophy

(Sosa 2017: 115–116), namely to improving one’s individual life and the life of one’s fellow beings. This, however, is to say that (such) knowledge is valuable because of its instrumental power, its usefulness for e.g. character-cultivation, for improving morality, and so on. In other words, Sosa thus perhaps willy-nilly extends his reliabilism to a combination of reliabilism and responsiblism, for the attempt to contribute to human flourishing is a function of character traits: perhaps of one’s disposition to strive for selfrespect and one’s humaneness/benevolence/altruism/empathy. If this does indeed follow from Sosa’s account of the value of knowledge, it would be a consequence very similar to the position advanced in the Xunzi. At least the value Sosa attributes to knowledge as conducive to human flourishing cannot be explained as being merely due to epistemic virtues like perception and logical thinking, etc. Sosa would probably agree with this. Xunzi emphasizes that intelligence and courage can (also) lead to robbery and murder (13, Dubs 1983: 161–162). As mentioned above, Xunzi attributes to Confucius the statement “The first [of the five worst traits] is having a heart [xin] with penetrating insight and using it for dangerous things.” Xunzi thus also argues that knowledge, as knowledge is understood by Sosa, is of no intrinsic value, and is of course no virtue. Going further, he even maintains that if one is no junzi, one “cannot have knowledge” (Xunzi 4, Dubs 1983: 66). Without ren and yi and without following li, even a scholar ought not to be regarded as a man of knowledge or wisdom (Xunzi 1, Dubs 1983: 18–20). Also, according to Xunzi, knowledge requires putting it into practice, etc. (Xunzi 8, Dubs 1983: 160) This raises the interesting question whether, according to Xunzi, only conceptions “verified” by “practical proof,” are knowledge. Pursuing this question is however beyond the scope of my analyses. In any case, according to Xunzi, a belief is knowledge (zhi ⸕) only if the belief is put into practice—which again may indicate a moral component of knowledge. Sosa calls his epistemology a “competence virtue epistemology” (2017: 141), implicitly thus referring to his definition that belief is knowledge since its truth is due to the epistemic competence of the believer. Again, such competence he calls a virtue because of its truth-conduciveness. However, in his “Two Forms of Virtue Epistemology,” he seems to restrict this notion to a notion of truth-constitutiveness. This is to say, as explained above, that epistemic faculties, such as e.g. perception, are virtues because they are not merely truth-conducive but also truth-constitutive. Does this mean that his epistemology is actually a theory of knowledge-constitution? But then one may ask: is a faculty like memory indeed constitutive of knowledge? In his continuing attempts to refine his notion of epistemic competence, Sosa expressly included personal skill, state, and situational adequacy among its characteristics, speaking of an SSS competence.17 Xunzi’s epistemology is of course much less explicit, though he also deals with the impact that skill, state, and situation has on epistemic performance. For example, he repeatedly discusses a junzi’s intelligence (skill) and courage (shape) in encountering/ criticizing a ruler (situation). In any case: Xunzi developed what could be called a comprehensive virtue epistemology although there was of course in his time no respective label. In one way or the other, Xunzi deals with and answers all five questions formulated above. To indicate his solutions: 1. A belief/speech is true if it accords with the dao (he yu dao ਸᯬ䚃) insofar as it is (i) free of contradictions (bu fu нᣲ, ci he yu shuo 䗝ਸᯬ䃚), and (ii) consistent

In Favor of a Comprehensive Virtue Epistemology?

2.

3.

4.

5.

59

with/fits reality (yuantian guan 㕈ཙᇈ, zhi qing 䌚䃻). However, it is, to say the least, difficult to identify in the Xunzi such a narrow notion of truth. Rather the Xunzi’s notion is a more complex one, referring to a truth that is also a moral or social achievement/good.18 Knowledge (zhi ⸕) is acceptable (ke ਟ) true belief manifest in one’s behavior and “actions.” It is an appropriate means to contribute to individual and social perfection/flourishing. Ideal knowledge—often called wisdom (zhi Ც)—is a kind of highly reflective and (almost) infallible knowledge. As reflective knowledge, it is a belief the believer is aware of, resultant from his conscious reflection on the dao and (as Sosa would put it) on his (the believer’s) skill, shape, and situation. There are also phrases similar to definitions of knowledge such as, e.g., “To know that the right is right and the wrong is wrong is wisdom” (Xunzi 2, Dubs 1983: 28);19 or, strongly reminiscent of Lunyu 2.17: “When you know, you say you know, when you do not know, you say you do not know” (Xunzi 8.21, Dubs 1983: 156, Hutton 2014: 63; also Xunzi 29.5).20 As to the relation between knowledge and reality, the Xunzi (22.2, Dubs 1983: 484, Hutton 2014: 236) states: “Knowledge which corresponds to reality is called wisdom.”21 Knowledge is gained by sensory perception, logical thinking, and learning, especially learning (xue ᆨ) from others/“teachers” and from (reliable) tradition/ transmission/ testimony, excluding as a source any kind of superstition. Like Sosa, the Xunzi regards e.g. testimony as an important source of knowledge, in this respect also advocating a “liberal” epistemology.22 Knowledge can be justified (or defended) by being (i) in accordance with and/or (ii) founded in the dao and reality. (iii) Actual justification is possible as (a) justification for oneself and/or (b) intersubjective justification. Though (b) presupposes or implies (a), (b) is of much greater social relevance. Thus, the Xunzi focuses on (b). This point qualifies Xunzi’s epistemology as both (i) a kind of coherentism and (ii) foundationalism. For, according to the Xunzi, justifiable knowledge is (i) consistent belief within a coherent whole of beliefs, all of which accord with the dao, and is (ii) founded in the dao and in reality. Intersubjective justification requires communication. Successful communication, however, usually presupposes more than mere truth of an argument. That is to say that, in many cases, truth, if understood as mere “accordance with reality” and/or logical consistency etc., is no sufficient criterion of acceptability. Sometimes, it may even be better to lie than to speak the truth (e.g. Lunyu 13.18, and 17.20).23 Also, truth/knowledge/understanding (zhi ⸕) may be used to successfully cause social unrest. Accordingly, the Xunzi distinguishes between (i) truth on the one hand and (ii) its usage on the other. As indicated, in communication, what one says ought to be acceptable for otherwise one’s attempt to convey to others and/or convince others of the truth of one’s knowledge, and/or to inter-subjectively justify it, is doomed to failure. This, in turn, demands that knowledge should be expressed in a becoming way: discernable as a manifestation of modesty, politeness (kuan ሜ, rong ᇩ), open-mindedness, honesty, truthfulness (xin ؑ, cheng 䃐), and (unpretentious) humaneness (e.g. Xunzi 22 and 23). Most

60

Ernest Sosa Encountering Chinese Philosophy importantly, (communicative) justification of knowledge requires supporting arguments with explanations and expressing oneself as clearly as possible,24 for otherwise one would rather display arrogance than e.g. honest interest in furthering the knowledge and wellbeing of others. Now, since without acceptable usage no belief can constitute knowledge and since such usage depends on one’s character traits, it is these character traits that are decisive. So, if the issue of (intersubjective) justification is regarded as a legitimate topic of epistemology, then character traits deserve comprehensive and detailed treatment in virtue epistemology. The more so, since in certain cases virtues like courage may perhaps (also) be constitutive of knowledge.25

Successful learning (xue ᆨ)—which is probably the most important source of acquiring truth and knowledge—may also depend on certain character traits, especially eagerness to learn, persistence, open-mindedness, and a kind of modesty and honesty (that excludes self-betrayal). For example, the Xunzi says that one should “listen with a heart of learning (yi xue xin ting ԕᆨᗳ㚭)” (Xunzi 22.12, Dubs 1983: 502, Köster 1967: 294; see also Xunzi 1). Thus, the question arises again whether, under certain normal conditions (in Sosa’s sense), certain character traits do not only facilitate, but also contribute to, constituting knowledge. In other words: are there normal situations in which one would not be able to acquire knowledge if one did not possess certain character traits? And if so, would this question be e.g. a merely psychological and/or social question, or (also) an epistemological one? Anyway, according to the Xunzi, communication of knowledge ought to be a function and/or manifestation of personal virtues.

6 In favor of both: specific and comprehensive virtue theories of knowledge? I need not emphasize that it is not only possible, but even desirable, to develop different forms of virtue epistemology, and that such forms can very well be compatible with each other, even fruitfully complementing each other, and that it is of course also desirable to expressly distinguish between e.g. moral and non-moral faculties—though one perhaps should not rule out that moral faculties, as far as they are (actually) conducive to knowledge, are also epistemic powers. Perhaps one should admit that there are epistemically and morally inseparably complex virtues? That, in certain cases, constitution of knowledge is due to and manifests both: intellectual virtues (in Sosa’s sense) and character virtues? Although one can (and should) analytically distinguish between different virtue concepts, actual (acquisition of) knowledge may be due to (an exercise of) insolubly combined virtues (or competences). Thus, Xunzi’s virtue epistemology can enrich the specter of ‘Western’ virtue epistemology. Xunzi does not deal with the faculties of perception and inference as separate means of gaining and justifying knowledge (or understanding), but as integral parts of a whole of interrelated and interdependent powers/virtues (which in Chinese could be called de ᗧ) that comprise epistemic and moral features.26 Further, Xunzi’s

In Favor of a Comprehensive Virtue Epistemology?

61

pragmatism and the communication theory it entails may also further the endeavor to develop a comprehensive virtue epistemology.27 Perhaps one should even try to develop a universally valid—transcultural—virtue theory of knowledge? This is what I would advocate. Or should one inquire into certain culturally relevant distinctions? Except for fundamentalist positions, there are probably none.28 At least one could further reflect on the methodology and criteria of a viable virtue epistemology.

Appendix Responsibilist reliabilism in the Xunzi? Hutton (2014) offers translations of important passages in the Xunzi that suggest impressive systematic concurrences with Sosa’s virtue epistemology, though the differences in Xunzi’s and Sosa’s notions of truth and knowledge/understanding (zhi ⸕ and zhi Ც) and the differences in their estimations of character traits/moral virtues as means of, and contribution to, knowledge remain. Hutton several times renders the Chinese chang ᑨ as “reliable,” or “reliably.” Since the basic meaning of chang is probably “constant/ly,” thereby often implying “regular/ly” (Mathews 1963: 25), Hutton’s choice is justified. Among the phrases of particular interest is the following exemplary/paradigmatic one, which Hutton (2014: 27) translates as Ren, yi, and virtuous conduct [de ᗧ, virtue, power] are reliable means to safety [Xunzi 4.9, ӱ㗙ᗧ㹼ˈᑨᆹѻ㺃ҏ], but that does not mean that one will never encounter danger.

This is to say, first, that certain virtues reliably (constantly, regularly) enable realizing safety. Second, and even more important, these virtues not only include character virtues, but also intellectual virtues. (In this respect one may regard Xunzi’s epistemology as “wider” or more comprehensive than Sosa’s.) Since “safety,” as a result of ren, yi, and de, also (ultimately) requires exercising e.g. correct perception and inference, and an even overall consistent reflection (for otherwise there would be unrest, etc.), the whole phrase can be justly understood as implying that epistemic virtues—no matter whether of a responsibilist or reliabilist kind—are reliable sources of knowledge (which is thus also justified). Since the Xunzi emphasizes the first kind, one may even speak of a responsibilist reliabilism, the more so, since notions such as e.g. ren comprise both kinds in an actually inseparable manner. One may even argue that specific notions of ren (which implies specific notions of character traits) logically imply the notion of correct perception and inference (i.e. a notion of intellectual virtue). Ultimately, Xunzi’s epistemology proves to be a virtue theory since—according to Xunzi, but wording it similarly to how Sosa would express it—belief is knowledge (zhi ⸕) only if it derives from and manifests (correct exercise of) the (comprehensive) virtue of a dao-based attitude. As a matter of consequence, knowledge thereby fits reality, is coherent, and is no question of luck. The specific difference, namely that zhi

62

Ernest Sosa Encountering Chinese Philosophy

includes respective practice, of course remains. For example, “understanding correct government” entails obeying humane government and criticizing tyranny.

Notes 1 2

3

4 5 6

7

I should like to thank Dr. Michael Poznic from Karlsruhe Institute of Technology (KIT) for his valuable suggestions. Large parts of these books are attributed to the philosophers Xunzi (third century BC), Mencius (fourth to third century BC), and Confucius (551–479). In my use of the names of these philosophers, I always refer to the respective texts. Sosa continues: “And if we include grasping the truth about one’s environment among the proper ends of a human being, then the faculty of sight would seem in a broad sense a virtue in human beings; and if grasping the truth is an intellectual matter, then that virtue is also in a straightforward sense an intellectual virtue.” In my opinion, this conclusion is problematic because of the following reasons: the faculty of sight may enable a murderer to successfully stab to death his victim. Is a faculty that enables such a deed a virtue? Sosa could perhaps reply, “Yes, an epistemic virtue.” Some virtue epistemologists reject such a use of the term “virtue” (Baehr n.d., Turri et al., n.d.). But if stabbing to death is a misuse of the epistemic virtue of sight, this implies that a virtue can be misused. But does it make sense to regard as a virtue a faculty that (easily, by the way) can be misused? Even if one distinguishes between different aspects? There is of course no problem with maintaining that it is rather the “right” usage of a faculty of sight than the faculty itself that has virtue but then Sosa’s intentions may be compromised. Sosa also uses the expressions “cognitive virtue” (1991: 275) and, often, “intellectual virtue.” Sosa argues that “knowledge [as] a form of action . . . is . . . subject to a normativity distinctive of action” (2017: 207). Although Sosa perhaps does not completely rule out the possibility that courage can constitute knowledge (Sosa 2015b: 68, point 7; Sosa 2017: 141–142, 144), he is inclined to answer in the negative. In his view, whereas a virtue that constitutes knowledge is necessarily manifest in this knowledge, this is not the case with e.g. courage (Sosa 2017: 144). However, if one gains a certain knowledge p only because of one’s courage, p may manifest this courage. For example, (the attainment of) certain knowledge gained by medieval scientists who knew very well that the Christian Church threatened to put to death anybody who held such belief, may not only manifest epistemic competence, but—even evidently—also courage. Reactions such as “What a courageous man!” or “What a display of courage!” may prove this. One may also argue, similar to Baehr (2015), that in certain cases it is not simply e.g. logical reasoning, but courageous logical reasoning that constitutes knowledge. This is to say that actual knowledge acquisition does not occur in two steps: first a display of courage that (only) “puts one in a position to know” (Sosa 2015b: 67, 69, Sosa 2017: 142, 144) and after that use of logic to get knowledge. Sosa (2017: 62) conceives of this problem as the question of whether “knowledge is tantamount to justified true belief.” As I would put the question: “Can unjustified and/ or unjustifiable belief be, and/or should be, called knowledge?”

In Favor of a Comprehensive Virtue Epistemology? 8

9

10

11

12

13

14

63

Often, the translation “heart-mind” is used, for xin not only designates a faculty and/ or organ to morally evaluate and motivate, but also an epistemic faculty and/or organ. However, some translators render xin just as “heart” or “mind.” Bian er bu shuo zhe zheng ye 䗟㘼н䃚㘵ҹҏ (Xunzi 4.2, Köster 1967: 30, Hutton 2014: 23). I follow Köster’s understanding of the Chinese phrase. A probably more “literal” rendering of the Chinese phrase would be: “debating without explaining is/ amounts to quarreling.” Cf. Xunzi 3.4 (Hutton 2014: 17), according to which the junzi engages in debates without being quarrelsome (bian er buzheng 䗟㘼н⡝) (see Paul 1990: 38–119, especially 82–83). I wrote that book without any knowledge that there existed something like a discipline called “virtue epistemology.” In Xunzi’s world, the epistemological skepticism that he criticized may indeed have been liable to causing harm, though such a hypothesis would need elaboration. On the other hand, even in his times there existed both, namely what could be called a realm of public communication and a more or less closed academic scene. After all, Xunzi was a member of a kind of academy. In the twenty-first century, the differences between the two realms have become extreme. In spite of all his detailed analyses of knowledge, Sosa too is of course aware that one should not try “too hard for [an] exact, fully general analysis” (2017: 138). However, his main reason is the impossibility of a certain kind of analyses, whereas, according to the Xunzi, such analyses and differentiations might ultimately even endanger morality and social harmony. In practice, person-centered and character-centered approaches may invite the rejecting of true but unwelcome beliefs/propositions as wrong/false by simply denouncing them as views of incompetent and/or morally deficient persons. On the other hand, restricting evaluation of a proposition or inference to merely judging its empirical correctness and/or its internal consistency without considering the proponent’s skill, state, and situation (what Sosa calls the proponent’s SSS competence) may lead to logical misunderstandings, and may even be unfair. One may argue that this question is mainly e.g. a psychological and/or ethical problem. Perhaps, Sosa would sympathize with such a view. Dougherty (2019) goes as far as maintaining that there are no epistemic virtues at all. However, if one wants to convince others of the truth of a certain proposition, one cannot avoid arguing that this proposition is epistemically valid (consistent, etc.). Sosa touches on this issue/topic when he speaks of “a social component of knowledge” and of “dependable informants and cognizers” (Sosa 1991: 275–282). However, thereby, his focus is on the relevant fields and conditions of (reliable, regular, nonaccidental) knowledge-conveyance without dealing (as the Xunzi does) with such character traits as e.g. honesty and trustworthiness that certainly contribute to making a subject a “dependable source of information” (ibid.: 282). Though these character virtues may not be constitutive of truth (since even a liar knows the truth), they can be necessary conditions for successful (and reliable) truth conveyance. The Buddhist “Science of Reasons” (Skt. hetuvidyā, Chin. yìnming ഐ᰾) (a discipline that comprises ontology, epistemology, logic, and argumentation-theory), which played an important role in China between about 650 and 1000, explicitly distinguishes between “inferences for oneself ” and “inferences for others,” listing rules of correct and thus successful communication as well as mistakes that ought to be avoided (in this regard similar to many treatises in the history of Aristotelian logic). Interestingly, Sosa distinguishes between two kinds of “affirmation” (e.g. of a proposition): “private affirmation to oneself ” and “public assertion” (Sosa 2017: 153, note 12, and 155; see also 1991: 277). This distinction indicates something similar to

64

15 16

17

18

Ernest Sosa Encountering Chinese Philosophy the mentioned distinction, namely that justification of one’s belief in communication with others is usually based on one’s prior “private affirmation,” i.e. one’s own conviction that one’s proposition in question is true. Perhaps virtue epistemology could, or even should, be developed further also along the thus indicated lines. Outstanding representatives of the “Science of Reason” insisted that true propositions must result from (direct) perception and/or (correct) logical reasoning. In his contribution to Battaly (2019), Baehr states that “the line between responsibilist and reliabilist approaches to virtue epistemology has begun to blur” (Baehr 2019: 96). Thereby, besides Sosa’s Judgment and Agency (Oxford University Press, 2015), he has probably in mind Sosa’s remark that “the exercise of such intellectual virtues [as e.g. character traits like persistence and resourcefulness] need not and normally will not [italics mine] constitute knowledge” (Sosa 2015a: 68, Baehr 2015: 77). Independent of Baehr, I have arrived at similar conclusions, though Sosa, in his “Telic Virtue Epistemology” (2019: 22), again emphasizes the differences between “‘auxiliary’ intellectual virtues [that are of central interest to responsibilism] . . . [and] ‘constitutive’ intellectual virtues of central interest to virtue reliabilism.” Actually, I have difficulties identifying a single definite definition of Sosa’s notion of epistemic virtue. According to Heather Battaly, Sosa argues that “an epistemic virtue is a stable disposition to reliably produce true beliefs” (2016: 99) or “a stable reliable disposition . . . [a] disposition that produces more true beliefs than false ones” (2016: 101, slightly changed). Sosa himself also several times speaks of epistemic virtues as dispositions. Perhaps the following formulation can serve as a summary definition taking into account Sosa’s explanations in “Two Forms of Virtue Epistemology”: an epistemic virtue is a stable and reliable disposition to constitute, in appropriate conditions, knowledge by exercising and manifesting epistemic competence/s, namely, “abilities[s] to succeed when one tries” (Sosa 2017: 191). Thereby, appropriate conditions require adequate personal skill, an adequate personal state, and an adequate situation (i.e. SSS competence). But again: could memory be a constitutive competence? And if only conducive to knowledge, would a definition that, instead of knowledge-constitutive virtues, speaks of virtues conducive to knowledge, not willy-nilly include character-traits? Sosa’s latest definitions of knowledge concur with earlier ones insofar as he characterizes knowledge as “apt belief,” e.g. “[true] belief whose correctness is attained through epistemic competence.” As pointed out, however, he supplements this characterization by adding that this belief has to be derived “in the right way” (Sosa 2017: 125), namely as a manifestation of a competence which is due to the believer’s skill, shape (the condition the believer is in) and the situation (he is in) (Sosa 2017: 131–134, 209–210, 218), Such SSS competence e.g. excludes luck as a source of knowledge. Noteworthy, the Xunzi too emphasizes that a belief that results from luck is no knowledge (8, Hutton 2014: 64–65). Sosa and Xunzi also agree that certain situations or conditions may preclude correct perception (and thus knowledge) (21.12, Hutton 2014: 232–233). As the Xunzi puts it: “conditions” may “disorder the senses.” In my Aspects of Confucianism (Paul 1990: 38–119, especially 75–96), I try to give a comprehensive and detailed account of the Xunzi’s epistemology and its notions of truth, knowledge, and communication. In sinology, it is sometimes disputed whether classics as e.g. the Xunzi formulate at all what could be called a notion of truth. Although I am convinced that the Xunzi does (though the text does not include a single word/term that could be translated as “truth,” but “merely” terms that refer to what “we” would call “truth” as something “adequate,” “appropriate,” and/or

In Favor of a Comprehensive Virtue Epistemology?

19

20

21

22

65

“acceptable”), I cannot discuss the issue here. The problem is discussed in Lenk and Paul (1993). As to zhi ⸕, “to know/understand,” “knowing/understanding,” “knowledge” etc., some explanations are in place too. In his contribution to the conference “Ernest Sosa Encountering Chinese Philosophy,” Christoph Harbsmeier, in a brilliant analysis, provided abundant evidence that, in Xi Kang ፷ᓧ (223–262), zhi implies, or refers to, “knowing how to” in the sense of “understanding how to act . . . on the assumption that P,” and/or “having a someway justified or justifiable certain, presupposed or shared belief that P.” In the Xunzi, zhi has the same basic meanings. Differing from Harbsmeier (2019), however, I do not share his criticism that Sosa’s epistemology is inacceptable, or at least highly questionable, since Sosa, according to Harbsmeier, uses English as a default language of epistemology. Dealing with such a critique, however, would demand discussing once more basic questions about the relationship between a particular language and the “thought” expressed by this language. I have discussed this problem in my publications about the relationship between logic and language and about how to avoid reading ideas distinctive of one’s own culture into a foreign text (e.g. Paul 2008). But one may also judge from the contributions to this volume whether or not it makes sense (proves fruitful, thought provoking, stimulating, “enlightening,” etc.) to compare, or even apply, Sosa’s epistemology to theories formulated in Chinese. By the way, some problems Harbsmeier mentions also arise if we compare e.g. the use and/or meaning of “knowing” in different English epistemologies, e.g. reliabilism and responsibilism. Shi shi fei fei wei zhi ᱟᱟ䶎䶎䄲ѻ⸕. Dubs’s translation indicates that knowledge implies epistemic and moral competence, perhaps in an inseparable way. A more literal translation may run as follows: “To say/hold ‘it is,’ if it is, and ‘it is not,’ if it is not, is knowledge/knowing/understanding/is to know.” Actually, translations of shi and fei deserve separate studies. In my opinion, Neo-Mohist influence on the Xunzi should thereby be taken into account. The Lunyu reads: zhi zhi wei zhi zhi, bu zhi wei bu zhi, shi zhi ye ⸕ѻ⡢⸕ѻ н⸕⡢н⸕ᱟ⸕ҏ. Legge (1872): 151 translates: “When you know a thing, to hold that you know it; and when you do not know a thing, to allow that you do not know it;—this is knowledge.” The Xunzi reads: zhi zhi yue zhi zhi, bu zhi yue bu zhi ⸕ѻᴠ⸕ѻн⸕ᴠн⸕.Dubs (1983): 156, translates: “When you know, you say you know, when you do not know, you say you do not know.” Legge would have translated: “When you know a thing, to say that you know it; and when you do not know a thing, to say that you do not know it [that is knowledge].” The only difference is that the Lunyu uses wei (“to hold”), whereas the Xunzi uses yue (“to say”). In his interpretation of Lunyu 2.17, Sosa rightly emphasizes that the Lunyu (implicitly) distinguishes between first-order knowledge and reflective knowledge (Sosa 2015: 325–330). This refers of course to a general level that abstracts e.g. from specific/particular meanings of “to know” and zhi. However, in my opinion, it would be difficult to maintain that, in 2.17, zhi does exclude “knowing that.” The Xunzi is extensively concerned with both kinds of knowledge, criticizing for instance superstitious beliefs as mistakes/errors on an object-level, and demanding again and again that one ought to strive for reflective knowledge. Zhi you suo he wei zhi zhi ⸕ᴹᡰਸ䄲ѻᲪ. Note the explicit distinction between knowledge and wisdom. There are, however, also translations that instead of reality speak of “facts” or (as e.g. Hutton) “things.” In contrast, the most outstanding representatives of the “Science of Reason” did regard only perception and inference as reliable sources of knowledge. Testimony had to

66

23

24 25

26

27

28

Ernest Sosa Encountering Chinese Philosophy accord with perception and had to be consistent. However, Xunzi would probably have agreed with these criteria of valid testimony. This attests to a pragmatic way in dealing with truth. However, this “pragmatism” is based on the conviction that morality is of higher value than truth. As to pragmatism in dealing with truth, cf. also Sosa (2015a): 73, and (2017): 153. Sosa (1991: 290) states: “Justification of a belief that p requires . . . use of reasons in favor of p.” Sosa may remark that his notion of an SSS competence (implicitly) excludes such cases from his definition of knowledge since the relevant epistemic competences do not meet the requirements of normal shape and situation (Sosa, 2007: 82–84, 2017, e.g.: 131–136, 191–203). However, according to Sosa’s terminological decisions, this could imply that a belief only acquirable under non-normal conditions, such as e.g. a belief attainable only by exercising outstanding courage, should not be called “knowledge,” since such a belief could not be “apt” (2007: 82). But if so, would this not again indicate that insistence on a more or less “pure” epistemological reliabilism would need too many subtle differentiations and supplementations? There is of course also the question of what should be regarded, and/or called, “normal.” In medieval Europe, scientists who pursued certain lines of research often ran the risk of being executed. Was this then a “normal situation?” Or think of the oppression of correct but unwelcome beliefs in Saudi Arabia. A “normal” situation? Baehr (2015) has discussed Sosa’s virtue epistemology in quite a similar way as I do. He mentions two points perhaps further supportive of my conclusions, namely that it is often not mere “good vision,” that constitutes knowledge but rather “attentive and careful visual perception,” and that it is often not mere “introspection” that leads one to correct one’s epistemic error, but one’s “intellectual honesty” (82). In such cases, Baehr contends, character traits do contribute to constituting knowledge. Though Xunzi does not go as far as Mencius (and probably Confucius too), who regarded both epistemic and moral dispositions as inborn/natural human characteristics, he insists that actual (valid) knowledge is a function of both, e.g. the faculties of the senses and heart/mind (perception and reasoning) on the one hand, and (acquired) moral competence on the other. I add this somewhat redundant explanation to point out that my interpretation(s) are compatible with Kim-chong Chong’s interpretation of the Xunzi in his contribution to this volume, “Xunzi, Zhuangzi and Virtue Epistemology” (Chapter 2). As indicated above (p. 1, note 2), Xunzi’s ethics is no mere utilitarianism, but also includes deontological features. It also is by no means a kind of traditionalism. Xunzi 29.8 (Hutton 2014: 335) attributes to Confucius the words that one of the outstanding virtues of the “great sage” (da sheng བྷ㚆) is the sage’s ability to change things (hua . . . wu ॆ . . . ⢙). Probably, Sosa would sympathize with this opinion. See e.g. Sosa (2007): 69.

References Alfonso, M., ed. 2015. Current Controversies in Virtue Theory. New York and London: Routledge. Baehr, J. n.d. Virtue Epistemology, https://www.iep.utm.edu/virtueep/ (accessed 20 January 2019).

In Favor of a Comprehensive Virtue Epistemology?

67

Baehr, J. 2015. “Character Virtues, Epistemic Agency, and Reflective Knowledge.” In M. Alfonso (ed.), Current Controversies in Virtue Theory. New York and London: Routledge, pp. 74–90. Battaly, H. 2016. “Epistemic Virtue and Vice: Reliabilism, Responsibilism, and Personalism.” In C. Mi, M. Slote and E. Sosa (eds), Moral and Intellectual Virtues in Western and Chinese Philosophy, New York: Routledge, pp. 99–120. Battaly, H. ed. 2019. The Routledge Handbook of Virtue Epistemology. New York and London: Routledge. Dougherty, T. 2019. “There Are No Epistemic Virtues.” In H. Battaly (ed.), The Routledge Handbook of Virtue Epistemology. New York and London: Routledge, pp. 127–137. Dubs, H. 1983. The Works of Hsüntze, Chinese and English. Taipei: Confucius Publishing Co. Hall, D. and Ames, R. 1987. Thinking through Confucius. New York: SUNY Press. Harbsmeier, C. 2019. “The Hazards of the Use of English as a Default Language in Analytic Philosophy.” https://www.academia.edu/39195863/The_Hazards_of_the_Use_ of_English_as_a_Default_Language_in_Analytic_Philosophy_A_Plea_for_ Philosophical_and_Historical_Cognitive_Ethnography. Hutton, E. 2014. Xunzi: The Complete Text. Princeton: Princeton University Press. Hutton, E. 2015. “On the ‘Virtue Turn’ and the Problem of Categorizing Chinese Thought.” Dao 14: 331–353. Köster, B. 1967. Hsün-tzu, Kaldenkirchen: Steyler. Legge, J. 1872. The Chinese Classics. Volume 1. Oxford: Clarendon Press. Lenk, H. and Paul, G. eds. 1993. Epistemological Issues in Classical Chinese Philosophy. New York: SUNY Press. Lunyu. https://ctext.org/analects. For translations see https://en.wikipedia.org/wiki/ Analects#Notable_translations. Mathews, R. H. 1963. Mathews’ Chinese–English Dictionary. Cambridge, MA: Harvard University Press. Mengzi. https://ctext.org/Mengzi (accessed August 2019). Mi, C., Slote, M., and Sosa, E. eds. 2016. Moral and Intellectual Virtues in Western and Chinese Philosophy. New York and London: Routledge. Paul, G. 1990. Aspects of Confucianism, Frankfurt/M: Peter Lang. Paul, G. 1991. “Reflections on the Usage of the Terms ‘Logic’ and ‘Logical’ in Comparative Philosophy.” Journal of Chinese Philosophy 18(1): 73–87. Paul, G. 1992. “Against Wanton Distortion: A Rejoinder to Hall’s and Ames’s Critique of my Views on Confucius and Logic.” Journal of Chinese Philosophy 19: 119–122. Paul, G. 2008. Einführung in die Interkulturelle Philosophie. Darmstadt: WBG. Paul, G. 2010. “Confucian Notions of Criticism: An Important Contribution to a Valid Theory of Humaneness.” In W. Teng and Ch. Shan (eds), Confucianism and Its Current Missions, vol. 4. Beijing: Jizhou Press, pp. 194–213. Paul, G. 2013. “Confucian Universalism as a Driving Force of Humanity: The Timeless Lesson of Japanese Confucianism.” In Confucianism in Intercultural Perspectives: Modern Developments. Papers from the Forth International Conference of Sincology. Taipei: Academia Sinica. Sosa, E. 1991. Knowledge in Perspective, Cambridge: Cambridge University Press. Sosa, E. 2007. A Virtue Epistemology. Oxford: Clarendon Press. Sosa, E. 2015a. “Confucius on Knowledge.” Dao 14: 325–330. Sosa, E. 2015b. “Virtue Epistemology: Character Versus Competence.” In M. Alfonso (ed.), Current Controversies in Virtue Theory. New York and London: Routledge, pp. 62–74.

68

Ernest Sosa Encountering Chinese Philosophy

Sosa, E. 2016. “Knowledge as Action.” In C. Mi, M. Slote and E. Sosa (eds), Moral and Intellectual Virtues in Western and Chinese Philosophy, New York: Routledge, pp. 5–15. Sosa, E. 2017. Epistemology. Princeton and Oxford: Princeton University Press. Sosa, E. 2019. “Telic Virtue Epistemology.” In H. Battaly (ed.), The Routledge Handbook of Virtue Epistemology. New York and London: Routledge, pp. 19–25. Turri, J., Alfano, M., and Greco, J. n.d. Virtue Epistemology, https://plato.stanford.edu/ entries/epistemology-virtue/ (accessed August 2019). Wright, S. 2019. “Are Epistemic Virtues a Kind of Skill?” In H. Battaly (ed.), The Routledge Handbook of Virtue Epistemology. New York and London: Routledge, pp. 58–68. Xunzi. Chinese Text Project, https://ctext.org/Xunzi (accessed August 2019).

5

Detachment A Trait-Reliabilist Virtue in Linji’s Chan Buddhism Tao Jiang

1 Introduction It has been forty years since Ernest Sosa published his landmark paper, “The Raft and the Pyramid: Coherence vs. Foundations in the Theory of Knowledge” (1980), which heralded a new approach to epistemology, i.e. virtue epistemology (VE). The popularity of VE has been attributed to its unique ability to tackle difficult philosophical problems that are central to epistemology in contemporary analytic philosophy, e.g. coherentism, foundationalism, internalism, externalism, and, of course, the Gettier problem, etc. This chapter is an attempt to bring the considerable conceptual resources developed in virtue epistemology to Chinese philosophy, especially Chinese Buddhist philosophy, in order to tread a new path in the study of the latter. In so doing, I also hope to complicate the landscape of contemporary virtue epistemology, which is divided between reliabilists and responsibilists. One important contribution of Sosa’s approach to virtue and epistemology, for the purpose of this chapter, is his call to differentiate moral and intellectual virtues in his 1980 article: We need to consider more carefully the concept of a virtue and the distinction between moral and intellectual virtues. In epistemology, there is reason to think that the most useful and illuminating notion of intellectual virtue will prove broader than our tradition would suggest and must give due weight not only to the subject and his intrinsic nature but also to his environment and to his epistemic community. Sosa 1980: 23

Indeed, studies of virtue within the field of Chinese and Buddhist philosophies tend to be overwhelmingly focused on moral virtues without paying much attention to intellectual virtues. In this connection, engaging virtue epistemologists like Sosa has offered me an opportunity to read some texts through a different lens, especially the transformation of Buddhism from India to China, and to see the broader Buddhist 69

70

Ernest Sosa Encountering Chinese Philosophy

project from a new perspective in a way that is unexpected and constructive at the same time. In this chapter, I will characterize the Chan/Zen Buddhist epistemology as a version of virtue epistemology, using some of the conceptual innovations developed in virtue epistemology to offer a new interpretation of the Chan project of enlightenment. More specifically, I will make the case that detachment, one of the key characteristics of enlightenment in Buddhism, should be understood as a reliabilist trait-virtue that is constitutive of the enlightened knowledge in Chan Buddhism, leading to the truth of emptiness celebrated in Mahāyāna Buddhist philosophy. The chapter has two goals in mind: first, it brings a particular epistemic virtue, detachment, articulated and celebrated within the Buddhist tradition, to a broader discussion on epistemic virtues in contemporary philosophy; second, it calls into question the way epistemic virtues are framed by reliabilists and responsibilists in the contemporary discourse. My discussion will be anchored on a general contour of the Buddhist approaches to knowledge, with the Linji lu 㠘ࣁ䤴, a Chan Buddhist text traditionally attributed to the famous ninth-century Chan master Linji 㠘☏, as the locus.

2 Debate about two kinds of epistemic virtues Virtue epistemology is an alternative to the traditional approach to the study of knowledge in contemporary analytic philosophy. Traditional epistemology is primarily concerned with propositional knowledge in the form of “S knows that p.” Such a knowledge is understood as justified true belief (JTB) in that knowledge has three individually necessary and collectively sufficient components: belief, truth, and justification, namely S knows that p if and only if p is true and S is justified in believing that p. The traditional approach to epistemology is concerned with a subject’s belief in examining whether it is true and justifiably so. Since the discovery of Gettier problems in the 1960s, epistemologists have attempted to offer an amended definition of knowledge, much of which has taken the form of JTB + X wherein X refers to whatever further condition or qualification is required to safeguard knowledge proper, although pinning down X still largely eludes the epistemic community. The Gettier-type problems have shaken the belief-based approach to epistemology, so much so that it is fair to say that epistemic luck presents one of the greatest, if not the single greatest, threat to knowledge in contemporary philosophical discussions. Virtue epistemology seeks to provide an alternative approach to the study of knowledge. Sosa, the earliest proponent of VE, proposes that knowledge should be grounded in the epistemic virtues of the agent: “Here primary justification would apply to intellectual virtues, to stable dispositions for belief acquisition, through their greater contribution toward getting us to the truth. Secondary justification would then attach to particular beliefs in virtue of their source in intellectual virtues or other such justified dispositions” (Sosa 1980: 23). This means that, for Sosa, the primary object of evaluation should be intellectual virtues and vices of an agent, whereas individual belief, the basis of knowledge in traditional epistemology as we have seen previously,

Detachment

71

should be relegated to secondary justification. Abrol Fairweather and Mark Alfano helpfully summarize the shift brought about by virtue epistemology: The essential shift in virtue epistemology is from belief-based epistemic norms to agent-based epistemic norms. The former confers epistemic good-making properties on agents due to the epistemic good-making properties of their beliefs, and the latter confers epistemic good-making properties on beliefs due to the epistemic good-making properties of the agent (their epistemic virtues). The direction of analysis where normative properties of agents confer normative properties on beliefs is essential to virtue epistemology. Some form of this virtuetheoretic direction of analysis is accepted by all virtue epistemologists. Fairweather and Alfano 2017: 8–9

Sosa has been widely acknowledged as someone who stands at the very beginning of such a major shift in epistemology. He is the proponent of what has come to be characterized as virtue reliabilism, which uses the competently successful performance as the criterion to evaluate a subject’s epistemic virtues and vices. The relationship between cognitive competence and successful performance determines the normative status of a performance, i.e. whether a particular performance is knowledge proper or not. In this effort, the nature of competence and the relationship between competence and performance are key to Sosa’s virtue epistemology (Vargas 2016: 3–4). With respect to the nature of competence, Sosa has proposed what he calls the SSS (skill, shape, and situation) framework in order to determine the epistemic competence of an agent;1 for the relationship between competence and performance, Sosa offers the AAA (accurate, adroit, and apt) framework to better understand the various normative relationships between the two.2 One area of dispute in VE is what kind of virtue should count as an epistemic virtue. In this respect, there are two competing camps among contemporary virtue epistemologists. Sosa’s reliabilist virtue epistemology is often cast against what has come to be known as virtue responsibilism, with the former focusing on epistemic competence in perception, memory, and inductive and deductive reasoning, and the latter on the moral character or moral virtues in epistemic pursuits, such as curiosity, open-mindedness, and intellectual courage, etc. John Greco provides a helpful summary of the rationale behind Sosa’s version of intellectual/epistemic virtue: “According to Sosa, an intellectual virtue is a reliable cognitive ability or power. Coherence-seeking reason is thus an intellectual virtue if reliable, but so are perception, memory, and introspection” (Greco 2002: 293). Since reliabilist virtues are cognitive faculties, they are often called faculty-virtues. In the following passage, Sosa lays out his case for touting the faculty-virtues: For example, it may be one’s faculty of sight operating in good light that generates one’s belief in the whiteness and roundness of a facing snowball. Is possession of such a faculty a “virtue”? Not in the narrow Aristotelian sense, of course, since it is no disposition to make deliberate choices. But there is a broader sense of “virtue,” still Greek, in which anything with a function—natural or artificial—does have

72

Ernest Sosa Encountering Chinese Philosophy virtues. The eye does, after all, have its virtues, and so does a knife. And if we include grasping the truth about one’s environment among the proper ends of a human being, then the faculty of sight would seem in a broad sense a virtue in human beings; and if grasping the truth is an intellectual matter then that virtue is also in a straightforward sense an intellectual virtue. Sosa 1991: 271

Clearly, Sosa is using the term virtue in a sense that is broader than its moral usage. Virtue here includes function, ability, power, potency, or competence. Such an employment of the term virtue is precisely what has prompted criticism from philosophers who are more aligned with virtue in its Aristotelian use. For example, Linda Zagzebski, one of the most prominent voices advocating what has come to be characterized as virtue responsibilism, criticizes Sosa’s use of the term virtue in this context: [Sosa] makes no attempt to integrate intellectual virtue into the broader context of a subject’s psychic structure in the way that has been done by many philosophers for the moral virtues. What’s more, Sosa’s examples of intellectual virtues are faculties such as eyesight and memory, which are not virtues at all in traditional virtue theory. It turns out, then, that his plea for a turn to the concept of intellectual virtue actually has little to do with the concept of intellectual virtue as a virtue in the classical sense. Zagzebski 1996: 8–9, original italics

In fact, Zagzebski accuses Sosa of confusing the function of a cognitive faculty with its virtue (Zagzebski 1996: 9, fn. 4). For her, the reliabilist faculty-virtues do not provide any added value to knowledge besides their conduciveness to truth. According to Zagzebski, “knowledge is a more valuable state than true belief. It follows that the value of the knowing state is more than the value of the truth that is thereby possessed. So what knowledge has in addition to true belief has value” (Zagzebski 1996: 301). In other words, it is insufficient for knowledge to be produced by a reliable belief-producing apparatus since as such there would be no additional value to that belief state than its truth. Therefore, Zagzebski defines intellectual virtues as acquired character traits of the agent that consist of two components: they are reliably conducive to truth and they reflect the agent’s virtuous motivation (Zagzebski 1996: 311). By contrast, “for Sosa the intellectual virtues are cognitive abilities rather than character traits, they need not be acquired, and their acquisition and use need not be under one’s control” (Greco 2002: 295). At the heart of Zagzebski’s critique of virtue reliabilism is her questioning of a strong distinction between intellectual and moral virtues, made by Aristotle. In fact, she argues that “[i]ntellectual virtues are best viewed as forms of moral virtue” (Zagzebski 1996: 139). By making intellectual virtues a subset of moral virtues, Zagzebski attempts to integrate intellectual and moral virtues into the agent’s overall psychic structure, something she thinks Sosa’s reliabilism fails to do. Sosa acknowledges the importance of trait-virtues favored by the responsibilists, e.g. curiosity, open-minded, intellectual courage, etc. However, as Sosa sees it, the

Detachment

73

responsibilists do not differentiate the virtues that are constitutive of knowledge from those that are merely facilitative: It is such knowledge-constitutive competences that are of main interest to a Competence Virtue Epistemology aiming to explain human knowledge. Other epistemically important traits—such as open-mindedness, intellectual courage, persistence, and even single-minded obsessiveness—are indeed of interest to a broader epistemology. They are of course worthy of serious study. But they are not in the charmed inner circle for traditional epistemology. They are only “auxiliary” intellectual virtues, by contrast with the “constitutive” intellectual virtues of central interest to virtue reliabilism. Sosa 2015: 43

Put simply, the faculty-virtues belong to the charmed inner circle of traditional epistemology in that they are knowledge-constitutive competences, whereas the trait-virtues are only auxiliary intellectual virtues that can assist the epistemic effort but are not constitutive of such effort. Greco echoes such an assessment when adjudicating the reliabilist/responsibilist virtues in pointing out that a simple case of perceptual knowledge does not have to involve Zagzebski-type intellectual virtues (Greco 2002: 296). However, there are important epistemic virtues that have largely remained outside the purview of the discussions between the reliabilists and the responsibilists. To make the case, I would like to look into how epistemic virtues are conceptualized within a very different intellectual context, namely that of Buddhism, especially Chan Buddhism in China. We will see that the way a core epistemic virtue, detachment, is articulated in Chan Buddhist philosophy does not fit the reliabilist-responsibilist framework; it is a reliabilist trait-virtue which takes on both reliabilist and responsibilist features. This implies that the trait-virtue of detachment does not settle into the auxiliary role within Chan Buddhist epistemology. Rather it is constitutive of a particular kind of knowledge, namely enlightened knowledge in Chan Buddhism. This trait-reliabilism, as opposed to faculty-reliabilism or trait-responsibilism, would complicate the disputation between virtue reliabilists and virtue responsibilists about what should count as a constitutive epistemic virtue and whether a reliabilist virtue can be a character trait instead of a faculty.

3 Detachment as a reliabilist virtue in Buddhist epistemology Many, if not most, traditional Chinese philosophers can probably be characterized as virtue epistemologists in the sense that “normative properties of agents confer normative properties on beliefs” (Fairweather and Alfano 2017: 9). Zhuangzi’s (late fourth-century to early third-century BCE) dictum “there is a true person, and afterwards there is true knowledge” is one of the most famous manifestos of such an orientation. Of course, questions about what is a true person and what is true knowledge in Zhuangzi’s context would immediately arise. However, I will not get into the

74

Ernest Sosa Encountering Chinese Philosophy

Zhuangist ideas of a true person and true knowledge in this chapter. Rather my focus will be on a later development of these ideas in the context of Chan Buddhism. The term “true person” is picked up by Chan Buddhist Linji (㠘☏, d. 866), whose deliberations of true person represent his most vigorous effort to sketch out the relationship between knowledge and agent in his project of enlightenment. Much of the text attributed to Linji, the Linji lu (㠘☏䤴; Recorded Sayings of Linji), is devoted to the portrayal of true persons as well as how to become one. Linji personifies the climax of Chinese Chan Buddhism that has come to define later iconoclastic expressions of Chan in Chinese and other East Asian cultures. He has been revered as the last, arguably the most famous, and certainly the most colorful, Chan patriarch in the “orthodox” Hongzhou ⍚ᐎ lineage during the so-called “golden age” of Chan Buddhism in Tang dynasty (618–907). Linji is legendary for his blasphemous and iconoclastic teachings as well as unconventional teaching methods, such as shouting at his disciples and hitting them with a stick, all of which have now become part of the stock images of Chan enlightenment. Chan Buddhists are privy to the intellectual legacies of both Buddhist and preBuddhist indigenous Chinese traditions.3 While Indian Mahāyāna Buddhist epistemologists, most famously Dignāna (c. 480–540 CE) and Dharmakīrti (c. 600– 660 CE), advocate some versions of coherentism and foundationalism, later Chinese Buddhists, especially Chan Buddhists like Linji, favor virtue epistemology of sort, likely due to influence of the broader orientation of the indigenous Chinese intellectual tradition mentioned earlier. This chapter will examine a character trait valorized in Buddhism, namely detachment, and investigate its place within the Buddhist project of enlightenment by exploring its epistemic qualities, inspired by the contemporary discourse on virtue epistemology. I will construe Linji’s Chan epistemology as one that is primarily geared toward cultivating the epistemic virtue of detachment, constitutive of the enlightened knowledge in Buddhism. More specifically, I will make the case that detachment is both a reliabilist virtue and a trait virtue in Linji’s Chan epistemology. The case will be made in two steps: first, I argue that detachment is reliabilist in nature by situating Linji’s thought within the broader context of Buddhist epistemology; second, I make the case that detachment is a character trait keenly cultivated in Linji’s approach to enlightenment. This section is devoted to the first step while the next section will deal with the second.

3.1 Root ignorance and attachment The primary motivation for Buddhist philosophers is how to overcome ignorance or illusion as the cause of suffering. Due to the critical role of ignorance and illusion in the Buddhist project of enlightenment, the Buddhists have devoted a significant amount of effort to investigating the nature of illusion and its causes. In this respect, they are mainly interested in two kinds of illusions: those due to faulty cognitive faculty or deceptive circumstances and those due to imputation of substance/essence to cognitive objects (empirical or mental). These two kinds can overlap under certain conditions, but they are treated as distinct cases here. Between the two kinds of illusion, the first is

Detachment

75

the more obvious one as it is very much part of our everyday experience, for example when we mistake something (e.g. a rope) for something else (e.g. a snake) in darkness or see something that is not out there, for example due to cataracts in the eyes. Although such illusions are not necessarily easy to deal with, they are relatively easy to understand. However, due to their supposed obviousness as an error—at least in theory, if not in practice—they are often employed as an analogy to make sense of the second kind of illusion, which is far more difficult to appreciate, let alone to overcome. Philosophically, the Buddhists are much more invested in the second kind of illusion. This kind of illusion does not reject the existence of empirical objects. Rather, it challenges our naïve views about the way of their existence. That is, the apparent existence of empirical objects tricks most of us into believing that they are substantive and independent of conditions at their core. In this sense, empirical objects are rather deceptive because they appear to be substantive and really out there on their own, independent of the conditions that make them possible. For the Buddhists, empirical objects are thoroughly dependent on conditions and causes, and outside such conditions and causes there is nothing in and of the objects themselves by way of substance or essence that is irreducible to the conditions and causes. Substance and essence—which are by definition, according to the Buddhists, unconditioned, uncaused, and cannot be reduced to anything else—are in fact illusory since their substantive existence cannot be empirically verified from the Buddhist perspective. In fact, the unconditioned and changeless substance and essence are anti-empirical since change and impermanence define empirical reality. This second kind of ignorance and illusion is what the Buddhists refer to as the root delusion or ignorance. As Dale Wright acutely observes: “Buddhists envision a systematic distortion that pervades all human understanding. Rather than establishing a framework for the discrimination of truth and falsity, Buddhists entertain the possibility that the frameworks we employ for the process of securing truth are themselves subject to the distorting impacts of desire and ignorance” (Wright 1998: 137). Root ignorance or illusion is such a systematic distortion of reality that is embedded in the way our cognitive apparatus is structured and habituated. For the Buddhists, the cause of root delusion or ignorance is the pernicious and ubiquitous attachment that permeates all of our cognitive activities. That is, we impute substance and essence to empirical objects through the mechanism of reification and attachment that is deeply entrenched in our cognitive apparatus. Reification is the culprit that distorts our cognition of the world whereas enlightenment is understood as a qualitatively different cognitive state wherein things and events are perceived without distortion or illusion. However, overcoming such a deeply entrenched reification and attachment in our conceptual and linguistic apparatus is an extraordinarily challenging endeavor and cannot be accomplished by a simple act of will. As Jay Garfield observes: Simply by resolving to abandon attachment one cannot thereby succeed in shedding it. It is difficult to accomplish this. Attachment arises as a consequence of the persistent, pervasive psychological, verbal, and physical habits that together constitute what Buddhist philosophers call the “root delusion,” the ignorance of

76

Ernest Sosa Encountering Chinese Philosophy the true nature of things. That delusion consists in confusing existence with inherent existence and issues inevitably in one of the two extreme views— reification or nihilism. Only through extensive meditation on the nature of phenomena and on the nature of emptiness can these habits be abandoned, and only through an understanding of the ultimate nature of things can the fruit of actions done through abandonment—that is, liberation from the suffering of cyclic existence—be attained. Garfield 1995: 236–237

In order to overcome such a systematic distortion of reification and attachment, the Buddhists propose a solution that trains our mind through vigorous meditation practices. It is through such vigorous meditation that a practitioner’s cognitive competence can be enhanced, enabling her to “see things as they are.”

3.2 Detachment and emptiness “See things as they are,” a famous Buddhist dictum, is to see all forms of existents as impermanent, dependently arisen, and non-substantive. These attributes would be coalesced later into the notion of emptiness in the hands of Mahāyāna Buddhists, which maintains that every existent in the world is empty of self-nature since it is thoroughly dependent on others such that there is nothing in and of itself outside of such dependency. The radical aspect of this vision has to do with the Buddhist recognition that there is a profound disconnect between the emptiness of the world (truth) and our engagement with it (knowledge) that is inevitably mediated by our conceptual, linguistic, and other cognitive apparatus that reify objects by endowing them with essence and substance. If knowledge can be broadly understood as the way to obtain truth, for the Buddhists the apparatus or instrument of knowledge we use in navigating the world is inadequate in achieving the enlightened knowledge that leads to realization of emptiness, the ultimate truth in Mahāyāna Buddhism. This means that for the Buddhists there is a fundamental disconnect between the ultimate truth of the emptiness of the world and our knowledge apparatus that hinders our realization of that truth. We should point out that the Buddhists are not skeptics or agnostics, at least not the radical kind, when it comes to knowledge. As Garfield points out, “there is a sense in which Buddhist epistemologists do want to undermine some of the pretensions of ordinary perceptual knowledge, but they do not want to reject the entire framework of everyday knowledge” (Garfield 2015: 218, original italics). Indeed, one of the challenges for the Buddhists is how to accommodate the everyday knowledge that is necessary for living in the world. Mahāyāna Buddhists often resort to what is known as the two-truth strategy, reserving the ultimate truth for the realization of emptiness and cognitive transparency while relegating truth in the everyday activity to the conventional realm. For the Buddhists, everyday knowledge, in their concrete particularities, is provisional at best. As Garfield explains, as he unpacks the notion of “convention” in the Buddhist philosophical discourse,

Detachment

77

for one thing . . . it depends upon our biology, our conceptual apparatus and our language. For another, it does not deliver reality in a way that withstands analysis, reality as it is independent of how we engage with it. Analysis reveals the properties we take ourselves to register to be mere imputations. And finally, in virtue of this, conception is always deceptive. While it is . . . an instrument of knowledge, it is a second-rate instrument, standing behind perception as a guide to reality, simply in virtue of always presenting itself as engaging with that which is not real. Garfield 2015: 222

This means that for the Buddhists the Gettier-type problems that have shaken up contemporary epistemology will never be solved satisfactorily since they are embedded in the very way our cognitive apparatus, our biology, and our habits operate. Furthermore, such problems are features of the world that is infinitely conditioned and ultimately out of our control, and as a result we can never really know the various possible layers of conditions of a given phenomenon. In this sense, the Buddhists are fallibilists when it comes to everyday or conventional knowledge. However, their attitude toward enlightened knowledge, which leads to the ultimate truth of emptiness, is different. At this juncture we need to take note of the fact that ultimate truth, especially the one formulated by Mahāyāna Buddhists, has very little to do with what are often considered the “big questions” in the history of philosophy and religion. In one of the most famous Buddhist parables, the Buddha remains silent when asked to shed light on questions like whether the universe has a beginning or not, whether it is finite or infinite, what happens to the Buddha after his death, etc. The Buddha brushes aside such questions, using a poisonous arrow parable to illustrate the irrelevance of such questions to the pressing problem of suffering that is central to the Buddhist project. As the Buddha narrates it, if someone is hit by a poisonous arrow, the most pressing need is to get the arrow out of the body, instead of wondering about the person who makes the shot (what that person is dressed in, what class he belongs to, what motivates the shooting, etc.). This is a clear indication that the Buddhists are primarily interested in the kind of knowledge that leads to enlightenment and the overcoming of suffering, not speculative knowledge that does not advance such a goal.

3.3 Mind versus belief What distinguishes the Buddhists from many contemporary philosophers in their epistemological effort is that the Buddhists locate the source of the deception within the structure and activities of the mind whereas many contemporary philosophers locate it outside, often under the guise of a deceptive Cartesian demon. Buddhist enlightened knowledge, what I have called a state of cognitive transparency (Jiang 2014), is free from the Gettier-type problems. Gettier-type problems pose a challenge to any claim of knowledge in a particular setting. Due to the intractable nature of circumstances, our seemingly justified true belief is not necessarily the equivalent of knowledge. However, the challenge posed by Gettier-type problems is not really a

78

Ernest Sosa Encountering Chinese Philosophy

problem for the Buddhists since what Gettier does is to simply add another layer of conditioning, often deceptive and unbeknown to the agent involved. This additional layer of conditioning and unpredictability, in any given particular circumstance, does in no way invalidate the emptiness of any object, whether it is empirically real or illusory. Although contemporary philosophers are still trying to come up with effective ways to deal with the challenges posed by the Gettier problem, the Buddhists have long moved on since they are fundamentally uninterested in the particularity of the intractable, therefore deceptive, empirical reality. What interests them is the deception that is caused by the deluded and ignorant mind. For the Buddhists, the stake about the deceptive empirical world, due to its infinite conditionality, is not as high as the deceptive mind that prevents us from enlightenment. This brings us to another critical distinction between Buddhist theory of knowledge and epistemological theories in contemporary analytic philosophy. That is, for the Mahāyāna Buddhists, knowledge is a mental episode (Stoltz 2007: 404), what I have called “cognitive transparency” (Jiang 2014), which presents the truth about the world, i.e. its emptiness; on the other hand, for most contemporary analytic philosophers knowledge is based on belief (JTB +) and is not a mental episode even though belief is. Furthermore, the separation between justification and truth at the heart of contemporary epistemology is hard for the Buddhists to accept. Even if the Buddhists can be persuaded about the usefulness of this separation in matters pertaining to the conventional or empirical realm, they would not accept it when it comes to enlightened knowledge about the emptiness of all forms of existences because such a separation would mean that there is always a possibility for divergence between truth and justification (Stoltz 2007: 397). For the Buddhists, the challenge about justification is not primarily due to the challenges posted by unpredictable circumstances, e.g. epistemic luck, but rather due to the root delusion or ignorance as we have discussed previously. So what kind of knowledge can get the Buddhists to the truth of emptiness? In this respect, they primarily rely on penetrating cognitions achieved in deep meditative states as that knowledge. In other words, cognitive transparency achieved in deep meditative states is the enlightened knowledge that avails a Buddhist practitioner of the ultimate truth of emptiness. Detachment is a constitutive component of such a cognitive state. Given the perniciousness of reification in almost every aspect of our cognitive activities, overcoming attachment becomes the central task in the Buddhist enlightenment project since detachment is a requirement for the cognitive transparency that leads to the truth of emptiness for the Buddhists. In this respect, it is interesting to note that while early Indian Buddhist epistemologists are primarily concerned with cultivating cognitive transparency, a meditative cognition that penetrates things-events in the realization of their thorough conditionality, there is a subtle but significant shift to the training of a character that cultivates the competence of detachment in the hands of Chinese Chan Buddhists. That is, whereas detachment is treated as concomitant with enlightened cognition in much of the Indian Buddhist tradition, later Chinese Buddhists like Linji regard it more as a character trait required for the embodiment of enlightened knowledge.

Detachment

79

Therefore, in the second step of my argument I will make the case that detachment is a trait virtue, by looking into the aspect of detachment as a character trait portrayed in one of the most important Chan Buddhist texts, the Linji lu.

4 Detachment as a cultivated character trait in the Linji lu4 The problem with attachment in Buddhism is often framed in terms of its objects, i.e., the (illusion of) self, sensuous pleasures, views/dogmas, etc. The reasoning is that attachment to those objects leads to suffering, as the Buddha’s Second Noble Truth points out. However, later Mahāyāna Buddhism has expanded the notion of attachment to the dualistic mechanism of subject/object structure in the apparatus of our everyday cognition. The element of grasping is added to the subject/object cognitive apparatus such that the everyday cognition is no longer merely cognitive, but also reifying. This “cognition + reification” characterizes the later Buddhist approach to cognition that problematizes the embedded reification component in the everyday cognitive activity. In this way, attachment is rearticulated as cognitive reification, not just an emotive (either positive or negative) investment in particular objects by a cognitive agent. Mahāyāna Buddhism promotes the cultivation of what are known as the six perfections: generosity or alms-giving (dāna ᐳᯭ), moral behaviors in accordance with Buddhist precepts (śīla ᤱᡂ), tolerance or forbearance (k·sānti ᗽ䗡), energy or vigor (vīrya ㋮䙢), meditative absorption (dhyāna ⿚ᇊ), and wisdom (prajñā 㡜㤕). Although detachment is not listed as one of the six perfections, it is considered a critical component of wisdom. So what is wisdom within the Buddhist tradition? For Mahāyāna Buddhists, wisdom refers to a specific kind of cognition, the enlightened cognition that is the realization of the empty and thoroughly dependent nature of all forms of existence and has the transformative power to overcome attachment understood to be the root cause of suffering in Buddhism. For the purpose of this chapter, it is critically important that detachment is regarded as constitutive of enlightened cognition or wisdom. Much of the Mahāyāna Buddhist discourse can be understood as directed at overcoming the reification component of the everyday cognition within the cognitive structure described above. What distinguishes the approaches proposed by Chan Buddhists like Linji is that they shift the focus of the Buddhist enlightenment project from enlightened cognition to enlightened character. The central character trait cultivated and celebrated in Linji’s Chan practice is the trait of detachment. Such a trait consists of three components: courage, confidence, and freedom. Let us take a closer look at how detachment is understood in the Linji lu.

4.1 The hurdles to enlightenment in the Linji lu Linji calls an enlightened person “a true person with no rank or position” (wuwei zhenren ❑սⵏӪ). This is someone who enjoys genuine spiritual freedom (ziyou 㠚⭡, 㠚൘, zizai or 䀓㝛 jietuo), unfettered by various traps in both the mundane (fan ࠑ) and sacred (sheng 㚆) realms. In the Linji lu, such a true person is someone who

80

Ernest Sosa Encountering Chinese Philosophy

realizes the ultimate truth of emptiness.5 My focus here is on what Linji considers to be the biggest challenge to enlightenment, namely, entrapments of a practitioner by jing ຳ. Jing is usually translated as circumstances, surroundings, environment, or objects in the Buddhist context, but in the Linji lu it takes on an outsized role, becoming an almost all-inclusive term that encompasses any situation or object, both mundane and sacred, that ensnares a Chan practitioner and prevents one from attaining enlightenment. It can be divided into two broad categories: past and present. Past jing refers to karma; it points to the fact that we are the products of karma and continue to be conditioned by the past (Sasaki trans.: 12). Present jing, which is the focus of the text, refers to the psychophysical constituent of the human existence: The grosser part of you is at the mercy of [the four elements:] earth, water, fire, and wind; the subtler part of you is at the mercy of the four phases: birth, being, decay, and death. Followers of the Way, you must right now apprehend the state in which the four elements [and four phases] are formless, so that you may avoid being buffeted about by jing. Sasaki trans.: 14, with modifications

The four elements of earth, water, fire, and wind are the traditional categories in the Buddhist discourse on the physical world. Here they refer to the constituents of the human body as well as its biological stages from birth to death. In the next passage, Linji expands the four elements to encapsulate mental activities by correlating them with specific mental phenomena: doubt with earth, lust with water, anger with fire, and joy with wind (Sasaki trans.: 14–15). The four elements are expanded to include both the physical and the psychological constituents of human beings. Therefore, jing in the Linji lu refers to both bodily and mental aspects of human existence. To overcome the entanglement by the four elements, a practitioner should strive to see their formlessness, synonymous with the famous Mahāyāna doctrine of emptiness that points to the insubstantiality and the thoroughly conditioned nature of all existence, including human existence. A more serious hurdle problematized in the Linji lu pertains to various forms of spiritual attachment in a Chan practitioner’s practices, i.e. scriptural studies and meditation. With regards to scriptural studies, to be a Chan Buddhist obviously requires one to follow the examples set by the Buddha and the patriarchs as well as their teachings. However, those Buddhist icons and ideals can themselves be reified and become objects of attachment. From Linji’s perspective, committed Chan followers can become slavish to Chan teachings, which is antithetical to the Buddhist project of enlightenment. Linji dismisses reified Buddhist teachings as “the words of some dead old guy” (Sasaki trans.: 27) and ridicules those who are attached to them as “blind idiots” (ibid.). Clearly, for Linji, rote learning and scholastic deftness are inadequate as far as achieving enlightenment is concerned. The cognitive and discursive approach to Buddhist teachings reifies those teachings by turning them into objects to be studied and memorized. Learning in such a fashion might enable a practitioner to engage in sophisticated conceptual games, but those games can become obstacles to reaching

Detachment

81

enlightenment if one is attached to them. Overcoming such an attachment requires cultivating a strong character of detachment.

4.2 Cultivating the character of detachment According to Linji, the key to overcoming such hurdles to enlightenment is to cultivate a strong character of detachment that can withstand our emotional volatility and to train one’s mind to be so agile and detached that it is not ensnared in any state associated with the four elements or attached to revered Buddhist icons. A practitioner with a strong character and a nimble mind is a person of freedom—free to go or stay as one pleases—who does not reify or attach to any of those states and is in the state of formlessness (❑⴨ຳ). Linji devotes much of his teaching to training his disciples how to act spontaneously, rather than how to think things through. For Linji enlightenment is more than enlightened cognition. Rather, it requires an enlightened character of detachment, marked by courage, confidence, and freedom. An enlightened character is one that spontaneously manifests itself in a Chan practitioner’s engagement with the world, especially under challenging circumstances. Such a singular focus on the practitioners’ character is also evident in Linji’s teaching on meditation. The practice of meditation is widely recognized as being central to the Chan project of enlightenment. However, as Linji sees it, the misunderstanding of meditation is rampant among Chan practitioners. Accordingly, many people mistake all the prescribed postures of the seated meditation—sitting down cross-legged with one’s back against a wall, tongue glued to the roof of one’s mouth, completely still and motionless—as the quintessential practice of Chan. He dismisses all of them as misguided as they direct practitioners toward obsessing over the external form (Sasaki trans.: 24–25). For him, Chan practice is about transforming a practitioner’s character, not the particular bodily posture or meditation-induced visions. Linji’s emphasis on the cultivation of a set of forceful character traits through meditation is an interesting contrast with the traditional Buddhist teaching that focuses more on the cognitive aspect. Focusing on the cognitive dimension of meditation in one’s Chan practice, for Linji, can easily lead to the reification of various kinds of meditation-induced visions. Linji sternly warns Chan practitioners of the grave danger posed by various meditationinduced hallucinations (S. māra; C.冄). In certain advanced meditative states, a practitioner can sometimes have a powerful experience of catching a glimpse of Buddhist icons like the Buddha or Chan patriarchs. Given the intensely meaningcharged nature of these icons for a Chan Buddhist, a practitioner can easily mistake such experiences in a meditative state as signs of enlightenment whereas they are actually manifestations of subtler reification and attachment at a more advanced level of the spiritual journey. Clearly, the extraordinarily demanding nature of Chan meditation practice means that it is not for those with a weak character of attachment and slavishness since they can be easily seduced and misled by certain images seen in meditation, especially those of the Buddha or Chan patriarchs, whereas all images should be dismissed as māra. This is critical in cultivating detachment to Buddhist icons that is at the heart of Linji’s

82

Ernest Sosa Encountering Chinese Philosophy

teaching against attachment to Buddhist icons and images, a particularly potent kind of attachment for a committed Chan follower: Someone asked, “What is Buddha-māra?” The master said, “One thought of doubt in your mind is māra. But if you realize that the ten thousand fas never come into being, that mind is like a phantom, that not a speck of dust nor a single thing exists, that there is no place that is not clean and pure—this is Buddha. Thus Buddha and māra are simply two states, one pure, the other impure. “In my view there is no Buddha, no sentient beings, no past, no present. Anything attained was already attained—no time is needed. There is nothing to practice, nothing to realize, nothing to gain, nothing to lose. Throughout all time there is no other fa than this. ‘If one claims there’s a fa surpassing this,’ I say that it’s like a dream, like a phantasm.” This is all I have to teach. Sasaki trans.: 12–13, with modifications

Interestingly, Linji appears to take two conflicting positions on the relationship between Buddha and māra here. In the first paragraph Linji characterizes the Buddha and the demon (māra) as two states of mind, pure and impure, respectively. On the other hand, he dismisses even the Buddha and argues that all is empty in the second paragraph. One way to account for the apparent inconsistency is, following Nāgārjuna’s famous teaching of two truths (Ҽ䄖) widely known to Chinese Buddhists, that the first passage explains Buddha versus māra from the conventional perspective which separates the Buddha from māra, whereas the second passage explains it from the ultimate perspective since both Buddha and māra are conventional constructs (all constructs are conventional) and are ultimately empty. In other words, any image experienced in meditative state is māra and only imagelessness and formlessness is the state of enlightenment wherein all reifications, gross and subtle, are overcome. Such an interpretation is consistent with Linji’s advice to cut off representations of enlightenment, i.e. the Buddha, the patriarchs, and arhats, as well as objects of mundane affection, i.e. parents and kinsmen. The following signature passage cements Linji as the ultimate iconoclast in the Buddhist tradition: Followers of the Way, if you want insight into fa as it is, just don’t be taken in by the deluded views of others. Whatever you encounter, either within or without, slay it at once. On meeting a buddha slay the buddha, on meeting a patriarch slay the patriarch, on meeting an arhat slay the arhat, on meeting your parents slay your parents, on meeting your kinsman slay your kinsman, and you attain emancipation. By not cleaving to things, you freely pass through. Sasaki trans.: 22, with modifications

In other words, attachments to both mundane and spiritual objects need to be overcome in order to attain enlightenment promised in Linji’s Chan teachings. Given the centrality of meditation in Chan practice, misunderstanding meditative experiences is an easy trap to fall into. It is critically important for Chan practitioners to be

Detachment

83

unwavering and resolute in the recognition that true enlightenment is formless and cannot be reified or attached to: “true buddha has no figure, true fa has no form” (Sasaki trans., 20, with modifications). The rather violent rhetoric in the above passage is obviously not to be taken literally, but should be taken as reflective of Linji’s wariness of the seductiveness of meditation-induced hallucinations that can be easily reified and clung to as signs of enlightenment as well as his clear-eyed awareness of the challenge in cultivating the trait of detachment. As I have argued elsewhere: What is central to Linji’s teaching is that true awakening is to transform this very structure of attachment, not just to substitute one set of attached objects for another. An attachment to “spiritual” objects does not, ultimately speaking, make the attachment better, since what is changed is simply the object of attachment while the underlying structure of attachment remains firmly entrenched and intact. Much of Linji’s teaching, as recorded in the Linji Lu, is geared toward helping his devout disciples to transform this structure of attachment. Jiang 2011: 259

To overcome attachment to spiritual ideals and to transform the underlying structure of attachment have to be extraordinarily difficult for Buddhist practitioners since those Buddhist icons and ideals represent the very fabric and structure of the Buddhist spiritual universe that gives meaning to the Buddhist practices. Therefore, to transcend a practitioner’s spiritual attachment and mundane affection requires a strong trait of detachment that can persevere in the course of the inevitably traumatic spiritual transformation, analogous to the overturning of one’s world: “Heaven and earth could turn upside down and he wouldn’t have a doubt; the buddhas of the ten directions could appear before him and he wouldn’t feel an instant of joy; the three hells could suddenly yawn at his feet and he wouldn’t feel an instant of fear” (Sasaki trans.: 20). Here Linji is pointing out that Chan practices are riddled with terrifying as well as seductive experiences wherein one’s established sense of self and the world would be turned upside down. A strong character of detachment can provide a secure anchor for a practitioner to explore perilous aspects of spiritual practices that are unavoidable in one’s spiritual journey. Clearly, the enlightened trait of detachment, characterized by courage, confidence, and freedom, is at the heart of Linji’s project of enlightenment.

5 Conclusion: detachment as a trait-reliabilist virtue In this chapter, I have used Sosa’s virtue epistemology and the debate between virtue reliabilists and virtue responsibilists on epistemic virtues to frame an inquiry of a characteristically Buddhist trait, namely detachment, especially as it is portrayed in the Linji lu. I have argued that detachment can be fruitfully understood as a trait-virtue that is constitutive of the enlightened knowledge in Buddhism, leading to the realization of the truth of emptiness about all forms of existence. As such, detachment is both a reliabilist virtue, in that it is constitutive of enlightened knowledge that leads to the

84

Ernest Sosa Encountering Chinese Philosophy

ultimate truth of emptiness, and a responsibilist virtue, in that it is not a faculty virtue but a trait virtue, encapsulating the components of courage, confidence, and freedom, that needs to be cultivated and vigorously trained in Buddhist practices. If my interpretation of the Buddhist detachment is plausible, it can provide a useful example of a trait-virtue that is reliabilist in nature, therefore blurring the line sharply drawn between reliabilist faculty-virtues versus responsibilist trait-virtues in the contemporary discourse on virtue epistemology.

Notes 1

Sosa often uses the example of driving to explain the SSS structure of competence: “[A] complete competence can be broken down into three components: the relevant Skill, Shape, and Situation. Consider such SSS competences, our concepts of these, and the induced SS and S correlates. Take, for example, our complete driving competence on a certain occasion, including (a) our basic driving skill (retained even when we sleep), along with (b) the shape we are in at the time (awake, sober, and so on), and (c) our situation (seated at the wheel, on a dry road, and so on). Drop the situation and you still have an inner SS competence. Drop both shape and situation and you still have an innermost S competence” (Sosa 2017: 131).

2

3

4

5

Sosa often uses archery as an example to illustrate the AAA structure: “A shot is accurate iff it hits the target. It is adroit iff it is an exercise of competence. It is apt iff it is accurate because adroit” (Sosa 2017: 72, Sosa’s italics). To what extent Chinese Buddhists were exposed to works by Indian Buddhist epistemologists is a complicated historical question that we cannot get into here. I am only claiming in this chapter that Chinese Buddhists were at least privy to the general contours of scholastic debates in India, including works by epistemologists like Dignāna, whose opus, Pramān·a-samuccaya 䳶䟿䄆, was translated into Chinese in 711 by Yi Jing 㗙␘ but was supposedly lost rather quickly. More recent scholarship has started to demonstrate that important epistemological works were produced by Chinese Buddhists, which show striking similarities as well as intriguing differences with their Indian predecessors. Given the intended readership of this chapter and this book, I will not get into those fascinating but complicated historical questions. My following discussion on Linji has utilized materials from an earlier article of mine, “Character Is the Way: The Path to Spiritual Freedom in the Linji Lu” (Jiang 2018), adapted for the current chapter. Interested readers can refer to my 2014 article for a detailed discussion of emptiness (Jiang 2014): “Incommensurability of Two Conceptions of Reality: Dependent Origination and Emptiness in Nāgārjuna’s MMK,” in Philosophy East & West 64(1): 25–48.

References Fairweather, Abrol and Alfano, Mark. eds. 2017. Epistemic Situationism. New York, NY: Oxford University Press.

Detachment Garfield, Jay, trans. and comment. 1995. The Fundamental Wisdom of the Middle Way: Nāgārjuna’s Mūlamadhyamakakārikā. New York, NY: Oxford University Press. Garfield, Jay. 2015. Engaging Buddhism: Why It Matters to Philosophy. New York, NY: Oxford University Press. Greco, John. 2002. “Virtues in Epistemology.” In Paul Moser (ed.), Oxford Handbook of Epistemology. New York, NY: Oxford University Press, pp. 287–315. Jiang, Tao. 2011. “Linji and William James on Mortality : Two Visions of Pragmatism.” In Amy Olberding and Philip J. Ivanhoe (eds), Mortality in Traditional China. Albany, NY: State University of New York Press, pp. 249–269. Jiang, Tao. 2014. “Incommensurability of Two Conceptions of Reality : Dependent Origination and Emptiness in Nāgārjuna’s MMK.” Philosophy East & West 64: 25–48. Jiang, Tao. 2018. “Character Is the Way : The Path to Spiritual Freedom in the Linji Lu.” In Youru Wang and Sandra Wawrytko (eds), Dao Companion to Chinese Buddhist Philosophy. Dordrecht, The Netherlands: Springer, pp. 399–415. Sasaki, Ruth Fuller, trans. 2009. The Record of Linji, edited by Thomas Yūhō Kirchner. Honolulu: University of Hawaii Press. Sosa, Ernest. 1980. “The Raft and the Pyramid: Coherence versus Foundations in the Theory of Knowledge.” Midwest Studies in Philosophy 5: 3–26. Sosa, Ernest. 1991. Knowledge in Perspective: Selected Essays in Epistemology. New York, NY: Cambridge University Press. Sosa, Ernest. 2015. Judgement and Agency. New York, NY: Oxford University Press. Sosa, Ernest. 2017. Epistemology. Princeton, NJ: Princeton University Press. Stoltz, Jonathan. 2007. “Gettier and Factivity in Indo-Tibetan Epistemology.” The Philosophical Quarterly 57: 394–415. Vargas, Miguel, ed. 2016. Performance Epistemology: Foundations and Applications. New York, NY: Oxford University Press. Wright, Dale. 1998. Philosophical Meditations on Zen Buddhism. New York, NY: Cambridge University Press. Zagzebski, Linda. 1996. Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge. New York, NY: Cambridge University Press.

85

86

6

Apt Performance as Unity of Knowledge and Action A Comparative Study on Ernest Sosa’s Virtue Epistemology and Wang Yangming’s Meta-ethics Xiang Huang

1 Introduction One important feature of Ernest Sosa’s virtue epistemology is that it takes performance and performers’ competence as the primary theoretical variable to understand the nature of knowledge. This character is not only unconventional in contemporary epistemology but also in Western philosophy in general. However, it may seem familiar in Chinese philosophy. As Shu-hsien Liu observes, “Chinese philosophy in general and Wang Yangming’s philosophy in particular are existential and practical in nature, very different from Greek cosmology, which is speculative in nature, though they do not lack cosmological implications when they were developed into comprehensive systems of philosophy” (Liu 2009: 406). The starting point for my comparative study on Sosa’s epistemology and Wang Yangming’s (1472–1529) meta-ethics is to explore some consequences of this existential and practical character. The first section argues that the AAA structure proposed by Sosa to model the ontological and normative structure of epistemological practices also underlies Wang’s meta-ethics. Addressing this point may help us clear up some confusion about Wang’s philosophy. The second and the third section aim to show that Wang’s philosophy can help us understand better the unconventional and even revolutionary character of Sosa’s virtue epistemology. In order to do this, first, in the second section, I introduce Wang’s doctrine of the unity of knowledge and action (UKA) and argue that Sosa’s virtue epistemology based on the AAA structure can be considered as a theory of the UKA. This may let us see more clearly why Sosa’s virtue epistemology allows a sort of epistemic luck. And second, in the third section, I explore in what sense the UKA and Wang’s meta-ethics can let us reconsider both the nature of intellectual virtues and the relationship between Sosa’s competence-based virtue epistemology and the virtueresponsibilism.

87

88

Ernest Sosa Encountering Chinese Philosophy

2 The AAA structure and Wang’s meta-ethics In Sosa’s virtue epistemology, a belief or a judgment can be taken as knowledge when it is an apt performance. A performance is apt when it falls under a so-called AAA structure: It is accurate iff it attains its aims. It is adroit iff it is an exercise of competence. It is apt iff it is accurate because of competence.

The content of a belief or a judgment is qualified as knowledge only if the belief and the judgment satisfy the AAA structure. A belief p of an agent S can be considered as knowledge when 1) p is accurate, i.e. it is true; 2) p is adroit, i.e. it manifests S’s epistemic virtue and competence; 3) p is apt, i.e. it is true because of S’s competence (Sosa 2007: 23). For example, S aptly believes and hence knows that her textbook is on the table when it is true that her textbook is on the table, and S’s visual faculty lets her see that her textbook is on the table, and S has the belief of her textbook on the table because she saw that her textbook was on the table. In this sense, animals can also have apt “belief ” and knowledge. A frog can know that its food is approaching when the frog successfully catches a coming fly, and the frog’s sensual capacity lets it detect the fly’s coming, and the frog catches the fly because of the frog’s detecting of the fly. We human beings share with animals this kind of knowledge, but we also have reflective knowledge that animals do not have. We can reflectively consider, when we saw the textbook was on the table, whether our perceptual faculties and environmental conditions were reliable. That is, our reflective capacity can let us aptly believe that our beliefs are apt. As a consequence, there are two kinds of knowledge: animal knowledge based on apt belief simpliciter and reflective knowledge based on apt belief aptly noted (Sosa 2007: 32). Both of them fall under the AAA structure. One important character of Sosa’s virtue epistemology based on the AAA structure is that it is an attempt to understand the nature of knowledge through understanding epistemic actions. It is a methodological character different from that of most other epistemological theories, which usually take beliefs as the primary theoretical variable, but it is more similar to moral theories that usually take action as the primary theoretical variable. In fact, as Sosa has pointed out, the AAA structure can be more general and is applicable to other performances besides epistemic ones (Sosa 2011: 23, n. 1; 2016: 6). In a simple application to a moral performance, we say that: It is accurate iff it attains its aim of being a good deed; It is adroit iff the performer has the competence of doing a good deed; and, It is apt iff it is accurate because of competence. This AAA structure applied to the moral performance can be an interesting resource for reflecting on the normative structure of moral principles. In the present section I continue to use the AAA structure to study the normative and ontological structure of Wang Yangming’s meta-ethics.1 The basic structure of Wang’s meta-ethics is summarized in the so-called “FourSentence Teaching” (sijujiao ഋਕᮉ). Wang put it forward in 1527, two years before his death, and therefore it has been considered as his final view. It is read as the following:

Apt Performance as Unity of Knowledge and Action

89

In the original substance of the mind there is no distinction of good and evil. When the will becomes active, however, such distinction exists. The faculty of innate knowledge is to know good and evil. The investigation of things is to do good and remove evil. Chan 1963b: 243

The four sentences describe four ethical statuses. In order to understand them and their relation with the AAA structure, some technical notions and their logical relations need to be clarified. I will use Wang’s Inquiry on the Great Learning (Daxuewen) to do the task. It is a short treatise written around 1527, almost at the same time as the Four-Sentence Teaching was articulated. In the first sentence, the notion “the original substance of the mind” (xin zhi ti ᗳѻ 億) refers to human nature. Mencius (371–289 BCE) is the first Confucian who explicitly claims that human nature is originally good. Wang accepts this thesis of Mencius’s. In order to prove it, he says that, when anyone with his original substance of the mind sees a child about to fall into a well, he cannot help a feeling of alarm and commiseration.2 This shows that his humanity (ren) forms one body with the child. It may be objected that the child belongs to the same species. Again, when he observes the pitiful cries and frightened appearance of birds and animals about to be slaughtered, he cannot help feeling an “inability to bear” their suffering. This shows that his humanity forms one body with birds and animals. It may be objected that birds and animals are sentient beings as he is. But when he sees plants broken and destroyed, he cannot help a feeling of pity. This shows that his humanity forms one body with plants. It may be said that plants are living things as he is. Yet even when he sees tiles and stones shattered and crushed, he cannot help a feeling of regret. This means that even the mind of the small man necessarily has the humanity that forms one body with all. Such a mind is rooted in his Heavenendowed nature, and is naturally intelligent, clear, and not beclouded. For this reason it is called the “clear character” (mingde). Chan 1963a: 659–660

The paragraph shows that: 1. The original substance of the mind is universally good. 2. It is the beginning of humanity. 3. It is an innate and internal intuition and everyone is capable of having it. The first sentence means that, as the original substance of the mind is universally good, there is no need for anyone who keeps it to make a distinction of good and evil, as what she does is always good. The second sentence identifies the origin of evil: it comes from the will. Says Wang: “what arises from the will may be good or evil, and unless there is a way to make clear the distinction between good and evil, there will be a confusion of truth and untruth” (Chan

90

Ernest Sosa Encountering Chinese Philosophy

1963a: 664). The reason why the will may diverge from the original substance of the mind is that it can be obscured by different sorts of selfish desires: “When [the mind] is aroused by desires and obscured by selfishness, compelled by greed for gain and fear of harm, and stirred by anger, he will destroy things, kill members of his own species, and will do everything. In extreme cases he will even slaughter his own brothers, and the humanity that forms one body will disappear completely” (Chan, 1963a: 660). The third sentence explains the nature of the most basic moral competence. This is a capacity of extending the innate knowledge of the good (zhi liangzhi 㠤㢟⸕). The innate knowledge of the good is the recovery of the original substance of the mind from the stirrings caused by the will. Here, the term “knowledge” does not refer to the knowledge of external objects but to the knowledge of moral principles of right and wrong. Facing the evil brought by the will, having innate knowledge of the good is having the competence of extending it: The extension of knowledge is not what later scholars understand as enriching and widening knowledge. It is simply extending one’s innate knowledge of the good to the utmost. This innate knowledge of the good is what Mencius meant when he said, “The sense of right and wrong is common to all men.” The sense of right and wrong requires no deliberation to know, nor does it depend on learning to function. This is why it is called innate knowledge. It is my nature endowed by Heaven, the original substance of my mind, naturally intelligent, shining, clear, and understanding. Chan 1963a: 664–665

Elsewhere, Wang describes the innate knowledge of the good as “the pure intelligence and clear intuition of the mind (lingming 䵸᰾).”3 It offers us the intuition of “the Principle of Nature” (tianli ཙ⨶), “which is not only the principle of right and wrong but also the principle that naturally extends” (Chan 1963a: 656). The fourth sentence explains the reflective part of moral practices. The reflective part is necessary for moral practices in order to avoid the following situation: when the will brings with it self-deception, our innate faculty can be deficient or obscured. Says Wang: Suppose I do not sincerely love [good] but instead turn away from it, I would then be regarding good as evil and obscuring my innate faculty which knows the good . . . . If I did not sincerely hate [evil] but instead carried it out, I would be regarding evil as good and obscuring my innate faculty which knows evil. In such cases what is supposed to be knowledge is really ignorance. Chan 1963a: 665

Now, the proper way to keep away from the ignorance is “the investigation of things” (gewu Ṭ⢙) suggested by the fourth sentence: A thing is an event. For every emanation of the will there must be an event corresponding to it. The event to which the will is directed is a thing. To investigate

Apt Performance as Unity of Knowledge and Action

91

is to rectify. It is to rectify that which is incorrect so it can return to its original correctness. To rectify that which is not correct is to get rid of evil, and to return to correctness is to do good. This is what is meant by investigation. Chan 1963a: 665–666

The four sentences put together demonstrate the basic structure of Wang’s meta-ethics: human nature is originally good; however, selfish desires may obscure it and bring the evil to us. The moral tasks for anyone who wants to be good consist in extending their innate knowledge of the good to the utmost. Besides, in order to extend efficiently their innate knowledge, they also need to investigate things. For some scholars, Wang’s theory is the most sophisticated and consistent metaethics that Chinese Confucian philosophy has ever developed. However, my task is not to evaluate it here but to relate it to Sosa’s virtue epistemology. The most direct way to establish the relation is in seeing how the AAA structure underlies the Four-Sentence Teaching. According to the first three sentences: A moral performance is accurate iff it is done by a performer who keeps her original substance of the mind; It is adroit iff the performer has the competence of extending her innate knowledge of the good; It is apt iff it is accurate because of the competence, that is, iff the performer’s doing good because of her moral practices of extending her innate knowledge.

The possibility of doing such an apt moral performance lays the foundation of moral practices. Now, what the first three sentences display is the innate part of moral practices. The capacity of extending one’s innate knowledge of the good is an innate and unlearned capacity that anyone is capable of having. However, the function of this capacity can be damaged by selfish desires. In order to prevent or repair the damage, Wang introduces the fourth sentence about the investigation of things as the reflective part of moral practices. The capacity of investigating things is a learned capacity that needs to be cultivated during the processes of moral practices. If the first three sentences correspond to the apt performance simpliciter, the fourth sentence corresponds to the apt performance aptly noted. This self-reflective stage can be reconstructed at the meta-level in the following way: The performer’s reflection on her performance is accurate iff she successfully thinks about it (or rectifies the performance in its original place) through the practice of the investigation of things. It is adroit iff the performer has the capacity of investigating the performance. It is apt iff it is accurate because of the competence, that is, iff the performer rectifies the performance in its original place because she has the capacity of investigating it.

In this sense, the fourth sentence can be understood as a process through which the AAA structure has been applied again at the meta-level.

92

Ernest Sosa Encountering Chinese Philosophy

Seeing that the AAA structure underlies the Four-Sentence Teaching in this way helps us understand more clearly the normative and ontological structure of Wang’s meta-ethics and dispels some doubts on it. Let’s see one example. Wing-Tsit Chan raises the following doubt: “Actually Wang’s theory is entirely subjective and confuses reality with value. It is difficult to accept his version of ko-wu (gewu, the investigation of things), for if the term means to rectify the mind, why should it be ko-wu (to ko things) instead of ko-hsin (to ko the mind)?” (Chan 1963a: 655). From the perspective of the AAA structure, Wang’s version of the investigation of things or ko-wu neither is entirely subjective nor a confusion of reality with value. It is a reflective stage of moral practices different from the mere function of the faculty of innate knowledge, just as in Sosa’s virtue epistemology, the reflective knowledge produced by apt belief aptly noted is different from the apt belief simpliciter. As we have seen, an apt belief that satisfies the three conditions of AAA, the accuracy, the adroitness, and the aptness, can be taken as knowledge. This is a kind of knowledge that human beings share with animals. Under normal conditions, appropriate functioning of the eyes can allow human beings and animals to gain visual knowledge. What human beings do, and animals do not, have is at a second level, where human beings can aptly reflect on the normalcy of conditions and the reliability of their competence. Similar to this epistemic configuration, in Wang’s meta-ethics, a performance that satisfies the three conditions of the AAA structure is an apt moral performance. As we have seen, we can call this kind of moral performance, delivered by extending innate knowledge, apt moral performance simpliciter. It is an innate stage of moral practices. Anyone, including the people who are affected by selfish desires, has the potential to deliver moral performances, as they have the innate faculty which knows the good; e.g. they have the innate capacity of feeling alarm and commiseration when they see a child who is about to fall into a well. However, the selfish desires can obstruct their capacity of extending innate knowledge so gravely that their performances fail to be apt. In order to repair the damage inflicted by selfish desires, the investigation of things as a reflective stage of moral practice is introduced. In this reflective stage, the object of investigation is the moral principles or the Principles of Nature, whose contents are not different from the innate knowledge when it is extended to the utmost. Once found, the moral principles can be used to rectify the incorrect so one’s performance can aptly return to its original correctness. In this sense, the investigation of things is an apt moral performance aptly noted. Seeing the investigation of things in this way, we can say that it is not entirely subjective, as the Principles of Nature in things can be shared among people who have the innate faculty of knowing the good. As a consequence, the investigation of things is a normative moral practice that need not lead to a confusion of reality with value, as Chan suspects.

3 The unity of knowledge and action If the AAA structure helps us understand the normative and the ontological structure of Wang’s meta-ethics as the above section shows, this and the next sections aim to show that Wang’s meta-ethics also helps us see more clearly the revolutionary character

Apt Performance as Unity of Knowledge and Action

93

of Sosa’s virtue epistemology. This character is concerned with the attitude toward the relationship between knowledge and action in the theoretical construction of epistemology. Different from traditional epistemological approaches that concentrate on epistemic statuses of beliefs, Sosa’s virtue epistemology takes actions as the primary theoretical variable. This difference has some profound consequences, and we can see them more clearly with Wang’s intuition of the relationship between knowledge and action. Wang’s doctrine of the unity of knowledge and action (UKA) has been considered one of his major contributions to Chinese philosophy (Chan 1963a: 656). Its basic idea is that knowledge is the beginning of action and action the completion of knowledge. The UKA has been articulated and explained in his conversation with his disciple and brother in law, Xu Ai (1487–1518) around 1517, ten years before the Four-Sentence Teaching. The conversation is included in Instructions for Practical Living, a collection of Wang’s conversations with his pupils, letters, and several short essays. The UKA is expressed in the following way: “I have said that knowledge is the direction for action and action the effort of knowledge, and that knowledge is the beginning of action and action the completion of knowledge. If this is understood, then when only knowledge is mentioned, action is included, and when only action is mentioned, knowledge is included” (Chan 1963a: 669–670). Wang gives the following examples: Therefore the Great Learning points to true knowledge and action for people to see, saying, they are “like loving beautiful colors and hating bad odors.” Seeing beautiful colors appertains to knowledge, while loving beautiful colors appertains to action. However, as soon as one sees that beautiful color, he has already loved it. It is not that he sees it first and then makes up his mind to love it. Smelling a bad odor appertains to knowledge, while hating a bad odor appertains to action. However, as soon as one smells a bad odor, he has already hated it. It is not that he smells it first and then makes up his mind to hate it. A person with his nose stuffed up does not smell the bad odor even if he sees a malodorous object before him, and so he does not hate it. This amounts to not knowing bad odor. Suppose we say that so-and-so knows filial piety and so-and-so knows brotherly respect. They must have actually practiced filial piety and brotherly respect before they can be said to know them. It will not do to say that they know filial piety and brotherly respect simply because they show them in words. Or take one’s knowledge of pain. Only after one has experienced pain can one know pain. The same is true of cold or hunger. How can knowledge and action be separated? Chan 1963a: 669

Unless one is capable of reacting appropriately to a bad odor, one cannot be taken as knowing what a bad odor is. Unless one is capable of reacting appropriately to one’s parents, one cannot be taken as knowing filial piety. This is a normative notion of “knowing” that specifies a necessary condition under which a cognitive status can be taken as a status of knowledge. Knowing something implies appropriate actions toward it. Without having the capacity of doing appropriately toward something, one cannot

94

Ernest Sosa Encountering Chinese Philosophy

be considered as knowing it. One question asked by Xu Ai is why we need the notions of knowledge and of action separately. Wang responds in this way: The reason why the ancients talked about knowledge and action separately is that there are people in the world who are confused and act on impulse without any sense of deliberation or self-examination, and who thus behave only blindly and erroneously. Therefore it is necessary to talk about knowledge to them before their action becomes correct. There are also those who are intellectually vague and undisciplined and think in a vacuum. They are not willing at all to try to practice . . . . They only pursue shadows and echoes, as it were. It is therefore necessary to talk about action to them before their knowledge becomes true. Chan 1963a: 670

The first kind of people do not allow moral principles to lead their actions and hence cannot have accurate performance. The second kind of people do not have enough competence of doing the good and hence cannot have adroit performance. Both of them fail to do apt moral performance. The notions of knowledge and of action are used in order to correct, persuade, and educate them. Wang puts forward the UKA, aiming to differentiate his meta-ethics from the predominant Confucian meta-ethical model at that time, especially the model of Zhu Xi (1130–1200), in which knowledge is prior to action. In Zhu Xi’s model, the empirical mind (renxin Ӫᗳ) from which evils ensue is different from the mind of Way (dioxin 䚃ᗳ) that acts in accordance with moral principles (tianli ཙ⨶). Moral principles come from one Ultimate Principle (taiji ཚᾥ) manifested in different forms underlying and within all things in the universe. A person has both kinds of mind, and in moral practices she needs to cultivate her mind of Way through learning and selfcultivation procedures called “investigation of things” (gewu) and “the extension of knowledge”(zhizhi). As Liu Shu-hsien describes, “in self-discipline, one has to take the gradual approach, investigating things in order to have better understanding of principles up to a certain extent that there is an enlightenment to get hold of the principle, the origin of various principles” (Liu 2009: 401). This is a dualist meta-ethical model in the sense that on the one hand moral principles built on some metaphysical and cosmological bases exist and do not act, and on the other hand the mind of an agent in moral practices needs to search actively for the moral principles through learning and self-cultivation. In this dualist model, one needs to obtain the knowledge of moral principles first and act accordingly later. Contrary to this dualist model, the UKA brings the moral principles back into the mind from the metaphysical and cosmological area. That is why the dualist model is usually called as “learning of principle” (lixue ⨶ᆨ) and Wang’s model as “learning of mind-heart” (xinxue ᗳᆨ). The way of bringing moral principles back to the mind is through recognizing that everyone can have innate knowledge of the good. As we have seen with the first three sentences of the Four-Sentence Teaching, a person’s doing morally accurate due to her competence of extending the innate knowledge of the good leads to her performance being aptly moral. The investigation of things is required only when one’s innate knowledge of the good is obscured by selfish desires and needs

Apt Performance as Unity of Knowledge and Action

95

to be repaired through a reflective process. Contrary to this order, in Zhu Xi’s dualist model, a moral practice starts with a process of investigation of things in order to find and know the moral principles, and only once the moral principles are already known, can they lead people to act morally. In order to relate the UKA to Sosa’s virtue epistemology, there is a point to be clarified. The term “knowledge” that Wang uses in the cited paragraphs refers to the knowledge of moral principles and not to the knowledge of external objects that an epistemological theory tries to understand. Moral practices and epistemic practices are different in many aspects. Different from practical knowledge, theoretical knowledge about some external objects need not necessarily require corresponding actions to them. At least many epistemologists think so. However, we can still relate the UKA to Sosa’s virtue epistemology by taking the UKA as a methodological strategy for theoretical construction. Just as Wang adopts the UKA as a strategy to differentiate his meta-ethics from the orthodox dualist model at the methodological level, when the AAA structure takes performance and performer’s competence as the primary theoretical variable, it also adopts an unconventional methodological strategy to understand the nature of knowledge with a resource from telic characters of actions to which many previous epistemologists fail to pay attention. As we have seen with the AAA structure, once knowledge in general is considered as apt performances or as cognitive achievements of the function of certain intellectual virtues, both knowledge of external objects and knowledge of moral principles share a similar normative and ontological structure of apt performance understood as achievements of the function of certain cognitive competences. It is in this sense that we can take Sosa’s virtue epistemology as a version of UKA. In the remaining space of the chapter, I will explore two consequences of taking Sosa’s virtue epistemology as a version of the UKA. First, in this section, I try to show that the UKA provides an interesting resource to reflect on the relationship between epistemic luck and epistemic competence. Second, in the next section, I argue that the UKA allows us to reconsider the nature of intellectual virtues and the relationship between Sosa’s virtue epistemology and virtue-responsibilism. From the point of view of the UKA, Sosa’s performance epistemology is quite revolutionary in the contemporary epistemological theories. This can be shown by Sosa’s attitude toward the question of epistemic luck. The UKA requires a kind of antiluck condition for both moral and epistemic practices, according to which the success of a practice should be attributable to the performer’s competence, so that the accuracy of the performance is a result of responsible efforts rather than mere luck. A deed done by mere luck is an irresponsible deed even though it obtains the same result of a deed that has been morally or epistemologically done. Let’s call this anti-luck requirement “the responsibility anti-luck requirement.” Both Wang’s meta-ethics and Sosa’s performance epistemology comply with this requirement. For Wang, failing to satisfy the responsibility anti-luck requirement is mainly due to the separation between knowledge and action. In the paragraph cited above about why we need the notion of knowledge and the notion of action, Wang distinguishes two kinds of people that lack the competence to act responsibly. The first are people “who are confused and act on impulse without any sense of deliberation or

96

Ernest Sosa Encountering Chinese Philosophy

self-examination.” These people’s innate knowledge of the good has been obscured by their selfish desires, and if they do a good deed, they do it blindly or luckily. They lack the knowledge of moral principles, and according to Wang, they need to obtain the moral principle through the extension of innate knowledge and the investigation of things. The second kind of people may know the literal meanings of moral principles but lack the motivation to put them into practice.4 If they happen to do a good deed, they do it by luck. Both kinds of people lack the competence of doing both responsibly and correctly. The remedy resides in putting the UKA in practices. These two kinds of irresponsible situations in moral practices can easily find analogues in epistemic practices. The first kind of people are those who, for example, give the correct answer in an examination with a lucky guess. An example of the second kind of people could be someone who has studied a textbook of swimming but has never put herself into water. If Wang gives us examples of the violation of the responsibility requirement, Sosa gives a positive safety condition of it in his principle C: C. For any correct belief that p, the correctness of that belief is attributable to a competence only if it derives from the exercise of that competence in appropriate conditions for its exercise, and that exercise in those conditions would not then too easily have issued a false belief. Sosa 2007: 33

This is a principle that specifies under what condition a correctness of a belief can be attributable to the corresponding competence or to the adroitness. The last phrase of the principle is a safety condition, which claims that if the competence had been exercised in appropriate conditions it would not easily lead to false beliefs. In this sense, when the competence is exercised in appropriate conditions, reaching the correct result will not be an accident. This safety condition is precisely what the responsibility anti-luck requirement wants. Let me paraphrase Principle C with Wang’s terminology for moral practices: C . For any morally right conduct c, the correctness of c is attributable to a competence only if it derives from the exercise of that competence (which consists in the extension of innate knowledge of the good, and the investigation of things) in appropriate conditions for its exercise, and that exercise in those conditions would not then too easily have led to morally wrong conducts.

It is not difficult to prove that the Four-Sentence Teaching satisfies C , as we have seen that the AAA structure underlies it, and the AAA structure certainly satisfies Principle C. The satisfaction of C will eliminate the lucky or irresponsible situations caused by the two kinds of people who fail to put the UKA into practice. The first kind of people lack the competence of extending the innate knowledge of the good and of investigating things. The second kind of people lack the motivation and resource to exercise their competence. In this sense, the responsibility anti-luck requirement guaranteed by the safety condition expressed in the last phrase of Principle C is what Sosa’s performance epistemology and Wang’s meta-ethics both are committed to.

Apt Performance as Unity of Knowledge and Action

97

However, some contemporary anti-luck theories tend to require a safety condition stronger than the satisfaction of the responsibility anti-luck requirement. More specifically, they require that the appropriateness of the exercise condition of a competence should be safe and no accident. That is, they require that when a competence is exercised, the conditions in which it is exercised would not easily be inappropriate. For example, Duncan Pritchard argues that the safety condition articulated in Principle C avoids only the type of luck that intervenes between ability and success, such as Gettier-style luck. But there is another type of luck called “environmental luck” that Principle C fails to deal with. It is a kind of factor that ensures that the agent could very easily have not been successful without intervening between her ability and success. Pritchard gives the following example: Suppose . . . that the archer chose her target at random from a range of targets on the range but that, unbeknownst to her, all of the targets bar the one that she actually chose contain a force field that repels anything that goes near it. As with the Gettier-style case . . . then, in which two freak gusts of wind interfere with the shot, the agent could very easily have missed. Crucially, however, environmental luck of this shot seems to in no way undermine the aptness of the shot. Pritchard [2009]2012: 121

In this case, if the archer’s shot hits the bull’s-eye, we still consider the accurate shooting as an apt performance, as it is a result of the archer’s competence, even though the appropriateness of the shooting environment depends on a twist of luck. With regard to this sort of environmental luck, Sosa explicitly allows it in the AAA structure (Sosa 2007: 121). Wang certainly also allows this kind of environmental luck, as it is well known from the beginning of Confucian philosophy. In a recent study, Sean Drysdale Walsh shows that Confucius, just as Aristotle, allowed luck in the development, maintenance, and exercise of moral virtues (Walsh 2017).5 A typical environmental luck that Confucian philosophy deals with is fate (ming ભ). As Chan observes, there is a Confucian doctrine of fate articulated by Mencius and firmly maintained in the history of Confucian philosophy. According to the doctrine, “man should exert his utmost in moral endeavor and leave whatever is beyond our control to fate. It frankly admits that there are things beyond our control but that is no reason why one should relax in his moral endeavor” (Chan 1963a: 79). Doing something morally right necessarily depends on a variety of environmental lucks. To be able to do something good, one should be lucky enough to be alive, the environment should be lucky enough to be appropriate, the objects involved in moral practices should be lucky enough to be accessible, etc. These are things beyond our control. Even though these environmental lucks always accompany everyone’s moral practices, one still can be considered to accomplish satisfactorily her practices as long as her performances are apt. The AAA structure that underlies Sosa’s performance epistemology and Wang’s meta-ethics do not give any more resource beyond the responsibility safety condition to eliminate this sort of environmental luck. The Confucian doctrine of fate gives us a good reason why a meta-ethical theory need not give a normative condition for eliminating it when it is out of our control. The reason can be extended to the area of empirical knowledge. Our

98

Ernest Sosa Encountering Chinese Philosophy

knowledge based on empirical evidence is defeasible. We are lucky enough that our perceptual faculties usually are reliable though always fallible. We are also lucky enough that the state of affairs supported by our empirical evidence is sufficiently durable though always changeable. This is the cognitive situation for human beings, including the most intelligent. In this situation, the AAA structure with the responsibility safety condition is sufficient to characterize empirical knowledge, without intending to resort to some externalist safety condition to eliminate the environmental luck.6 This is an important difference between Sosa’s virtue epistemology and Pritchard’s anti-luck virtue epistemology. At this point, the doctrine of fate adopted by Wang’s meta-ethics helps us to see why the former is preferable.

4 On intellectual virtues It is widely known that there are two kinds of virtue epistemology corresponding to different ways of understanding the intellectual virtues. The faculty-based or reliabilist approach conceives of intellectual virtues as reliable agential faculties like vision, hearing, memory, introspection, and reason. The character-based or responsibilist approach argues that intellectual virtues are certain states of character like openmindedness, fair-mindedness, inquisitiveness, attentiveness, carefulness, thoroughness in inquiry, intellectual courage, honesty, etc. Sosa is the leader of virtue reliabilism. He proposes that the two approaches can be related. According to his proposal, reliabilist intellectual virtues are competences that constitute knowledge, and, “a competence can constitute knowledge only if it is a disposition to believe correctly, one that can then be manifest in the correctness of a belief ” (Sosa and Baehr 2015: 69). These virtues belong to “the charmed inner circle for traditional epistemology.” Character virtues are located outside this inner circle, and they manifest dispositions that reliably put one in a position to know. As they cannot constitute knowledge, “we can best understand the responsibilist, character-based intellectual virtues highlighted by responsibilists as auxiliary to the virtues that are a special case of reliable competence intellectual virtue” (Sosa and Baehr 2015: 63). This way of accommodating virtue-reponsibilism into virtue reliabilism is controversial. Jason Baehr contends that the division between the constitutive facultybased virtues and the auxiliary character-based virtues is far from clear, and that some character virtues do constitute knowledge. For example, if we try to reflect in the second order about a first-order judgment, we certainly need to avoid intellectual vices like hastiness, shallowness, superficiality, provincialism, cowardice, etc. In this case, intellectual characteristics like honesty, fairness, carefulness, thoroughness, openness, courage, etc., must be constitutive for having reflective knowledge (Sosa and Baehr 2015: 83). What I want to argue here is that the UKA provides to us with a new perspective from which to see this controversy between Sosa and Baehr over the relationship between the two approaches of virtue epistemology. In order to see this, let us first consider what Wang thinks about the moral virtues. In Wang’s meta-ethics, having innate knowledge of the good is a most fundamental virtue. We ordinary people can

Apt Performance as Unity of Knowledge and Action

99

have it. But, due to our selfish desires, we need to extend our innate faculty through concrete moral practices. As we have seen, as an example to support the UKA, Wang claims that one who knows filial piety and one who knows brotherly respect must have actually practiced filial piety and brotherly respect before they can be said to know them. Now, what are moral virtues in the practices of filial piety according to Wang? We can see this through Wang’s response to a question asked by his disciple Xu Ai: Take . . . the matter of serving one’s parents. The filial son is to care for their comfort both in winter and summer, and to inquire after their health every morning and evening. These things involve many actual details. Should we not endeavor to investigate them? The Teacher [Wang] said, “Why not endeavor to investigate them? The main thing is to have a basis. The main thing is to endeavor to investigate them by ridding the mind of selfish human desires and preserving the Principle of Nature. For instance, to investigate the provision of warmth for parents in the winter is none other than the extension of the filial piety of this mind to the utmost, for fear that a trifle of human selfish desires might creep in, and to investigate the provision of coolness for parents in the summer is none other than the extension of the filial piety of this mind to the utmost, for fear that a trifle of selfish human desires might creep in. It is merely to investigate this mind. If the mind is free from selfish human desires and has become completely identical with the Principle of Nature, and if it is the mind that is sincere in its filial piety to parents, then in the winter it will naturally think of the cold of parents and seek a way to provide warmth for them, and in the summer it will naturally think of the heat of parents and seek a way to provide coolness for them. These are all offshoots of the mind that is sincere in its filial piety. Nevertheless, there must first be such a mind before there can be these offshoots. Compared to the tree, the mind with sincere filial piety is the root, whereas the offshoots are the leaves and branches. There must first be roots before there can be leaves and branches.” Chan, 1963a: 667–668

Some interesting points relevant to our discussion can be taken from this paragraph: 1. Moral virtues in the practices of filial piety include caring for the provision of warmth for parents in the winter and caring for the provision of coolness for parents in the summer. They are dispositions whose exercises manifest the actions of filial piety. We cannot say anyone who is not capable of exercising these sorts of virtues knows filial piety. 2. When selfish desires creep in, we need to investigate these virtues and keep the innate knowledge of the good that underlies filial piety. That means that these virtues often need reflective cultivation. 3. The innate knowledge of the good that underlies filial piety is like the root from which these virtues grow like branches and leaves. From this root, different sorts of branches and leaves grow up in different domains of moral practices. The innate knowledge of the good that underlies brotherly respect grows virtues in respect to

100

Ernest Sosa Encountering Chinese Philosophy

the practices of brotherly relations. In Wang’s philosophy, the relationship between the root (the innate knowledge of the good) and the branches/leaves is the relationship between his meta-ethics articulated by the Four-Sentence Teaching (hence by the AAA structure) on the one hand, and the concrete moral practices on the other. What interests us the most is this relationship between the root and the branches/ leaves. It gives us a new model to understand both the relationship between Sosa’s virtue epistemology and different sorts of intellectual virtues, and the relationship between virtue reliabilism and virtue responsibilism. From the point of view of this model, the root is the virtue epistemology based on the AAA structure, and the branches and leaves are different sorts of intellectual virtues functioning as norms that guide concrete epistemic practices. Now, questions with regard to the nature of these intellectual virtues can be asked. Are they faculty-based or character-based virtues, or both? If both, what is the relationship between them? More specifically, are facultybased virtues necessarily knowledge constitutive and character-based virtues necessarily knowledge auxiliary, as Sosa claims and Baehr denies? In order to answer these questions, we need to realize the importance of the contextualist character of the root and branches/leaves model. According to this character, in different domains of epistemic practices, from the same metaepistemological root based on the AAA structure, different sorts of intellectual virtues as branches and leaves can be developed. Let’s look at the example given by Sosa. Suppose a mysterious closed box lies before us, and we wonder what it contains. One direct way to unravel the mystery is to open the lid. In this case, faculty-based virtues like perception certainly play a constitutive role for having apt belief, as Sosa correctly points out (Sosa and Baehr 2015: 68). However, in a slightly different practice, the situation can change. During security inspections at airports or train stations, the workers need to check passengers’ luggage. The suitcases and bags are sent to the X-ray detection equipment, and the workers check their contents without opening them but through seeing the detector’s screen, trying to find out from different item images shown on the screen the illicit items such as knives, weapons, lighters, etc. In the checking process, the workers need to pay attention constantly to the images of moving items appearing on the screen trying to distinguish among them the illicit items. In this case, both the faculty virtues like visual perception and the character virtues like attentiveness of the workers play the constitutive role for having apt beliefs. Also, the machine’s capacity and its reliability also contribute to having apt beliefs. Now, let’s continue to suppose that some X-ray security inspection machines have installed an artificial intelligent program that can automatically identify the images of illicit items. Once an illicit item is identified, its location is marked with a colored arrow on the screen, and the alarm bell buzzes. What the workers need to do is, on hearing the alarm, go to pick out the luggage already marked by the colored arrow and turn off the alarm. In this case, the machine’s capacity becomes more prominently constitutive for having the apt belief of the luggage’s content. This example shows that different epistemic practices may require different sorts of intellectual virtues and different combinations of these virtues. In a determinate

Apt Performance as Unity of Knowledge and Action

101

epistemic practice, we can identify which intellectual virtues are constitutive for having apt beliefs and which play only an auxiliary role. But when the context has been modified, some sorts of intellectual virtues that are auxiliary in one practice can be constitutive in other practices. In this sense, the root and the branches/leaves model supports Baehr’s claim that some character virtues can constitute knowledge in certain epistemic contexts. Besides, character virtues such as attentiveness, open-mindedness, inquisitiveness, etc. need not play a normative role in all epistemic practices. As Sosa shows, someone’s obsessive pursuit of truth, at the cost of malnourishment and depression, should not be considered as epistemic competence on many occasions. Terms such as “attentiveness”, “inquisitiveness” etc. refer to different cognitive statuses. In some epistemic practices these statuses are cognitive excellences, while in other practices they can be irrelevant, and in some extreme situations they can even be intellectual vices. The contextualist character of the root and the branches/leaves model suggests a way alternative to that of Sosa and of Baehr, if not totally incompatible with them, to understand the nature of intellectual virtues: it is the concrete epistemic practice that determines which sort of virtues constitute knowledge, just as the concrete moral practice of caring for parents determines the virtues of filial piety. Corresponding to a particular epistemic practice, some cognitive systems of excellent dispositions or competences can be identified as intellectual virtues. These intellectual virtues are of different sorts. They can be faculty-based or character-based virtues. Some of them can be reliabilist prosthetic instruments like telescopes, eyeglasses, X-ray security inspection equipment, electron-positron colliders, etc. Some of them can be reliabilist social resources such as collaborative deliberation, information sharing, etc. In a determinate epistemic practice, some of these intellectual virtues and their combinations can be constitutive, but in other epistemic practices, they need not be. That is, no intellectual virtue can always be knowledge constitutive in all epistemic practices, even though in a determinate epistemic practice we can always determine some virtues are constitutive and some are auxiliary. What all epistemic practices share is the AAA structure, the root that can explain different sorts of intellectual virtues as the branches and leaves in different practices. The root and the branches/leaves model has at least two methodological advantages worth mentioning here. First, without losing the basic insight of the AAA structure, it opens up the possibility of including a broader range of recent developments of different versions of virtue epistemology. Taking cognitive systems composed by different sorts of intellectual virtues and their different combinations as the branches and leaves growing up in different epistemic practices allows the approaches that emphasize not only the individual, but also the material and social resources, and hence the so-called epistemic situationism can be related to Sosa’s performance epistemology through the root and branches/leaves model (Turri 2017: sect. 9). In this sense, Jonathan Kvanvig’s classic study on how intellectual virtues develop in social contexts can also be considered as a way to investigate how some branches and leaves grow up (Kvanvig 1992). Incorporating the material and social components into this boarder sense of intellectual virtues will give Sosa’s performance epistemology more resources to deal with the cases in which knowledge credits go beyond the individual

102

Ernest Sosa Encountering Chinese Philosophy

agential virtues, such as testimonial knowledge, knowledge obtained from reliable equipment, etc. Second, the root and the branches/leaves model can provide a way to connect epistemology and the philosophy of science. In the first half of the twentieth century, philosophers of science such as Carnap, Popper, Sellars, Lakatos, etc. considered their own studies also to be epistemological studies. But after knowledge being understood as justified true belief, epistemology and the philosophy of science began to talk about different issues. Whereas epistemology focuses largely on the epistemic condition under which a piece of true belief converts into knowledge, philosophy of science investigates epistemological, methodological, and ontological questions of science. Now, the root and the branches/leaves relation can be used to relate virtue epistemology based on the AAA structure as the root on the one hand and the epistemology of science as branches and leaves on the other. One possible objection to our way to understand the nature of intellectual virtues is the following. One reason why the faculty-based virtues are considered as knowledge constitutive and the character-based virtues as knowledge auxiliary is that the former are in the charmed inner circle of traditional epistemology and the latter are not (Sosa and Baehr 2015: 69). To this objection I have two observations. First, it is a fact that the faculty virtues like perception, memory, reason, etc. have been discussed frequently in traditional epistemology as basic sources of knowledge, and the character virtues have been comparatively less discussed. However, this fact does not lead to the conclusion that the character virtues cannot be knowledge constitutive, if we define knowledge constitutive as dispositions manifest in the correctness of beliefs. As we have seen, non-faculty virtues can manifest and explain saliently the correctness of beliefs. Second, taking Sosa’s virtue epistemology as a meta-theory of knowledge committed to the UKA allows us to reconsider its relationship with traditional epistemology. It is true that when Sosa introduced the notion of intellectual virtue into analytic epistemology, he tried to use it to resolve many classic problems, including the justification conditions, the Gettier problem, the skeptic problems, etc. It is an approach that already diverges from previous approaches that are belief-based. As Heather Battaly characterizes, “in belief-based epistemology, beliefs are the primary objects of epistemic evaluation, and knowledge and justification, which are evaluation of beliefs, are the fundamental concepts and properties in epistemology” (Battaly [2008]2012: 4). In contrast, in virtue epistemology, agents’ competences and volitional epistemic characters rather than beliefs become the primary objects of epistemic evaluation. When the AAA structure is put forward, it becomes clearer that appropriate actions and their telic aspects should be taken as the primary object of epistemic evaluation. The methodological modification in the meta-level of epistemic practices may suggest new epistemological problems different from those of the traditional epistemology. That is why some virtue epistemologists claim that directions and inquiries of virtue epistemology should be “largely independent of traditional concerns about the nature, limits, and source of knowledge” (Baehr, [2008]2012: 36). I do not think that the problems of traditional epistemology are not worth investigating. It is important to diagnose why traditional epistemology produces the problems and why it lacks resources to solve them. However, it is also important to see that when the AAA structure takes appropriate actions as the primary object of epistemic evaluation, it

Apt Performance as Unity of Knowledge and Action

103

already suggests a new, if not totally different, direction of inquiry. In consequence, the inner circle of traditional epistemology need not be considered an important constraint on settling the inner circle of virtue epistemology.

5 Closing remarks Through this comparative study, I have drawn two conclusions. First, the AAA structure is a power telic theory that articulates the normative structures of both epistemic and moral practices. Understanding of this telic normativity helps us to see more clearly the ontological configuration of Wang’s meta-ethics. Second, taking telic notions, such as competence, aptness, etc., as theoretical variables, Sosa’s virtue epistemology is unconventional and revolutionary in comparison with many traditional epistemological approaches, and Wang’s notion of the UKA can help us see this more clearly. This second conclusion is more controversial. Sosa will not deny that his virtue epistemology is unconventional as it does develops an approach alternative to other ones such as indirect realism, the knowledge-first approach, contextualism, etc. However, he does not think it is as revolutionary as I think. He insists that his virtue epistemology still keeps the gnoseological conscientiousness that aims to advance a satisfactory theory of knowledge, capable of not only explaining the nature and the scope of knowledge but also responding to the skeptical challenges. It is according to this conscientiousness that the charmed inner circle for traditional epistemology is formed. I agree that this gnoseological conscientiousness has been the basic driving force for the development of both epistemology and philosophy in general in the West, and that Sosa’s virtue epistemology represents a solid achievement of it. It is certainly legitimate to keep on holding this conscientiousness for future investigations. However, I still have two observations about this point. First, the gnoseological conscientiousness does not give a free ride to the telic virtue epistemology. Many epistemological approaches such as foundationalism, coherentism, infinitism, knowledge-analysis theories based on a sensitivity condition or on a safety condition, the knowledge-first approach, contextualism, etc., do not take competence, aptness, and other telic resources as the theoretical variables, though they seriously take for granted this conscientiousness. Recently, Sosa pointed out that Descartes can be interpreted as a virtue reliabilist (Sosa 2017: ch. 1). There is no doubt that this interpretation helps to prompt a renewed push on telic understanding of the normativity of epistemology. However, it is still an interesting question why, in the history of the gnoseological development, the telic character of epistemic faculty in Descartes’ epistemology has not been well noticed, and what has been highlighted is the rationalist character of it as something contrasted with, or opposed to, or independent of, empiricist resources. Different forms of non-telic understandings of epistemological normativity have been quite common for the gnoseological conscientiousness. The existence of so many belief-based and non-telic epistemological approaches is one of the reasons for thinking that Sosa’s telic virtue epistemology can be considered as unconventional and even revolutionary.

104

Ernest Sosa Encountering Chinese Philosophy

Second, even though it is a historical fact that epistemological approaches based on the gnoseological conscientiousness have achieved quite profound and systematic understandings about the nature of knowledge and about the strategies to respond to the Skepticism, this does not mean that the inner circle of traditional epistemology cannot be modified, as the scope and the focus of the gnoseological conscientiousness are historically formed. New conceptual resources introduced in different historical moments, such as rationalism vs. empiricism, foundationalism vs. coherentism, internalism vs. externalism, contextualism, sensitivity or safety condition, knowledge first, etc., have expanded the scope and even changed the focus of the gnoseological conscientiousness in different ways. The shift from the non-telic understanding of epistemological normativity to the telic one through Sosa’s virtue epistemology is one important effort of this kind of expansion. In last the two sections I have intended to explore some consequences of this effort with Wang’s notion of the UKA, and I find that this effort can be quite revolutionary.

Notes 1 2 3 4

5

6

For a short introduction to Wang Yangming’s philosophy in general, see Liu (2009), Chan (1963a: 654–658). The example of the child about to fall into a well originally comes from Mencius 2A6, showing the idea that all men are capable of having humanity (Chan 1963a: 65). Chan’s expression is “the pure intelligence and clear consciousness of the mind” (Chan 1963a: 656). But, “intuition” is a better translation than “consciousness.” This second kind of people are in the position that Yong Huang takes in his chapter (Chapter 12) of what creates the third Platonic problem. Here I have no space to deal with the problem in detail. But I think we can read the paragraph cited above (cited also by Yong Huang) in the following way: for Wang, this kind of people fail to deliver an apt performance when they lack the motivation to do what they are convinced to do. Wang’s metaphorical descriptions of their performance as being “intellectually vague and undisciplined,” thinking “in a vacuum,” and pursuing “shadows and echoes,” etc. certainly imply that this performance cannot be apt. Their failure of doing aptly is due to the loss of their innate knowledge of the good (or “the moral knowledge” in Yong Huang’s translation), hence they need to recover or extend their innate knowledge of the good through learning and self-cultivation. Walsh claims that Confucius also allows luck in acting well despite lacking virtue. This kind of luck, if existing, has nothing to do with the environmental luck that we are talking about, as the acting well without virtue cannot be an apt performance, and hence cannot be a morally responsible action. The example that Walsh gives is Confucius’s political ideal in which a government should rectify or correct the lowest people who lack virtue to do good (Walsh 2013: 116). This example is questionable, as what the Confucian politics emphasizes is not the external coercion from the government to force people to do what they do not want to do, but the education and persuasion needed to make people become virtuous. Sosa is not the only philosopher to see this. See the following paragraph of John McDowell: “What the intuition suggests is that we conceive knowledge of the right kind of truth as a sort of continuant. With fully eternal truths, such a conception is

Apt Performance as Unity of Knowledge and Action

105

unproblematic. But the intuition indicates that we extend the conception, more interestingly, to knowledge of changeable, though reasonably durable, states of affairs. Like a living thing, such knowledge needs something analogous to nutrition from time to time, in the shape of intermittent confirmation that the state of affairs known to obtain does still obtain. But the persistence of knowledge does not need the constant operation of a sustaining cause; between the intermittent confirmations, we allow a kind of inertia to operate in the dynamics of epistemic life. If someone counts, as some time, as having a state of affairs of the right kind within his cognitive grasp, say by seeing that things are thus and so, we allow that that epistemic status can outlast the original mode of access to the known fact” (McDowell 1998: 426). Here, the subject who has a state of affairs of the right kind within his cognitive grasp gets an apt belief, and the inertia to operate in the dynamics of his epistemic life necessarily carries with it different sorts of environmental luck.

References Baehr, Jason. [2008]2012. “Four Varieties of Character-Based Virtue Epistemology.” The Southern Journal of Philosophy 46. Reprinted in John Greco and John Turri (eds). 2012. Virtue Epistemology—Contemporary Readings. New York: The MIT Press, pp. 33–69. Battaly, Heather. [2008]2012. “Virtue Epistemology.” Philosophy Compass 3. Reprinted in John Greco and John Turri (eds). 2012. Virtue Epistemology—Contemporary Readings. New York: The MIT Press, pp. 3–32. Chan, Wing-tsit (trans. and comp.) 1963a. A Source Book in Chinese Philosophy. Princeton NJ: Princeton University Press. Chan, Wing-tsit (trans.) 1963b. Instructions for Practical Learning and Other NeoConfucian Writings of Wang Yang-ming. New York: Columbia University Press. Kvanvig, Jonathan K. 1992. The Intellectual Virtues and the Life of the Mind. Savage, MD: Rowman and Littlefield. Liu, Shu-hsien. 2009. “Neo-Confucianism (II): From Lu Liu-Yuan to Wang Yang-Ming.” In Bo Mou (ed.), History of Chinese Philosophy. London and New York: Routledge, pp. 396–428. McDowell, John. 1998. “Knowledge by Hearsay.” In Meaning, Knowledge, and Reality, Harvard University Press, pp. 414–444. Pritchard, Duncan. [2009]2012. “Apt Performance and Epistemic Value.” Philosophical Studies 143. Reprinted in John Greco and John Turri (eds). 2012. Virtue Epistemology— Contemporary Readings. New York: The MIT Press pp. 117–129. Sosa, Ernest. 2007. A Virtue Epistemology—Apt Belief Reflective Knowledge, vol. I. Oxford: Clarendon Press. Sosa, Ernest. 2011. Knowing Full Well. Princeton and Oxford: Princeton University Press. Sosa, Ernest. 2016. “Knowledge as Action.” In Mi Chienkuo, Michael Slote, and Ernest Sosa (eds), Moral and Intellectual Virtues in Western and Chinese Philosophy—The Turn Toward Virtue. New York and London: Routledge, pp. 5–15. Sosa, Ernest. 2017. Epistemology. Princeton and Oxford: Princeton University Press. Sosa, Ernest and Jason Baehr. 2015. “How Are Virtues and Knowledge Related?” In Mark Alfano (ed.), Current Controversies in Virtue Theory. New York and London: Routledge, pp. 61–90. Turri, John. 2017. “Virtue Epistemology.” Stanford Encyclopedia of Philosophy, https:// plato.stanford.edu/entries/epistemology-virtue/

106

7

The Hearer’s Conditions for Accepting Testimony Winnie Sung

1 Sosa’s account of testimony In “Knowledge: Instrumental and Testimonial,” Sosa (2010) argues that the kind of knowledge one gains from testimony is the same kind of knowledge that one gains from relying on instruments such as GPS devices and cellular phones. Both instruments and testimony are sources that reliably deliver a true proposition. As Sosa explains: “A deliverance of a proposition by an instrument is epistemically reliable only if that proposition belongs to a field, and that instrument is so constituted and situated, that not easily would it then deliver any falsehood in that field” (Sosa 2010: 219). Instruments are constituted and situated in such a way that they would not easily deliver false propositions. Hence, one who consults instruments can gain knowledge by believing the outputs the instruments yield. Similarly, people who provide testimony would not easily deliver false propositions. A hearer can gain knowledge by believing what the speaker says. In Sosa’s words, Interpretive knowledge of what a speaker thinks (says) is thus instrumental knowledge that uses the instrument of language. Language is a double-sided instrument serving both speaker and audience. Hearers rely on the systematic safety of the relevant deliverances. Not easily would the speaker’s utterance deliver that the speaker thinks (says) that such and such without the speaker’s indeed thinking (saying) that such and such. Sosa 2010: 135

Sosa also maintains that instrumental knowledge cannot be reduced to testimonial knowledge. Instead, “[h]uman testimony stands with the senses in providing default rational justification” (Sosa 2010: 139). What is interesting about Sosa’s proposal is that, on one hand, he compares testimonial knowledge to instrumental knowledge. This seems to suggest that Sosa is on the side of the reductivists, who maintain that testimonial justification is reducible to the reliability of deliverances from instruments. And yet, on the other hand, insofar as Sosa also maintains that testimony is like perception in providing default traditional justification, it seems that he is on the side of the non-reductivists, who maintain that testimonial justification is unique (Faulkner 107

108

Ernest Sosa Encountering Chinese Philosophy

2007). Hence, on the standard reductionist versus non-reductionist way of carving up the debate, Sosa’s account of testimony falls in the middle. In this chapter, I will draw on insights from Xunzi’s thought and suggest a way of understanding testimony that leaves open the possibility that the claim that testimony and perception provide default justification and the claim that testimonial knowledge is a kind of instrumental knowledge are compatible.

2 Xunzi’s remarks on testimony The Confucians have not explicitly discussed questions concerning the epistemology of testimony. However, Xunzi, an early Confucian thinker, makes a remark that hearing something is not as good as seeing something: “Not having heard something is not as good as having heard it; having heard it is not as good as having seen it; having seen it is not as good as knowing it; knowing it is not as good as putting it into practice” (Xunzi 8; Knoblock trans. 8.11). Xunzi seems to be expressing here the worry that hearing something is not as good as seeing something, suggesting that there is something less, perhaps less direct, about hearing that p than seeing that p. A clue to why Xunzi might think that comes from his discussion of the role of the heart/mind (xin ᗳ). It is said that: The heart/mind is ruler of the body and the master of the spiritual intelligence. It issues orders but does not receive orders. On its own it prohibits or causes, renounces or selects, acts or stops . . . . The heart/mind cannot be forced to change its thoughts. If the heart/mind thinks something right, it will accept it; but if it thinks something wrong, then it will reject it. Xunzi 21/44–45; Knoblock trans. 21.6a modified

The point of this passage, I take it, is that if the heart/mind registers that a piece of information might change one’s view (regardless of the reliability of the source or content), necessarily, the heart/mind goes through the steps of rejecting or accepting it. Applying this to the case of testimony, Xunzi probably has in mind that the heart/ mind will always make a decision whether it accepts or rejects a piece of testimony. To do that, the heart/mind has to first register that some thoughts are being presented to it for acceptance or rejection. This suggests that it is possible that a hearer’s heart/mind does not register a piece of information as being presented as testimony that is up for acceptance. Since her heart/mind does not register that something is presented to it for acceptance, her heart/mind will not initiate the state that considers whether the piece of testimony should be accepted or rejected.

3 Accepting testimony In the following, I attempt to develop Xunzi’s idea that the heart/mind has to accept or reject any idea that is presented to it. The questions I raise might not be the questions that Xunzi himself was interested in. My aim is not to do Xunzi exegesis but to develop

The Hearer’s Conditions for Accepting Testimony

109

Xunzi’s remarks and speculate on what Xunzi might have had to say about testimony. To understand Xunzi’s thoughts that the heart/mind has to accept or reject any thought that is presented to it, I will suggest a distinction between a hearer’s forming a belief on the basis of an utterance, which is in fact a piece of testimony, and a hearer’s forming a belief on the basis of what she takes to be a piece of testimony. If a hearer does not register that something is being presented as a piece of testimony, she might hear an utterance and believe it without accepting it as testimony. I will argue that the acceptance of testimony necessarily occurs at a reflective level where the hearer concludes that “it is the case that p.” This leaves open the possibility that one’s conclusion is made just on the basis that it is a piece of testimony. My focus is different from the kind of questions that divide the reductivist and nonreductivist theories of testimony. Traditionally, non-reductivists argue that testimony is a distinctive kind of knowledge. When one knows that p from testimony, one acquires testimonial knowledge of p because testimony to p itself provides warrant for the belief that p. We cannot further reduce the reason for believing testimony to p other than it is a piece of testimony. For this reason, testimony is like perception in being its unique source of warrant. Reductivists agree that we can gain knowledge from accepting testimony, but deny that we gain a distinctive kind of knowledge. We can reduce our reason for believing testimony to p to other reasons, such as that testimony is normally reliable. The focus of the debates in testimony thus tends to be on the justification question, namely, whether testimony is itself a source of justification or warrant. The focus of this chapter is not on how testimony justifies the beliefs a hearer acquires from it, or on whether testimony transmits knowledge, or on whether there is a distinctive kind of testimonial knowledge. The focus is on how a hearer acquires beliefs when she accepts a piece of testimony. Sorting out this question will also have a bearing on the way we understand knowledge gained from testimony. The kind of acceptance that is relevant here is the kind that potentially results in the hearer’s believing that p. We will exclude instances of accepting testimonies that only issue in a state of acceptance and make no difference to the hearer’s belief about p.1 In order to understand what it means to accept testimony, we need to first understand what counts as testimony for the hearer. Note that I am not saying what in fact counts as a piece of testimony. I am only thinking of what counts as testimony for the hearer who accepts it as testimony. For one to accept something, one has to first recognize that something is being offered. One cannot accept a proposal if one is never offered a proposal. One cannot accept a job if one is never offered a job. Similarly, for a hearer to accept a piece of testimony, she has to at least think that she is being offered a piece of testimony. In other words, the hearer has to first register that a speech act providing testimony has occurred before she can be in the state of accepting testimony. I suggest the following principle for a hearer’s registering a speech act as a piece of testimony: A hearer H registers what she received from giver G as a piece of testimony to p only if (1) H believes that G made an utterance U, (2) H believes that G means by U that p, and (3) H believes that G believes what she means by U.

110

Ernest Sosa Encountering Chinese Philosophy

I shall specify these three necessary conditions in turn. (1) H must register that an utterance is made. An utterance is a series of actions, which can be verbal or nonverbal, by means of which G means something (Grice 1957). This can accommodate a wide range of circumstances in which H accepts testimony, such as from drawings and gestures. For example, I ask my colleague if she has a pen during a meeting and she shakes her head. I take my colleague to be offering me testimony that she does not have a pen. The utterance does not have to be directed at H or anyone in particular (Sosa 1991). H can count what she overhears from a lecture as testimony or what she read from a private diary as testimony. (2) H believes that G means by U that p. Grice draws a distinction between what an utterance means and what a speaker means by his utterance. For example, when a professor writes a reference letter and comments only on the student’s English ability, what the professor means by the utterance is that the student is not good at philosophy. But in order for H to register an utterance as a piece of testimony to p to H, it is not necessary for G’s utterance to mean p or for G to mean p by her utterance. What matters is that H takes the utterance to mean p. In order for H to register that she is given a piece of testimony to p, she has to take the utterance to mean p. When G utters “My brother is sitting on the fence,” G does mean by her utterance that her brother has not made up his mind. H can register the utterance as a piece of testimony to “G’s brother has not made up his mind,” instead of a testimony to “G’s brother is sitting on a railing.” There are also cases where G by her utterance does not mean p but H can still register that a piece of testimony to p is offered. For example, someone with red–green color blindness says, “It is a blue ball” and he means by his utterance that it is a blue ball. And suppose it is a blue ball. H knows that G sees purple hues as blue hues and mistakenly takes the utterance as a piece of testimony that there is a purple ball. In this case, the hearer thought that she was presented with the testimony that there is a purple ball when in fact she was given testimony that there is a blue ball. Nonetheless, this does not affect the fact that to H’s mind, H is presented with testimony that there is a purple ball. By conditions (1) and (2) alone, H does not have to attribute a belief to the speaker. If I see someone putting on their jacket, I might believe that she is cold without believing that she believes that she is cold. (3) requires H to form a belief about G’s belief. For H to register an utterance as a piece of testimony to p, H has to believe that G believes what she means by her utterance. H does not take utterances by which the speaker is not committed to a view about the world, such as raising a question, performing in theatre, and relaying messages, as testimonies.2 If I hear someone utter, “It is raining” in the context of a play, I will not take her to be offering a piece of testimony. This does not mean that H has to believe that G is presenting p as true. H does not have to take G to be offering evidence for p to count the utterance as testimony to p for H (Graham 1997: 27). G herself might not be aware that she is presenting something to be the case to an audience. For example, when Anne writes in her private diary that she had a fever on 1 January 1945, she is not presenting it as true to anyone. Still, a reader can take this diary entry as providing testimony that Anne had a fever on 1 January 1945. Note that (3) does not require H to believe that G believes that p. It only requires G to believe what her utterance means. Imagine case A:

The Hearer’s Conditions for Accepting Testimony

111

There are contact lenses fitted in G such that he sees red hues as green and green hues as red. There is a red ball. H knows that there are lenses fitted in G and that G does not know that there are lenses fitted. G tells H that there is a green ball. H registers G’s utterance as testimony that there is a red ball. H takes G to mean by his utterance that there is a green ball. H herself takes the utterance to mean that there is a red ball.

(3) is satisfied because H believes that G believes what he means by his utterance, namely, there is a green ball. Hence, for H to register that testimony to p is presented, H does not have to believe that G believes that p. H only needs to believe that G believes what he means by the utterance. Whether G in fact believes what he means by his utterance does not make a difference to whether H registers an utterance as a piece of testimony. Suppose, in case B: Neither G nor H knows that there are lenses fitted in G. G lies and tells H that there is a red ball when in fact G sees a green ball. H still registers G’s utterance as testimony that there is a red ball.

In everyday life, there are many cases where one accepts what one takes to be testimony which then turns out not to be false.3 This does not affect the point that, to H’s mind, she is taking G to be providing testimony to p even though G himself does not believe what he means by his utterance. It is a separate question whether H’s belief is justified. Our focus here is on whether H takes an utterance as a piece of testimony. These are two different questions. Contrast a different case, C: H knows that there are lenses fitted in G and that G does not know that there are lenses fitted. H also knows that G lies to her. G tells H that there is a red ball. On the basis of G’s utterance and her background knowledge, H believes that there is a red ball.4

This case does not count as accepting testimony even though G presents p as true and H ends up acquiring the true belief that p. To H’s mind, G never offered her testimony to p because G does not believe what he means by his utterance. By his utterance, G means that there is a red ball when in fact he believes that there is a green ball. Some might think that there are cases where, even though the speaker does not believe p, the hearer can acquire knowledge of p on the basis of testimony to p (Lackey 1999). But that is a different consideration, one about whether one can acquire knowledge of p on testimony in cases where the speaker does not believe that p. The point here is that, to the hearer’s mind, if she does not believe that the speaker believes what she means by her utterance, the hearer would not have registered the utterance as a piece of testimony. If a student believes that her teacher does not believe evolutionary theory, then she will not take the teacher as giving testimony. Even if the student herself ends up believing evolutionary theory, to the hearer’s mind, she has never accepted testimony from the teacher. Rather, she comes to believe evolutionary theory from

112

Ernest Sosa Encountering Chinese Philosophy

making up her own mind. This is like how a student might come to believe a philosophical theory T from a professor’s talk whose goal is to reject T. The professor’s explanation of T is so clear that it helps the student to understand the theory. The student then makes up her own mind and believes that T. Even though the student hears about T from the professor’s talk, she does not take what the professor says as testimony because she believes that the professor does not believe that T is true. My point here is different from the distinction, drawn by Fricker, between a firstlevel belief of the hearer and a second-level belief (Fricker 1987: 69). For Fricker, a first-level belief is the belief in the content of what is asserted. A second-level belief is the belief as to what speech act has been executed. The relevant second-level belief for testimony is the belief that the speaker has made an assertion. Our focus here is on the side of the hearer accepting testimony. What concerns us is whether the hearer believes that a piece of testimony is provided. Hence, in addition to the hearer’s believing that an assertion has been made, she also has to attribute the relevant belief to the speaker. If the hearer does not attribute the relevant belief to the speaker, then even if the speaker is in fact providing testimony, the hearer will not register the speaker’s utterance as a speech act of providing testimony and hence will not initiate the state of acceptance. Since accepting testimony requires a hearer to attribute the relevant belief to the speaker, if a hearer fails to meet the above three conditions, then even if the speaker is in fact providing testimony to p, the hearer will not register the speaker’s utterance as a speech act of providing testimony and hence will not consider whether she should accept testimony to p. Among the three conditions, (3) is probably the more demanding. It requires more than just the hearer monitoring the speaker’s trustworthiness. Elizabeth Fricker, for example, thinks that ordinary hearers typically monitor for signs of untrustworthiness at a “non-conscious level” (Fricker 1994: 150). Even if this is right, such sub-personal monitoring of untrustworthiness is not enough for the hearer to satisfy (3). Monitoring untrustworthiness at a sub-personal level does not require the hearer to form a belief about the speaker’s belief. It only requires the hearer to monitor for some tell-tale signs of untrustworthiness. Such monitoring may occur at a sub-personal level. When I see someone pressing her fingers to her lips when she tells me that p, I might detect a sign of dishonesty. But my detecting the sign of dishonesty does not always turn into consciously believing that the speaker believes that not-p.5 It is possible that, despite the hearer’s monitoring of trustworthiness at the sub-personal level, she still fails to form the corresponding belief about the speaker’s belief. Once a hearer registers that she is being offered testimony to p, she will have to decide whether to change her mind. Accepting testimony is an activity. The hearer has to draw a conclusion about p on the basis of the testimony presented. The activity of accepting testimony terminates when the hearer has concluded that p is the case. This does not mean that the hearer has to provide reasons to justify for the truth of p.6 It depends on how we fill in an account of the testimonial justification we adopt. If we subscribe to a theory which says that testimony provides irreducible warrant, then on that picture, as soon as a hearer registers that a piece of testimony is offered, she will take the piece of testimony to be decisive evidence for p. The reason for her acceptance is as simple as “because she said so” or “because I trust her.” The mind will not have to

The Hearer’s Conditions for Accepting Testimony

113

engage in further reasoning about what to make of the piece of testimony. As soon as she registers that it is a piece of testimony, she concludes that “it is the case that p.” If we subscribe to a theory which says that acquiring beliefs from testimony is normally reliable, then we may think that, once a hearer registers that testimony is provided, she will then deliberate about whether this particular piece of testimony is to be accepted. Regardless of whether testimony is an irreducible warrant or is normally reliable or something else, on the proposed account the hearer still has to infer that “it is the case that p” because she is at a level on which she has to decide whether she should change her mind. This process is analogous to how an official has to sign a document to make it effective. She might sign all the documents simply because they are certain kinds of documents or she might sign only those that meet certain standards. The structure of the process is one to which she must put her signature in order to conclude the process. Similarly, a hearer has to conclude “it is the case that p” to conclude the process of accepting testimony.

4 Believing p without accepting testimony to p There might be a worry that this account sounds overly intellectual. Intuitively, hearers do not normally take a lot of effort to accept testimonies, such as how to get to the train station or whether the tickets are sold out. Two responses may be made. One response is that the phenomenological character of meeting the conditions for accepting testimony does not have to be one that is noticeably slower and more effortful. From the hearer’s perspective, it does not make much of a phenomenological difference to attribute a belief to someone who says, “I just finished teaching my class,” or to someone who says, “The economy is now in recession.” One might not have the same level of readiness to accept these two, but in both cases the hearer can process two utterances as testimonies in a way that is not significantly phenomenologically different for her. The second response is this. I agree that, in many everyday exchanges, we often readily believe what we are told under normal circumstances without thinking too much about whether the speaker believes what she told us. My explanation for this phenomenon is that hearers need not construe the utterance as a speech act of providing testimony. Perhaps, more often than not, hearers do not construe what they hear as pieces of testimonies. Imagine that Anne has never been to Singapore and she watches a film set in Singapore. The film begins with many skyline shots. When watching the film, Anne acquires many beliefs about Singapore. In one scene, two characters are depicted as talking about visiting an island near Singapore called Bintan. When watching the scene, Anne forms the belief there is an island near Singapore called Bintan. In another scene, the same two characters are depicted as talking about visiting another island near Singapore called Wayan. When watching this scene, Anne also forms the belief there is an island near Singapore called Wayan. It turns out that the part about Bintan island is true, but the part about Wayan island is fictional. But this does not affect the point that, in both cases, Anne can come to believe that p after hearing that p without accepting what she heard as testimony to p. It does not matter why Anne believes that

114

Ernest Sosa Encountering Chinese Philosophy

p here. Perhaps the vividness of depiction somehow has led Anne to think that p is the case. Even though Anne knows that the film is fictional and that the actors are not providing testimonies about Singapore, she still forms the belief that there is an island near Singapore called Bintan. Her belief results from what she heard, not from her mind accepting what she heard as testimony. The above example shows that it is possible for a hearer’s mind to move from hearing that p to believing that p without accepting testimony to p. For an average epistemic responsible agent, chances are that she will at some point believe what she heard without accepting what she heard as testimony. This could happen when she watches films, reads fiction, hears a commercial on the radio, or glances at the headlines of gossip magazines. She might later reflect on her beliefs and correct them, but that is a separate issue. Similarly, the way a hearer’s mind moves when she hears something from what we standardly take to be sources of testimonies could be the same as the way her mind moves when she watches a film. She believes what she heard without accepting what she heard as testimony. This could happen when she watches documentaries, reads non-fiction, hears news on the radio, or glances at the titles of peer-reviewed science articles. Suppose Anne reads in a history textbook that “Raffles landed in Singapore in 1819.” Her mind moves in the same way as her mind moves when she watches the film about Singapore. She comes to believe that Raffles landed in Singapore in 1819 just as she comes to believe that there is an island called Bintan near Singapore from watching the film. Just because the history textbook is a source of testimony, it does not mean that Anne has accepted what she reads as testimony. So even if she ends up believing that Raffles landed in Singapore in 1819 through what she reads, she never processed what she read as testimony. Contrast this with another reader, Ben. When he reads “Raffles landed in Singapore in 1819” in the same history textbook, he takes the history textbook as a source of testimony and believes that Raffles landed in Singapore in 1819. In Ben’s case, his belief that Raffles landed in Singapore in 1819 results from his accepting testimony that Raffles landed in Singapore in 1819. Even though both Anne and Ben come to believe the same thing from reading the same sentence from the same book, their minds move differently. Ann’s mind moves from hearing to believing p without accepting testimony to p, whereas Ben moves directly from hearing to accepting testimony to p. It is another question whether for both Anne and Ben their beliefs that p are formed on the basis of testimony to p. We should not confuse the question about the nature of testimony with the question about whether a hearer accepts what she heard as testimony. If we focus on how a hearer accepts testimony, we do not have to assume that just because a hearer acquires the belief that p through what is in fact a piece of testimony the hearer must have accepted the testimony. On my account, in many of the mundane cases where the hearers do not seem to process testimony in an overly intellectualized manner, the hearers simply form beliefs through what they heard without taking what they heard as testimony. Suppose I am on the plane and my fellow traveler tells me that she lives in London. I come to believe that she is from London. It is possible that my processing of her utterance and her trustworthiness all happened at a sub-personal level. I think she is trustworthy. I

The Hearer’s Conditions for Accepting Testimony

115

understand what she is saying. And I come to believe what she said. But I am not monitoring what she believes and have not attributed a belief to her. In this case, even though I come to believe what she said, I have not accepted what she said as testimony. It is not that I do not accept the testimony; it is just that it has not occurred to me that her speech act is one of testifying. This is unlike in a job interview situation where I am the interviewer. In such a situation, I am poised to consciously monitor her beliefs and take what she said as testimonies. Hence, I form beliefs about what the interviewee told me from what I take to be testimonies from her.

5 Two senses of uptake When Fricker talks about how one may gain knowledge from testimony, she writes: when two individuals S and H are masters of a common language the following can happen: A belief of S gives rise to an utterance by him, which utterance produces in his audience H a belief with the same content; and all this happens in such a way that, if S’s belief is knowledge, then we may allow that title to H’s belief too. Fricker 1987: 57

I agree with Fricker that this is how a speaker may pass a belief to a hearer. The speaker believes her belief that p to the hearer. The hearer upon hearing the utterance comes to believe that p. But this is only a broad description. This is like saying how a postal worker can pass a parcel and the parcel ends up in my house. This is a broad description of what happened. We can further describe more precisely the way in which the parcel ends up in my house. This can happen when the postal worker just slips the parcel through the mail slot on the door; this can also happen when I accept the parcel and put it in my house. A hearer who successfully registers an assertion at the sub-personal might not register the assertion as a piece of testimony. This is analogous to saying how one might acquire a parcel without ever accepting the parcel. It is only when a hearer registers an assertion as a piece of testimony that she is in a position to accept the testimony. This is analogous to saying it is only when she registers that it is a parcel for her that she will then go through the step of accepting the parcel. In summary, the transmission of belief from a speaker to a hearer may happen in at least two different ways: (a) Hearing “p” → Believing that p. (b) Hearing “p” → Accepting testimony to p → Believing that p. “→” here indicates a sequence. If I am right about the distinction between (a) and (b), then we may talk about “uptake of testimony” in two different senses. By “uptake,” I mean one’s believing that p as a result of hearing that p. We may refer to (a) as the broad sense of uptake and (b) as the narrow sense of uptake. On both (a) and (b), a hearer’s mind is changed by a speaker’s utterance. The difference is that on (a), one’s mind is

116

Ernest Sosa Encountering Chinese Philosophy

changed without the hearer accepting the speaker’s utterance as testimony; on (b), one’s mind is changed as a result of the hearer accepting the speaker’s utterance as testimony. An account of how we gain knowledge from testimony must explain both (a) and (b). At this point, we may leave open the question whether a hearer can gain knowledge from (a). Minimally, a belief formed in the sense of (b) accepting testimony counts as testimonial knowledge. I will call it “testimonial knowledge from acceptance.”

6 Perception and instrument In the above, I developed Xunzi’s remarks on testimony and suggested that there is a distinction between testimonial uptake that results from accepting testimony to p and testimonial uptake that results from hearing what is in fact a piece of testimony that p without the hearer registering it as testimony that p. Testimonial knowledge from acceptance requires the hearer to register that the conditions for testimony are met and accept the piece of testimony. On such an understanding of testimonial knowledge from acceptance, it is possible to make both the claims that testimony provides justification like perception and that testimonial knowledge is a kind of instrumental knowledge as Sosa has suggested. The non-reductivists maintain that testimony is a source of warrant like perception. John McDowell and Elizabeth Fricker, for example, maintain that hearers’ discernment of speakers’ utterances is an instance of perception (McDowell 1981). Fricker summarizes McDowell’s two claims as follows. The first claim is about the phenomenology of language use: “a hearer competent in the language used by speaker will hear an utterance as the speech act it is – will experience it as such.” In the case of testimony, the hearer experiences utterances as assertions. The second claim is about epistemology: experiencing an utterance as the speech act it is “provides a basis for knowledge in much the same way that our experiences as of the instantiations of less special states of affairs provide perceptual knowledge of them” (Fricker 1987: 70). The proposed account is not affected by the first claim. Let us grant that a hearer perceives a speaker’s utterance as assertions. In order for a hearer to discern an utterance as one that is providing testimony, it is not enough for a hearer to register that an assertion is made or to form a belief about the speaker’s trustworthiness. The hearer has to experience the speech act as one of giving testimony. Taking an utterance as a piece of testimony is more than taking an utterance as an assertion. The hearer has also to believe that the speaker believes that p. With regard to the epistemological claim, my account does not allow the stronger epistemological claim that testimony provides a basis for knowledge in much the same way as perception. Unlike perception, accepting testimony requires the hearer to reason her way from “because she said p” to “p is the case.” That said, the proposed account can allow for a weaker claim that says something like testimony provides a basis for knowledge in a way similar to perception in that something being a piece of testimony itself is a reason for acceptance. Hence, my account leaves room for the claim that testimony is analogous to perception in being a source of warrant.

The Hearer’s Conditions for Accepting Testimony

117

The proposed account is similar to a reductivist account to the extent that it holds that the belief that a hearer acquires from accepting testimony to p is inferential. The hearer has to come to the conclusion that p on the basis of testimony presented. In this respect, the way a hearer’s mind moves when she accepts testimony is the same as the way her mind moves when she accepts deliverances from instruments. Both testimonies and deliverances from instructions serve as evidence for a hearer. Even if the evidence is of a kind that by itself serves as decisive evidence, the hearer’s mind still has to treat it as evidence and draw a conclusion on the basis of the evidence. In this regard, accepting testimony that p is substantively different from perceiving that p. Even if testimony provides default justification, the hearer’s mind still has to move from hearing that p via accepting testimony to p to believing that p. It is perhaps due to this mediated route that Xunzi thinks that learning from testimony that p is not as direct or good as learning from seeing that p. There are still many details that need to be filled in, but I hope that I have borrowed Xunzi’s insights and showed one way in which we can understand Sosa’s view on testimony. The proposed account differs from mainstream accounts of testimony in that it focuses on the hearer’s conditions in which she accepts testimony. This shift in focus helps us cut the nature of testimonial uptake at a different place from the traditional debates between the reductivists and the non-reductivists. Once we are not restricted by the traditional divide between reductivists and non-reductivists, the apparent tension in the thought that testimony is a kind of instrumental knowledge that provides default justification like perception is also loosened.

Notes 1 These instances might occur when a hearer already has the belief that p prior to accepting the testimony. The audience might also act as if p and/or uses p in her reasoning in a certain context but abandons p when the context ends. There are also situations where one passes on testimony that p without believing that p. This can happen when p is inaccessible to the subject, such as an unconscious belief, or when the hearer plans to spread gossip she does not believe (Faulkner 2011), or when a teacher is forced to teach something she does not believe (Lackey 1999). 2 For example, I can infer from the question “Were you the one who opened the window?” that the window is open even though the speaker is not providing testimony that the window is open. (Faulkner’s example 2011: 16, modified). 3 We can imagine a different case, D, where G knows that he has lenses fitted and that H knows that he has lenses fitted. G also knows that H does not know that G knows he has lenses fitted. G lies and tells H that there is a red ball. This is also a case where H thinks that she is offered with testimony, even though she ends up forming a false belief. 4 Lackey (1999) introduces the example of a teacher who does not believe evolutionary theory but teaches it out of obligation. If we imagine that a student knows that the teacher does not believe evolutionary theory, but after hearing the lecture, the student herself comes to believe evolutionary theory, then this structurally parallels this case. 5 Recent research shows that ordinary hearers’ “less-conscious processing” of signs of unworthiness might be more accurate than “more-conscious processing.” This suggests

118

Ernest Sosa Encountering Chinese Philosophy

that sub-personal monitoring and conscious monitoring can come apart. In some cases, our higher-order conscious responses can interfere with the less-conscious processing turning up in higher-order consciousness (ten Brinke et al. 2014). 6 This is different from e.g. Coady (1992), Burge (1993), McDowell (1980).

References Burge, T. 1993. “Content Preservation.” Philosophical Review 102(4): 457–488. Coady, C. A. J. 1992. Testimony: A Philosophical Study. Oxford: Oxford University Press. Faulkner, P. 2007. “Review of The Epistemology of Testimony by Jennifer Lackey and Ernest Sosa.” Mind 116(464): 1136–1139. Faulkner, P. 2011. Knowledge on Trust. Oxford: Oxford University Press. Fricker, E. 1987. “The Epistemology of Testimony.” Proceedings of the Aristotelian Society Supplementary 61: 57–83. Fricker, E. 1994. “Against Gullibility.” In B. K. Matilal and A. Chakrabarti (eds), Knowing from Words. Boston: Kluwer, pp. 125–161. Graham, P. 1997. “What is Testimony?” The Philosophical Quarterly 47(187): 227–232. Grice, H. P. 1957. “Meaning.” The Philosophical Review 66(3): 377–388. Knoblock, J. 1988, 1990, 1994. Xunzi: A Translation and Study of the Complete Works, 3 vols. Stanford: Stanford University Press. Lackey, J. 1999. “Testimonial Knowledge and Transmission.” The Philosophical Quarterly 49(197): 471–490. McDowell, J. 1980. “Meaning, Communication, and Knowledge.” In Zak Van Straaten (ed.), Philosophical Subjects: Essays Presented to P. F. Strawson. Oxford: Claredon Press, pp. 117–139. Sosa, E. 1991. Knowledge in Perspective. Cambridge: Cambridge University Press. Sosa, E. 2010. Knowing Full Well. Princeton: Princeton University Press. ten Brinke, L., Stimson, D., and Carney, D. R. 2014. “Some Evidence for Unconscious Lie Detection.” Psychological Science, 25(5): 1098–1105.

8

Sosa’s Reflective Knowledge and Xunzi’s Knowledge of the Dao Leo K. C. Cheung

1 Sosa on animal and reflective knowledge Ernest Sosa seems to have introduced the notion of intellectual virtue into philosophy in the following passage from his paper “The Raft and the Pyramid” (1980):1 “Here primary justification would apply to intellectual virtues, to stable dispositions for belief acquisition, through their greater contribution toward getting us to the truth. Secondary justification would then attach to particular beliefs in virtue of their source in intellectual virtues or other such justified dispositions” (Sosa 1980: 23). Here, Sosa already puts forward the view that there are two levels of justification— one is the exercise of intellectual virtues for the reliabilistic stabilization of dispositions for belief acquisition, and another a second justification attached to the beliefs acquired in virtue of their source in intellectual virtues. This seems to be the first introduction of virtue epistemology, via his brand “virtue perspectivism,” into philosophy. Sosa relates intellectual virtues or, rather, his virtue epistemology with justification, as explained in his paper “Knowledge and Intellectual Virtue” (1985): “An intellectual virtue is a subject-grounded ability to tell truth from error infallibly or at least reliably in a correlated field. To be epistemically justified in believing is to believe out of intellectual virtue” (Sosa 1985: 243). Clearly, Sosa is relating justifications with the reliability of intellectual virtues in such a way that an intellectual virtue, as a faculty of the relevant subject, is able to enhance the reliability of the belief-acquisition process. The latter is actually determined by the reliability of the subject with the relevant intellectual virtue or faculty. It also seems that Sosa first introduces the distinction between animal knowledge and reflective knowledge into philosophy in the following:2 “One has animal knowledge about one’s environment, one’s past, and one’s own experience if one’s judgments and beliefs about these are direct responses to their impact—e.g., through perception or memory—with little or no benefit of reflection or understanding. One has reflective knowledge if one’s judgment or belief manifests not only such direct response to the fact known but also understanding of its place in a wider whole

119

120

Ernest Sosa Encountering Chinese Philosophy

that includes one’s belief and knowledge of it and how these come about” (Sosa 1985: 241–242). Thus, Sosa (1985) holds that mere animal knowledge is constituted by justified true beliefs acquired by direct responses to impacts via perception or memory. It has no benefit of reflection. On the other hand, reflective knowledge is animal knowledge with reflective understanding of the place and source of the relevant beliefs in the relevant environment. For Sosa (1985), both animal knowledge and reflective knowledge involve justified beliefs. The difference between them is that “[a]nimal knowledge yielded by reaction to the relevant field unaided by reflection on the one’s belief and its object within one’s wider view” (Sosa 1985: 243), while reflective knowledge is aided by reflection. Sosa (1985) also holds that reflective knowledge is better justified than, and thus better than, animal knowledge: Since a direct response supplemented by such understanding would in general have a better chance of being right, reflective knowledge is better justified than corresponding animal knowledge . . . . . . A reason-endowed being automatically monitors his background information and his sensory input for contrary evidence and automatically opts for the most coherent hypothesis even when he responds most directly to sensory stimuli. Sosa 1985: 242

On the same page (Sosa 1985: 242), Sosa also discusses how the subject with the relevant intellectual virtue may be able to monitor any possible overriders. He therefore seems to think that reflective knowledge is reliably better than animal knowledge because of the following two reasons: (1.1)

(1.2)

The reflecting subject with the relevant intellectual virtue monitors contrary evidence or overriders (and this makes reflective knowledge reliably better than animal knowledge). The reflecting subject automatically opts for the most coherent hypotheses (and this fulfils a certain internal coherence requirement for knowledge and thus makes reflective knowledge reliably better than animal knowledge).

That Sosa holds (1.1) is further supported by the following passage: “Since a direct response supplemented by such understanding would in general have a better chance of being right, reflective knowledge is better justified than corresponding animal knowledge” (Sosa 1985: 242). That Sosa holds (1.2) is further supported by a later passage in the paper “Reflective Knowledge in the Best Circles” (1997): “This is not to deny that there is a kind of ‘animal knowledge’ untouched by broad coherence. It is rather only to affirm that beyond ‘animal knowledge’ there is a better knowledge. This reflective knowledge does require broad coherence, including one’s ability to place one’s first-level knowledge in epistemic perspective” (Sosa 1997: 422).

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao

121

Sosa (1985) thinks it is difficult to “bypass justification” (Sosa 1985: 244). However, in the paper “Reliabilism and Intellectual Virtue” (1991), he does bypass justification at the level of animal knowledge via a distinction between “justification” and “aptness”: (a) The “justification” of a belief B requires that B have a basis in its inference or coherence relations to other beliefs in the believer’s mind—as in the “justification” of a belief derived from deeper principles, and thus “justified,” or the “justification” of a belief adopted through cognizance of its according with the subject’s principles, including principles as to what beliefs are permissible in the circumstances as viewed by that subject. (b) The “aptness” of a belief B relative to an environment E requires that B derive from what relative to E is an intellectual virtue, i.e., a way of arriving at belief that yields an appropriate preponderance of truth over error (in the field of propositions in question, in the sort of context involved). Sosa 1991: 144

Accordingly, an apt belief only requires the derivation of it from an intellectual virtue (of the subject), while a justified belief also requires the justification of it via the secondary coherence relations to other beliefs in the subject’s mind or epistemic perspectives. Sosa (1991) puts this neatly as follows: Virtue perspectivism distinguishes between aptness and justification of belief, where a belief is apt if it derives from a faculty or virtue, but is justified only if it fits coherently within the epistemic perspective of the believer—perhaps by being connected to adequate reasons in the mind of the believer in such a way that the believer follows adequate or even impeccable intellectual procedure. Sosa 1991: 145

Sosa (1991) then characterizes animal knowledge and reflective knowledge as follows: Virtue perspectivism distinguishes between animal and reflective knowledge. For animal knowledge one needs only belief that is apt and derives from an intellectual virtue or faculty. By contrast, reflective knowledge always requires belief that not only is apt but also has a kind of justification, since it must be belief that fits coherently within the epistemic perspective of the believer. ibid.

In other words, for Sosa (1991), animal knowledge consists of apt beliefs, while reflective knowledge consists of apt beliefs with justifications bringing forth coherent relations holding among other beliefs within the epistemic perspective of the subject. Later, Sosa lifts the justification requirement from reflective knowledge as well. In A Virtue Epistemology (2007) and Reflective Knowledge (2009b), he defines animal knowledge and reflective knowledge such that justification has no status whatsoever in either of them. To achieve this, Sosa (2007, 2009b) first defines aptness and animal knowledge as follows:

122

Ernest Sosa Encountering Chinese Philosophy

A performance is apt if, and only if, it is correct attributably to a competence exercised by the performer, in conditions appropriate for its exercise. Sosa 2007: 92 Beliefs are a special case of performances, epistemic performances. When a belief is correct attributably to a competence exercised in its appropriate conditions, it counts as apt and as knowledge of a sort, animal knowledge. Sosa 2007: 93 Animal knowledge is essentially apt belief. 2007: 24

Sosa (2007, 2009b) then defines reflective knowledge as follows: Indeed, despite leaving the word “knows” undefined, one might proceed in three stages as follows:

(a) affirm that knowledge entails belief; (b) understand “animal” knowledge as requiring apt belief without requiring defensibly apt belief, i.e., apt belief that the subject aptly believes to be apt, and whose aptness the subject can therefore defend against relevant skeptical doubts; and (c) understand “reflective” knowledge as requiring not only apt belief but also defensibly apt belief. Sosa 2007: 24

Sosa’s mature theory of animal knowledge and reflective knowledge consists in the views that, first, an apt belief, or one that is correct attributably to a competence, as an intellectual virtue, exercised in its appropriate conditions, is an instance of animal knowledge. Second, an apt belief that the subject aptly believes to be apt, or a defensibly apt belief, is an instance of reflective knowledge. This second point needs to be further explained here. Sosa amended the paper “Reflective Knowledge in the Best Circles” (1997) with some alterations to form chapter 9 of Sosa (2009b). The passage in the section “Virtue and Coherence” (1997: 422) quoted before was retained, with a few insignificant changes, as follows: This is not to deny that there is a kind of “animal knowledge” that owes little to such broad coherence. It is rather only to affirm that beyond “animal knowledge” there is a better knowledge. This reflective knowledge does require a good measure of broad coherence, including one’s ability to place one’s first-level knowledge in epistemic perspective. Sosa 2009b: 193

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao

123

This shows that, for Sosa (2007, 2009b), when the subject aptly believes a certain apt belief to be apt, and thus produces a defensibly apt belief, the apt belief coheres among other beliefs within the epistemic perspective of the subject—fulfilling a certain internalist coherence requirement for knowledge. At this juncture, three remarks are in order. First, reflective knowledge, as apt beliefs aptly believed to be apt, is animal knowledge, but not vice versa. Second, if Sosa’s distinction between mere animal knowledge and reflective knowledge is tenable, then this provides a middle ground between reliabilism, as a kind of externalism, on one side, and internalism on the other. This is because apt beliefs, as beliefs correctly attributed to a competence or an intellectual virtue, are reliably produced. Mere animal knowledge, consisting of apt beliefs, is reliabilist and thus externalist. Moreover, reflective knowledge, consisting of defensibly apt beliefs, fulfils the internalist requirement that a defensibly apt belief coheres with other beliefs within the epistemic perspective of the subject. Reflective knowledge is then internalist. Third, if Sosa’s view that mere animal knowledge is externalist, while reflective knowledge is internalist, is correct, then it can contribute to the resolutions of the new evil demon problem and the meta-incoherence problem.3 Before proceeding to the discussion of Xunzi’s views, I would like to point out that Sosa holds that the distinction between animal knowledge and reflective knowledge, however, faces the problem of forbearing or withholding (Sosa 2009a, 2011). He introduces the notions of narrow and broad aptness, and attempts to solve the problem of forbearing by granting that the forbearing is merely narrowly apt. In this regard, he also introduces the notions of meta-aptness and full aptness, and then that of knowing full well. A performance is fully apt if it is in virtue of being meta-apt that it is apt. An instance of knowing full well is a special case of fully apt performance. Accordingly, an instance of reflective knowledge is not an instance of knowing full well if the relevant apt belief is not apt in virtue of the aptness of the meta-belief that it is apt (even if that meta-belief is indeed apt). The theory of knowing full well is therefore a further development of the distinction between animal knowledge and reflective knowledge. However, since the notion of knowing full well is not my main concern, I will not go into detail here.

2 Xunzi’s unification of nature and heaven via the dao I am now going to argue that Xunzi, as the author of the Xunzi, holds that there are two different types of knowledge; namely, knowledge of instances of change (hereafter “knowledge of changes”) and knowledge of the dao. I shall argue that, for Xunzi, knowledge of changes may be externalist or internalist, while knowledge of the dao must be internalist. I believe Xunzi has given one of the best definitions of dao in Pre-Qin philosophy in Book 21: Dispelling Blindness (Jiebi 䀓㭭) of the Xunzi (my translation):4 The dao is in itself constant and governs all changes (or all instances of change). ཛ䚃㘵ˈ億ᑨ㘼ⴑ䆺DŽ

124

Ernest Sosa Encountering Chinese Philosophy

I have argued that, for the Xunzi, there are at least two different classifications of changes (Cheung 2001). According to the first classification, there are these two different types of change: (2.1) (2.2)

The change of Heaven and Earth or, hereafter, the change of the Nature. The change of human affairs.

Instances such as the fall of stars and the groaning of trees mentioned in Book 17: Discourse on Heaven (tian lun ཙ䄆ㇷ) belong to the change of Heaven (and Earth), or the Nature. The following passage from Book 8: The Teachings of the Ru (ruxiao ݂᭸), which is on the Duke of Zhou, clearly shows that Xunzi accepts that there are instances of the change of human affairs: Since for the sake of the peace of the empire, the duke had carried on the tasks of Wen and Wu, he made clear the principle about the proper relationship between the branch and main lines of a family. These are also changes and transformations.5 ഐཙлѻ઼ˈ䙲᮷↖ѻᾝˈ᰾᷍ѫѻ㗙ˈᣁӖ䆺ॆ⸓DŽ

According to the second classification, there are the following two different types of change: (2.3) (2.4)

Begetting (sheng ⭏) Transformation (hua ॆ)

I shall explain that, according to Xunzi, “transformation” can be further classified into the following three different types. Before doing that, however, I shall explain what is meant by “begetting” first. Consider the following passage from Book 19: Discourse on Ritual Principles (Li lun ⿞䄆ㇷ): when Heaven and Earth conjoin, the myriad things are begot when the Yin and Yang principles combine, transformations and transmutations are produced; when inborn nature and conscious activity are joined, the world is made orderly. Knoblock 1999: 67; emphasis mine ཙൠਸ㘼㩜⢙⭏ˈ䲠䲭᧕㘼䆺ॆ䎧ˈᙗ‫ڭ‬ ਸ㘼ཙл⋫DŽ

Because the begetting of a thing, or a thing’s being begot, is clearly an instance of change, begetting (2.3) is a type of change. The second type of change, according to the second classification, is transformation (2.4). Commentators seem to have overlooked the fact that the Xunzi has a very clear definition of the notion of transformation in Book 22: On the Correct Use of Names (zhengming ↓਽): Where the appearance undergoes metamorphosis, but there is no distinction in the reality, yet they are deemed different, it is called “transformation.” Knoblock 1999: 131 ⣰䆺㘼ሖ❑ࡕ㘼⛪⮠㘵ˈ䄲ѻॆDŽ

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao

125

Transformation does not involve the change of “reality” or, rather, substance, but it is still a type of change. Although the Xunzi does not specify what it means by “distinction in reality,” it is reasonable to say that begetting makes a distinction in reality. Then transformation and begetting are two different types of change. Transformation can be further divided into three different types: (2.4A) Transformation of yin and yang (yin yang zhi hua 䲠䲭ѻॆ) (2.4B) Well-order (zhi ⋫) (2.4C) Anarchy (luan Ҳ) The first type of transformation is the transformation of yin and yang (2.4A). That Xunzi holds that there are instances of the transformation of yin and yang is proven by the passage in Book 19: Discourse on Ritual Principles, already quoted above, where it says, “when the Yin and Yang principles combine, transformations and transmutations are produced.” Book 17: Discourse on Heaven also takes the fall of stars and the groaning of trees to be unusual events due to the transformation of yin and yang. Clearly, for Xunzi, these events only involve the change of motion or combination, and not begetting, and thus are instances of transformation. The other two different types of transformation are Well-order (2.4B) and Chaos (2.4C). It should be noted that the above quotation from Book 19: Discourse on Ritual Principles deliberately writes, “when inborn nature and conscious activity are joined, the world is made orderly [zhi]” alongside the two sentences about begetting and the transformation of yin and yang, respectively. This to a certain extent shows that wellorder is regarded as a type of transformation. Moreover, according to Book 12: On the Way of a Lord, the impossibility of appropriate response to changes of human affairs6 is sufficient to cause anarchy [Ҳ luan]. Knoblock 1999: 126 н㜭៹㩜һѻ䆺ˈ䏣ԕҲ⸓DŽ

The purpose of responding to changes of human affairs is to avoid anarchy. And wellorder and anarchy are always put into contrast. This is proven by, among other things, the following passage from Book 14: On Attracting Scholars: when the gentleman is obtained, there is order (zhi ⋫); when the gentleman is lost, there is anarchy (luan Ҳ). Knoblock 1990: 207; emphasis mine ᗇѻࡷ⋫ˈཡѻࡷҲDŽ

This shows that “non-anarchy” simply means “well-order.” Moreover, it is important to note from the passage in Book 3: Nothing Indecorous: “well ordered [well-order]” refers to ritual and moral principles and that “chaotic [anarchy]” refers to what is contrary to them. Knoblock 1999: 176–177 ⿞㗙ѻ䄲⋫ˈ䶎⿞㗙ѻ䄲Ҳ

Another supporting passage is from Book 19: Discourse on Ritual Principles:

126

Ernest Sosa Encountering Chinese Philosophy

Through rites, Heaven and Earth are conjoined, the sun and moon shine brightly . . . . The myriad things change but not chaotic (luan Ҳ),7 but if one is divided in his loyalty to them, he will be brought to ruin. Surely it is true that the rites are indeed perfection! Knoblock 1999: 60 㩜⢙䆺㘼нҲˈ䋣ѻࡷ௚ҏˈ⿞䉸н㠣⸓ૹDŽ

Thus, well-order [⋫ zhi], as a function of ritual principles is neither constant nor unchanged, but, rather, “change but not anarchic.” Well-order (that is, change but not anarchic), or (2.4B), is then a type of change. Because [Ҳ luan] is simply change and anarchic, anarchy, or (2.4B), is a type of change. Let us make a conclusion at this stage that, for the Xunzi, there is the classification of change into the change of the Nature and the change of human affairs. There is also the classification of change into begetting and transformation, where transformation can be further divided into at least the transformation of yin and yang, well-order and anarchy. We can now turn back to Xunzi’s definition of the dao: “The dao is in itself constant and governs all changes.” We may take the relevant statement of the constant, or the regularity, intrinsic to the dao to be the principle or the law induced by the dao. Because the dao governs all instances of change, it induces the most general principle or law. In this way, the dao is shown to be something nomic or nomological. With respect to a certain field of instances of change, we have the following definition of a specific dao: (2.5)

A dao is that which is in itself constant and governs all changes in a certain field.

We may take the relevant statement of the constant, or the regularity, intrinsic to a specific dao to be the principle or the law induced by the specific dao. If we employ the first classification of instances of change, then we have the following definitions: (2.6) (2.7)

The dao of the Nature is that which is in itself constant and governs all instances of change of the Nature. The dao of human affairs is that which is in itself constant and governs all instances of change of human affairs.

The dao of the Nature induces the general law of the Nature. The general law of the Nature is the general law of nature constituted by the laws of nature understood in the sense of modern science. This shows that Xunzi holds the unification of natural sciences, and that there is the general law of nature or the most general scientific law. Because Xunzi holds that there is the dao of human affairs, he would also hold that there is the general law of human affairs, which governs all the instances of change of human affairs, induced by the dao of human affairs. According to Xunzi, the most general principle induced by the dao is li ⿞, or ritual. This is shown by the following claim in Book 17: Discourse on Heaven that li is the marker of the dao:

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao

127

When men cross the water at fords, they mark the deep places; but if their markers are unclear, those who come after will drown. Those who govern the people mark out the Way, but if the markers are not clear, then the people will fall into disorder [luan]. Ritual principles are such markers. Knoblock 1999: 21 ≤㹼㘵㺘␡ˈ㺘н᰾ࡷ䲧DŽ⋫≁㘵㺘䚃ˈ㺘н᰾ࡷҲDŽ ⿞㘵ˈ㺘ҏDŽ

The particular example in this passage is the li, or the general law, of human affairs. Actually, Xunzi also talks about the li, or the general law, of the Nature. For example, he writes in Book 19: Discourse on Ritual Principles, Through rites, Heaven and Earth are conjoined, the sun and moon shine brightly, the four seasons observe their natural precedence, the stars and planets move in ranks, the rivers and streams flow, and the myriad things prosper. Through them, love and hate are tempered, and joy and anger made to fit the occasion. They are used to make inferiors obedient and to make superiors enlightened. The myriad things change but not chaotic [luan]; but if one is divided in his loyalty to them, he will be brought to ruin. Surely it is true that the rites are indeed perfection! Knoblock 1999: 60 ཙൠԕਸˈᰕᴸԕ᰾ˈഋᱲԕᒿˈᱏ䗠ԕ㹼ˈ⊏⋣ԕ ⍱ˈ㩜⢙ԕ᰼˗ྭᜑԕㇰˈௌᙂԕ⮦˗ԕ⛪лࡷ丶ˈԕ⛪кࡷ᰾ˈ㩜⢙䆺 㘼нҲˈ䋣ѻࡷ௚ҏˈ⿞䉸н㠣⸓ૹDŽ

Clearly, Xunzi holds that there is li, as the general law, of the Nature. Since Xunzi takes the dao to be that which is constant and governs all changes, including changes of the Nature and that of human affairs, he would hold that the li, or general law, of the nature and the li of human affairs are unified via the li of the dao— the most general li. It is important to note that, according to such a view, even though the li of the Nature is descriptive while the li of human affairs is normative, they are unified via the same general li. It is therefore proven that the dao in the Xunzi is nomic, and that Xunzi holds, what I call, “Xunzi’s unification thesis,” that the dao of the Nature and the dao of human affairs are unified via the dao. Suppose we call the totality of empirical and scientific facts and normative human affairs “reality.” The unification thesis can also be formulated as follows: (2.8)

The (great) dao is the generalization of the dao of reality.

3 Xunzi’s nomic-transcendentalism, knowledge of change and knowledge of the dao Xunzi’s unification thesis holds that the dao induces the most general law, or li, of reality, governing all instances of change, including that of the Nature and that of human beings. In particular, the (descriptive) dao of the Nature and the (normative) dao of human affairs are unified via the (great) dao.

128

Ernest Sosa Encountering Chinese Philosophy

Moreover, according to Book 31 Duke Ai, the great sage (hereafter “the sage”) is the one who knows the great dao (that is, “the dao”): Those who are called great sages are persons who have an awareness (zhi ⸕; knowledge) that extends to the Great Way [dao], who are limitlessly responsive to every change. Knoblock 1999: 975 ᡰ䄲བྷ㚆㘵ˈ⸕䙊Ѿབྷ䚃ˈ៹䆺㘼нマ

This passage also says, among other things, that the sage knows the dao in order to be responsive to all kinds of changes. Given nomic-transcendentalism, reasonably, he would hold that the sage responds to changes by means of nomological arguments (including deductive, inductive and deontic arguments), based on the general li, as the marker of the dao: 1. An instance of change, or a cluster of instances of change, obtains. 2. A description of the li of the dao. -------------------------------------------------------------------- [necessarily, probably or deontically] 3. Another instance of change, or another cluster of instances of change, obtains. In order for the sage to be able to respond to instances of change, she must be able to acquire both knowledge of the obtaining and non-obtaining of instances of change (hereafter “knowledge of changes”) and knowledge of the dao, or the li. The former involves knowledge of [1], while the latter [2]. Xunzi explains how one can acquire knowledge of changes in the following passages in Book 1 An Exhortation to Learning: The gentleman says: “Learning must never be concluded.” . . . In broadening his learning, the gentleman each day examines himself so that his awareness will be discerning and his actions without excess. Knoblock 1988: 135 ੋᆀᴠ˖ᆨнਟԕᐢDŽੋᆀঊᆨ㘼ᰕ৳ⴱѾᐡˈ ࡷᲪ᰾㘼㹼❑䙾⸓DŽ Truly if you do not climb a high mountain, you will be unaware of the height of the sky. If you do not look down into a deep gorge, you will be unaware of the thickness of the earth. If you have not heard the words inherited from the Ancient Kings, you will be unaware of the greatness of learning and inquiry. Knoblock 1988: 135 ᭵нⲫ儈ኡˈн⸕ཙѻ儈ҏ˗н㠘␡䉯ˈ н⸕ൠѻ৊ҏ˗н㚎‫⦻ݸ‬ѻ䚪䀰ˈн⸕ᆨ୿ѻབྷҏDŽ I once spent a whole day in thought, but it was not so valuable as a moment in study. I once stood on my tiptoes to look out into the distance, but it was not so effective as climbing up to a high place for a broader vista. Knoblock 1988: 135 ੮ే㍲ᰕ㘼ᙍ⸓ˈнྲ丸㠮ѻᡰᆨҏΧ ੮ే䏲㘼ᵋ⸓ˈнྲⲫ儈ѻঊ㾻ҏDŽ

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao

129

What kind of learning is this? What is the nature of the knowledge one would obtain? Clearly, according to the above passages, Xunzi’s views are that the learning involved is empirical, and that knowledge thus acquired is knowledge of changes. Moreover, according to Sosa’s virtue epistemology, this holds: (3.1)

Knowledge of changes is animal knowledge, reflective knowledge or knowing full well.

What then is the nature of knowledge of the dao? Xunzi’s answer is to be found in the following passage in Book 21: Dispelling Blindness: What do men use to know the Way [dao] ? I say that it is the mind. How does the mind know? I say by its emptiness, unity, and stillness (xu yi er jing 㲋༩㘼䶌) . . .. Not allowing what has previously been stored to interfere with what is being received in the mind is called emptiness (xu 㲋) .. . . Not allowing the one thing to interfere with the other is called unity (yi а) . . .. Not allowing dreams and fantasies to bring disorder to awareness is called stillness (jing 䶌) . . .. A person who knows the Way (zhi dao ⸕䚃) and discerns it and puts it into practice embodies the Way [dao]. Emptiness, unity, and stillness are called the Great Pure Understanding (da qing ming བྷ␵᰾). Knoblock 1999: 683, 685 Ӫօԕ⸕䚃˛ᴠ˖ᗳDŽᗳօԕ⸕˛ᴠ˖㲋༩㘼 䶌DŽнԕᡰᐢ㠗ᇣᡰሷਇ䄲ѻ㲋DŽнԕཛаᇣ↔аˈ䄲ѻ ༩DŽнԕདྷࢷҲ⸕䄲ѻ䶌DŽ⸕䚃ሏˈ⸕䚃㹼ˈ 億䚃㘵ҏDŽ㲋༩㘼䶌ˈ䄲ѻབྷ␵᰾DŽ

The gentleman can know the dao by making the heart-mind endowed with emptiness, unity, and stillness, so as to attain the Great Pure Understanding. When one attains the Great Pure Understanding, one acquires knowledge of the dao. To possess emptiness, unity, and stillness, one should not allow the previously stored to interfere with what is being received in the mind, the one thing to interfere with the other, and dreams and fantasies to bring disorder to awareness, respectively. Therefore, the Great Pure Understanding is independent of anything external to the heart-mind. It follows that, for Xunzi, the following hold: (3.2) (3.3)

Knowledge of the dao is purely internalist. Knowledge of the dao is based on introspective reflection entirely.

In particular, because of (3.2) and (3.3), the Xunzi adopts knowledge internalism (hereafter, “K-internalism”) with respect to knowledge of the dao. Here, I would like to further explain (3.2) and (3.3), in order to locate the philosophical thesis supporting them. Remember that the dao, as the most general law, or principle, of reality (the Nature and human affairs) governing all instances of change, is representing the most general content of all instances of change. It is important to note that the most general content of all instances of change is formal—formal-logical, formal-descriptive-empirical-scientific, and normatic. Xunzi’s K-internalism holds

130

Ernest Sosa Encountering Chinese Philosophy

that in order to know the most general formal content of all changes, the heart-mind, as the internal cognitive agent, must detach herself from any specific instances of change, and thence (3.2) and (3.3). The crucial question for Xunzi here is obviously this: How can the heart-mind know the dao merely by pure internal introspection? The possibility of knowledge of the dao in this sense can only be guaranteed by a certain metaphysical position. More precisely, to support his K-internalism with respect to knowledge of the dao, Xunzi must hold the following: (3.4)

Knowledge of the dao is possible only if the heart-mind and reality (including the Nature and human affairs) share the dao.

Let me call the thesis that the heart-mind and reality share the dao “Xunzi’s nomictranscendentalism,” or simply “nomic-transcendentalism.” In other words, nomictranscendentalism holds that the dao of the heart-mind and the dao of reality are the same formal (formal-logical, formal-descriptive-empirical-scientific, and normative) dao, that is, the same great dao governing all instances of change. Thus, one can know the dao by merely internally introspecting the dao of the heart-mind, because the dao of the heart-mind and the dao of reality share the same great dao; that is, nomictranscendentalism holds. Therefore, Xunzi holds this: (3.5)

Nomic-transcendentalism grounds knowledge of the dao.

This can be formulated by means of an argument:8 1. The dao is the generalization of the dao of reality; that is, the unification thesis holds. 2. The dao of the heart-mind and the dao of reality (the Nature and human affairs) share the same general dao; that is, nomic-transcendentalism holds. Therefore, 3. If the heart-mind knows the dao of reality by merely internally introspecting its own formal contents—its own dao, then she knows the great dao which unifies the dao of the heart-mind and the dao of reality.

4 The result of a comparison: a new generalized definition of reflective knowledge According to Sosa’s mature theory, an apt belief—a belief that is correct attributably to a competence, as an intellectual virtue, exercised in its appropriate conditions—is an instance of animal knowledge, while an apt belief that the subject aptly believes to be apt, or a defensibly apt belief, is an instance of reflective knowledge. Here, I would like to draw attention to the several points I made at the end of Section 1 above, and I shall reformulate them as follows:

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao (4.1)

(4.2) (4.3)

131

When the subject aptly believes a certain apt belief to be apt, the apt belief coheres among other beliefs within the epistemic perspective of the subject, and therefore reflective knowledge fulfills a certain internalist coherence requirement for knowledge. Reflective knowledge is animal knowledge, but not vice versa. Sosa’s distinction provides a middle ground between reliabilism (mere animal knowledge) and internalism (reflective knowledge).

Consider now the notions of knowledge of changes and knowledge of the dao found in the Xunzi. Clearly, knowledge of changes, as knowledge of the obtaining and non-obtaining of instances of change, is animal knowledge, reflective knowledge, or knowing full well. How about knowledge of the dao? Surely, it is not animal knowledge, because it is purely internalist. It is also not reflective knowledge, nor knowing full well, understood by Sosa. This is because it is not animal knowledge and thus violates (4.2). However, as I shall argue below, there is good reason to generalize Sosa’s definition of reflective knowledge to include knowledge of the dao as reflective knowledge. An essential factor for a belief to be an instance of reflective knowledge is of course the faculty of reflection of the subject. In Sosa’s definition of reflective knowledge, which takes an instance of reflective knowledge to be an apt belief that the subject aptly believes to be apt, the faculty of reflection involved is characterized by the object of reflection and the aptness of reflection. Now the aptness of the subject’s believing an apt belief—the first order belief—to be apt should be essential, as that should be what makes the instance of knowledge to be reflective. However, must the object of the second order belief, which is the first order belief, be apt, in order for the second order belief to be able to be apt? I don’t think so—at least for cases concerning formal knowledge. Therefore, I suggest relaxing the requirement that the first order belief be apt, in order to generalize the notion of reflective knowledge. The generalization brings forth the following generalized definition of reflective knowledge: (4.4)

An instance of reflective knowledge is a true belief that the subject aptly believes to be true, if the content of the subject’s belief is formal, or apt, if otherwise; and vice versa.

Of course, according to this new definition, reflective knowledge, understood in the sense of Sosa’s definition, is still reflective knowledge defined by (4.4). But the new definition allows some other kinds of knowledge, including Xunzi’s knowledge of the dao, to be reflective knowledge. Let me explain why Xunzi’s knowledge of the dao is reflective knowledge under the new definition. First, emptiness, unity, and stillness should be considered as intellectual virtues of the heart-mind or the cognitive subject. They are also qualities of the faculty of reflection. So, when the heart-mind exercises the intellectual virtues of emptiness, unity, and stillness to attain the Great Pure Understanding, she forms the belief that a certain belief of law (regularity)—the first order belief with formal (formal-logical, formal-descriptive, and normatic) content—is true. (A belief with formal content is

132

Ernest Sosa Encountering Chinese Philosophy

hereafter called “a formal belief.”) In other words, in that case, the heart-mind aptly believes that a certain formal belief is true. The apt belief that the true first order formal belief is true is an instance of reflective knowledge. If Xunzi’s nomic-transcendentalism holds, then when the heart-mind attains the Great Pure Understanding, it grasps the most general coherence, or regularity, intrinsic to the dao of the heart-mind, and shared by the dao of reality (the Nature and human affairs). In this way, the heart-mind forms a true formal belief concerning the dao, and thus acquires an instance of knowledge of the dao, which is also an instance of reflective knowledge under the new generalized definition. The above also shows that if nomic-transcendentalism holds, then when the heartmind (or the cognitive subject) aptly—that is, so to speak, Great-Pure-Understandingly— believes a certain true formal belief to be true, the apt belief coheres among other beliefs within the epistemic perspective of the heart-mind, and thence a certain internalist coherence requirement for knowledge is fulfilled. The new generalized definition of reflective knowledge therefore retains the merit of Sosa’s definition that it provides a middle ground between reliabilism (mere animal knowledge) and internalism (reflective knowledge), as stated in (4.3). If nomic-transcendentalism holds, then there are instances of knowledge of the dao, which are also that of reflective knowledge by the new definition. It should be noted that in order for the generalized notion of reflective knowledge to be intelligible, it does not require nomic-transcendentalism to be true, but only metaphysically possible. For in a metaphysically possible world in which the dao of the heart-mind and the dao of reality share the dao, reflective knowledge is possible and thus the relevant notion is intelligible. Now it seems clear that nomictranscendentalism is at least metaphysically possible. But, of course, we are not sure if the metaphysical possibility actualizes (and I believe that it does not); and yet that there is such a metaphysical possibility is already a sufficiently strong reason for the generalization of Sosa’s definition of reflective knowledge to the new definition I put forward here.

Notes 1 Jason Baehr (2004) holds that Sosa first introduces the notion of intellectual virtue into philosophy in (Sosa 1980). 2 Hilary Kornblith (2004) holds that Sosa first introduces the distinction into philosophy in Sosa (1985). 3 For helpful discussions of the relevant issues, I would like to refer to Sosa (1991) and Greco’s “Introduction” to Sosa (2004). 4 I have argued for this translation, and interpretation, of the sentence (Cheung 2001), though I use “the Way,” instead of “the dao,” to translate the Chinese word “䚃” in that paper. 5 Knoblock’s translation of the last part of the passage is “main lines of a family may indeed also be interchanged.” I use a different translation because, first, the Chinese original contains the term “bian hua 䆺ॆ.” Second, I take the last phrase “yi yi bian hau yi ᣁӖ䆺ॆ⸓” as meaning that these are also changes and transformations. This

Sosa’s Reflective Knowledge, Xunzi’s Knowledge of the Dao

133

is at least a plausible interpretation. Knoblock’s translation “may indeed also be interchanged” has no ground at all. Why “interchanged”? 6 Instead of “the changes of human affairs,” Knoblock translates the term “shi bian” as “evolving affairs.” I take it to be clear from the context that “shi bian” refers to human affairs. 7 I do not follow Knoblock’s translation of the last sentence as “Through a myriad transformations nothing becomes disorderly” here. Instead, I translate the sentence as “The myriad things change but not anarchic [Ҳ luan].” This is because, first, my translation corrects Knoblock’s mistake of rendering “bian” as “transformation” instead of “change.” Second, “change but not chaotic” matches the meaning of the Chinese original much better than Knoblock’s “nothing becomes disorderly.” 8 I would like to thank Kasaki Masashi for having helped me formulate the argument.

References Baehr, Jason. 2004. “Virtue Epistemology.” Internet Encyclopedia of Philosophy. http://www. iep.utm.edu/virtueep/. Cheung, Leo K. C. 2001. “The Way of the Xunzi.” The Journal of Chinese Philosophy 28(3): 301–320. Greco, John. ed. 2004. Ernest Sosa: And His Critics. Oxford: Blackwell. Knoblock, John. 1988. Xunzi: A Translation and Study of the Complete Works, Volume I. California: Stanford University Press. Knoblock, John. 1990. Xunzi: A Translation and Study of the Complete Works, Volume II. Stanford, CA: Stanford University Press. Knoblock, John. 1999. Xunzi: A Translation and Study of the Complete Works, Volume III. Stanford, CA: Stanford University Press. Kornblith, Hilary. 2004. “Sosa on Human and Animal Knowledge.” In John Greco (ed.), Ernest Sosa: And His Critics. Oxford, Blackwell, pp. 126–134. Sosa, Ernest. 1980. “The Raft and the Pyramid: Coherence versus Foundations in the Theory of Knowledge.” Midwest Studies in Philosophy 5(1): 3–25. Sosa, Ernest. 1985. “Knowledge and Intellectual Virtue.” The Monist 68(2): 226–245. Sosa, Ernest. 1991. “Reliabilism and Intellectual Virtue.” In Ernest Sosa, Knowledge in Perspective: Selected Essays in Epistemology. Cambridge: Cambridge University Press, pp. 131–145. Sosa, Ernest. 1997. “Reflective Knowledge in the Best Circles.” The Journal of Philosophy 94(8): 410–430. Sosa, Ernest. 2007. A Virtue Epistemology—Apt Belief and Reflective Knowledge Volume One. Oxford: Oxford University Press. Sosa, Ernest. 2009a. “Knowing Full Well: The Normativity of Beliefs as Performances.” Philosophical Studies 142: 5–15. Sosa, Ernest. 2009b. Reflective Knowledge—Apt Belief and Reflective Knowledge Volume Two. Oxford: Oxford University Press. Sosa, Ernest. 2011. Knowing Full Well. Princeton, NJ: Princeton University Press.

134

9

Epistemic Virtues, the Gettier Problem, and the Rectification of Names1 Yingjin Xu

1 Introduction The most salient difference between Ernest Sosa’s treatment of beliefs and traditional treatment of beliefs in contemporary epistemology, among others, is that Sosa’s treatment is based on an analogy between “performance” in the most general sense, and beliefs, which are viewed as a specific type of performances. Hence, according to Sosa, just like one can meaningfully evaluate a performance like archery in terms of “accuracy,” “adroitness,” and “aptness” (or the “AAA-model” in short), one can also evaluate a performance like someone’s holding a belief in terms of “accuracy,” “adroitness,” and “aptness.” And in the epistemological context, the “accuracy” of a belief, according to Sosa, means its truth, the “adroitness” of producing a belief means its manifesting epistemic virtue or competence, whereas the “aptness” of a belief means that it is true because its holder is competent in producing it (Sosa 2007: 23). It is noteworthy that, in Confucian philosophy, the analogy between performance involving, say, Six Arts (namely, rites, music, archery, charioteering, calligraphy, and mathematics) and cognition also plays a salient role, although by “cognition” Confucian scholars basically mean the mental processes required by the treatment of moral/ political issues (e.g. how to pick up the most qualified officer from the candidate pool, or how to find the most suitable solution to a given moral dilemma in a certain context). Thus, given that descriptions like “accuracy,” “adroitness,” and “aptness” are fairly attributable to Six-Art-related performances, there will be no reason for a Confucian philosopher not to use this AAA-model to evaluate the functioning of a beliefproducing system as well as its products. Hence, prima facie, the building of a joint epistemological position both assimilating Sosa’s and Confucian resources may appear effortless, despite the seemingly troublesome issue of how to extend the scope of the typical Confucian narrative from political/moral topics to typical epistemological topics. Nonetheless, readers unfamiliar with the Chinese philosophical resources may still ask why the construction of the putative Sosa–Confucius alliance would be beneficial to Western philosophers, given that it is contemporary Anglophone virtue epistemology 135

136

Ernest Sosa Encountering Chinese Philosophy

that provides the most expressive meta-language via which the reconstruction of Confucian resources is possible, not vice versa. However, in my view, during the reconstruction of the relevant Confucian resources, the meta-language of virtue epistemology can be only employed on a high level, and when moved to a lower level, Confucian scholars may have a different point of view of the criteria on the aptness or adroitness of a belief-holding cognizer. Such difference itself is related to an underlying difference between the typical Anglophone treatment of the inter-belief relations and its Confucian counterpart treatment. To be more specific, from the contemporary Anglophone view inherited from the Fregean/Early-Wittgensteinian tradition, interbelief relations have to be construed in the light of truth-value-tables, the application of which is neutral with respect to the semantic content of the involved propositions. Hence, a belief ’s being deduced from another in accordance with the truth-function theory can be viewed as evidence for the first one’s justifiedness. By contrast, the Confucian counterpart notion of “justification,” namely, “the rectification of names,” mainly exploits the semantic relation between the subject-term of the belief and its predicate-term, and such relation is to be evaluated in the light of the norms embedded in corresponding predicate-terms. This treatment itself is underpinned by the Confucian point of view of language,2 according to which the epistemologically primary statuses of beliefs as truth-value-bearers have to be transferred to Chinese characters as constituting elements of sentences. Since the pictorial nature of characters is significantly related to character-cognizers’ relevant pictorial experiences, it would be extremely unnatural for these character-cognizers to merely use experience-neutral truth-tables or other logical devices to judge whether a belief is justifiably based on another. Hence, typical puzzles related to the notorious Gettier problem (Gettier 1963), which is problematic only when the employment of experience-neutral logical devices is taken for granted, would no longer bother Confucian scholars. Moreover, provided that Gettierized counterexamples also pose a threat to Sosa’s theory of knowledge attribution, a Confucian treatment of a Gettierized case, when assimilated into the putative Sosa–Confucius alliance, can make Sosa’s explanation of the competence in knowledge attribution even more convincing. However, before formally introducing the Confucian philosophy of language, it may still be helpful to provide a high-level sketch of Sosa’s theory, especially on his general attitude toward knowledge attribution. Meanwhile, I will suspend a formal treatment of the Gettier cases, since in the context of Sosa’s philosophy such a treatment is just a by-product of his speculations on the nature of knowledge attribution.

2 Sosa’s theory reformulated on a high level Obviously virtue epistemologists need to defend the commonsensical notion of knowledge, i.e. we human beings do know a lot of things concerning the external world, since the key concept addressed by virtue epistemology, namely “adroitness/ competence (of doing something related to the environment),” cannot make any sense if such knowledge is suspended. Hence, before being engaged with the Gettierized

Epistemic Virtues, Gettier Problem, Rectification of Names

137

cases, virtue epistemologists first need to handle the challenges posed by skeptics, from whose point of view no AAA-narrative can be meaningfully organized. Sosa’s refutation of brain-in-the-vat-type skepticism is that knowledge is commonsensically committed to basis-related safety, rather than basis-related sensitivity, whereas skeptics, in his view, mistakenly formulate the condition for knowing in terms of sensitivity, rather than safety (Sosa 2007: 27–28). Here “basisrelated sensitivity” mainly means that a belief has a basis that it would (likely) have if and only if the belief is true (Sosa 2007: 26), e.g. a warrior can easily have a perceptual basis of the belief that “this is a chariot” if and only if this really is a chariot. (Incidentally, there is a chance that such a warrior may be having a hallucination of the chariot in question, but the likelihood is too small to be seriously considered.) In contrast, “basisrelated sensitivity” means that a belief p has a basis such that if p were untrue, it would not be easy for that believer to believe p on the foregoing basis (ibid.); as, for instance, if the “chariot” that the warrior sees is a paper model of chariot, then it would not be easy for him to believe that it is a real one. However, what if, as a skeptic may continue, the appearance of the model of the chariot were amazingly similar to the real one? Under such a condition, to tell the difference between them would be a very intellectually demanding task, and hence, the difficulty of completing this task will cast doubt on a considerably huge part of our perceptually accumulated commonsensical knowledge, especially when a perceptual object is in contrast with its Cartesian-demon-inducing counterpart. But a Sosaadherent may simply eschew this objection by saying that to know the existence of a perceptual object does not require the capacity to tell the difference between it and its Cartesian-demon-inducing counterpart, since this requirement is already beyond the scope of the basis-related safety, rather than basis-related sensitivity. Or, in other words, given the difficulty of physically realizing the Cartesian-demon-inducing scenarios, to embrace these scenarios is of little practical value in ordinary activities of knowledge attribution. It is worth noting that Sosa’s objection to basis-related-safety-based skepticism is not the same as the so-called Moorean response to skepticism (cf. Black 2002) and the Pritchard-type safety theory (cf. Pritchard 2005), since Sosa is not apparently appealing to the notion of “possible world” as both Mooreans and Pritchard would be likely to do. (According to Mooreans, possible worlds including Cartesian-demon-manipulating scenarios have to be precluded from knowledge attribution mainly because they are too remote from the actual world; according to Pritchard, a belief ’s being safe can be defined by virtue of its being true not only in the actual world but also in its closest possible worlds.) The marginalization of possible-world-related conceptual tools in Sosa’s account, as a matter of fact, goes hand in hand with his highlights on the cognizer’s epistemic virtues, the accessibility of which to the cognizer’s own mental activities is in a sharp contrast with the externalist flavor of any possible-worldsemantics-based narrative. (It is worth noting that only an externally working knowledge ascriber, rather than a subject within the vignette, can fully achieve information required for any possible-world-based knowledge attribution, such as the information related to the individuation of possible worlds and the estimation of interworld distances.) However, on the other hand, Sosa’s virtue-based narrative is also

138

Ernest Sosa Encountering Chinese Philosophy

minimally open to externalism in the sense that the degree of any cognizer’s AAAfeatures is attributable only within a social context, within which a credit-calculating system works in accordance with its own blueprint (Sosa 2007: 95). Hence, a virtuebased narrative on knowledge attribution can overarch the gap between extreme internalism and extreme externalism. Sosa’s reluctance to take sides in the internalism–externalism debate has also influenced his treatment of the Gettier problem, which is supposed to be treating any conceivable internalism-oriented account of knowledge as justified true beliefs (JTB). Facing Gettierized counterexamples, the remedy of the definition of knowledge in the light of standard safety theory has to appeal to conceptual tools related to possibleworld semantics. However, given their externalist nature, one may naturally ask: how could these external factors, which may be inaccessible to a subject’s awareness, be easily cooperating with her internal mental life, in which the process of justification can be unpacked? I don’t think it is a question easy to answer for standard safetytheorists. By contrast, due to his distance from the standard safety-theory, Sosa can easily answer this question by resorting to the dispositional notion of “aptness,” which is not apparently externalist. Or, to be more specific, one’s aptness, according to Sosa, means that her belief is true just because of her epistemic competence, and to be competent is further defined in terms of “being steadily capable of basing the target belief on the right type of evidential belief.” Hence, a cognizer working in accordance with Sosa’s AAA-model can identify a typically Gettierized belief as a piece of nonknowledge mainly because her aptness in employing her epistemic competence allows her to do so. This epistemological picture may be appreciated from the Confucian point of view for the following reasons. First, like virtue epistemologists, Confucian scholars need to assume the reliability of commonsensical knowledge, given that the performance of, say, charioteering, has to be based on a solid ontological commitment of the very chariot in sight. Incidentally, although skepticism did not pose a sustaining threat to commonsensical beliefs in Chinese culture, in certain historical contexts wherein this threat appeared to be salient (due to, say, the invasion of Buddhist ideas), Confucian scholars would not hesitate to seriously defend commonsense, as both contemporary Mooreans and virtue epistemologists would do.3 Thus, Confucian scholars may share the same commonsense-friendly metaphysical intuition with Sosa. In my view, it is a very good starting point for building the putative Sosa–Confucius alliance. Second, like virtue epistemologists, Confucian scholars are more interested in ranking concrete persons’ intellectual characters or virtues rather than evaluating the deliverances of their cognitive processes. One of the most recommended virtues, according to Confucianism, is identified as a rarely seen capacity of applying the doctrine of the Mean, which is the highest order (Analects, 6:29). Given that the doctrine of the Mean is nothing but a requirement of maintaining balance and harmony wherein the mind could be directed to a state of constant equilibrium, it would not be so hard to derive an epistemologically interesting version of such a doctrine (e.g. such a derived doctrine may require that an apt cognizer has to maintain the balance between basing a true belief on a large evidence base and basing it on a very small evidence base, and so on.) Hence, in principle, one can extend the scope of

Epistemic Virtues, Gettier Problem, Rectification of Names

139

the traditional Confucian virtue-based narrative to cover virtues more relevant to knowledge attribution. Third, paralleled to Sosa’s identification of the process of earning more credits related to AAA-features as a socially confirmable process, Confucian scholars would also like to highlight the importance of having one’s virtue nurtured in a social context. As a matter of fact, most of the Six Arts listed in the Confucian philosophy of education are obviously social-context-involving: rites involves social hierarchy, music involves the audience, and charioteering involves teamwork, etc. Hence, Confucians may also feel comfortable about Sosa’s comment that, just like a single football player’s success is based on the whole team’s competence, one cognizer’s epistemic success is also somehow related to certain social environments like the reliability of the testimonial links distributed in the community (Sosa 2007: 95). But Confucians still have something more to say to enrich Sosa’s account. Sosa still owes his reader a more detailed account of the Gettier cases. His high-level strategy to refute Gettierized counterexamples is to appeal to the explanans–explanandum pair, namely, a belief is justifiably held only when its basing belief explains the very content of the target belief. However, Gettier case-builders, according to Sosa, have mistakenly regarded the fact that there being my true belief as an explanation of the fact that my belief is true, whereas the explanans–explanandum relation does not really hold there (Sosa 2007: 97). And an apt knowledge attributor is supposed to be able to debunk Gettier case-builders’ tricks competently. My worry about his strategy is: the explanans–explanandum pair is originally used in the narrative of philosophy of science, which largely involves reflective knowledge, whereas Sosa (Sosa 2007: 96 n. 1) still intends to treat the most typical Gettier cases by appealing to animal knowledge (namely, gut feeling-based knowledge) alone. Hence, Sosa still owes readers a further account of how Gettier case-related animal knowledge can be based on gut feelings when the explanans–explanandum pair is evoked. And this is the right place where Confucians can come to help. The Confucian solution is: to think by virtue of “the rectification of names” will suffice for the animal knowledge needed by the treatment of Gettier cases.

3 Gettier’s cases revisited in the light of “the rectification of names” Now it is the right time to scrutinize the two most discussed Gettier cases (Gettier 1963), as follows: ●

Case I: Smith and Jones have applied for the same job. But Smith has been told by the company manager that Jones will win the job. Smith combines this testimony with his observational evidence of there being ten coins in Jones’s pocket. And he proceeds to infer that whoever will get the job has ten coins in their pocket (let us call it belief b1). As it happens, too, belief b1 is true—although not in the way in which Smith was expecting it to be true. For it is Smith who will get the job, and Smith himself has ten coins in his pocket. So, does Smith know that b1? By

140

●

Ernest Sosa Encountering Chinese Philosophy

intuition, most readers of this vignette would say he does not know that; but he has to know that b1, which is a JTB, if a JTB-based definition of knowledge is assumed. Hence, most readers’ intuition cannot be compatible with a JTB-based definition of knowledge. Case II: Smith possesses good evidence in favor of the proposition that Jones owns a Ford. Smith also has a friend, Brown. Where is Brown to be found at the moment? Smith does not know. Nonetheless, on the basis of his accepting that Jones owns a Ford, he infers—and accepts—this disjunctive proposition: Either Jones owns a Ford, or Brown is in Boston (we will call this belief b2). Luckily, this proposition is true, but not because Jones really has a Ford (actually Jones lost his Ford due to a recent accident), but because Brown is really in Boston, although Smith has no evidence of where Brown is. Now the question becomes: does Smith know that b2? By intuition, most readers of this vignette would say that he does not know that; but he is supposed to know that b2, which is a JTB, if a JTB-based definition of knowledge is assumed. Hence, most readers’ intuition cannot be compatible with a JTB-based definition of knowledge.

A Sosa-inspired response to both cases may be like this: in both cases, Smith bases the acceptance of b1 on a wrong evidential belief, hence, the evidential beliefs cannot be really supportive of the target belief, and hence, neither b1 nor b2 is a piece of knowledge due to Smith’s lack of aptness of making the right explanans–explanandum pairs. By contrast, a Confucian scholar would simply cast doubt on the JTB-statuses of the target beliefs themselves by asking a further question: under what circumstances can a Gettier case-builder say that both target beliefs are justified? A Gettier case-builder may answer this question by appealing to deductive rules. To be more specific, b1 is justified because one can deduce it from Smith’s belief that Jones will get the job and Jones has ten coins, and b2 is justified because one can deduce it from Smith’s belief that Jones owns a Ford. Nonetheless, such operation is possible only when deductive validity is viewed as a core criterion for justification, and this is definitely not what a Confucian philosopher would like to endorse. Rather, a Confucian approach to the nature of justification is greatly based on the semantic relations between certain components of an integrated belief, rather than formal inter-belief relations, which could be easily highlighted in a Fregean framework. But do Confucians really have a serious theory of epistemic justification, given the marginal status of any epistemological topic in the typical Confucian narrative? I think we could reconstruct such a theory by making it paralleled to the Confucian notion of moral or political justification. In the arena of moral/political debates, the Confucian notion of “justification” is nothing but the notion of “the rectification of names.” The key citation in Analects relevant to my reconstruction is as follows: Zilu ᆀ䐟 asked, “If the Duke of Wei were to employ you to serve in the government of the state, what would be your first priority?” The Master answered, “It would, of course, be the rectification of names. . . . If names are not rectified, things will not be successfully accomplished. When things are not successfully accomplished, ritual practice and music will fail to flourish;

Epistemic Virtues, Gettier Problem, Rectification of Names

141

when ritual and music fail to flourish, punishments and penalties will miss the mark. And when punishments and penalties miss the mark, the common people will be at a loss as to what to do with themselves. This is why the gentleman only applies names that can be properly spoken and assures that what he says can be properly put into action.” Confucian Analects: 13.34

The core idea in Confucius’s words is basically that actual states of affairs in the world should exist in a way that is in accordance with “names” that are routinely used to depict them, otherwise their existence cannot be endorsed from a normative point of view. Or, put in another way, according to Confucius, if wrong names were to be attached to objects, wrong norms embedded in these wrongly attached names would be applied to the reality as well, and mismatching of this type would systematically produce social disorders of one form or another.5 A tentative epistemology-oriented reformulation of the foregoing notion of “the rectification of names” is as follows: Definition-1 (D-1): To say that “S is justified in holding a belief of the form R(a)” is tantamount to saying that “S believes that a is rectified under the name of R,” and this saying is true if and only if (hereafter “iff ”) other descriptions of a (other than “R(a)”) actually acquired by S are instantiations of norms encapsulated in the function R(x). Here “a” may refer to persons, physical objects (including artifacts), social constitutions, or even actions. And norms related to R(x) are assumed to be norms held by qualified justification-ascribers.6

Now we may use this pattern to analyze Case I of the Gettier cases. Prima facie, the target belief in Case I is that whoever will get the job has ten coins in their pocket. But Confucians routinely encourage a context-sensitive interpretation of the target belief rather than construe it according to its face-value. Hence, based on the information from the vignette, a Confucian scholar may simply reinterpret b1 as “Jones, who is supposed to be the job-winner, has ten coins in his pocket.”7 And thereby a b1holder has to presuppose that Jones will get the job. But according to D-1, if this presupposing belief is justified, the subject term “Jones” has to be rectified under the name of “job-competition winner,” hence, Jones has to be instantiating norms imbedded in the function “Job-competition-winner(x).” And no matter what norms will be involved here, according to D-1, they have to involve a social hierarchy in which a reliable testimonial link can be attributed, or in other words they have to be based on testimonies with enough social authority. In this sense, the belief that Jones will get the job is unjustified, because it is based merely on the manager’s oral report rather than some formal contract. Therefore, b1, which is based on this presupposing belief, is also unjustified. Now let’s turn to the Confucian analysis of Case II. This case is a bit more complicated, since what is involved here is a disjunctive proposition. Suppose that a Confucian scholar now barely accepts the deductive rule that a disjunctive proposition is true if and only if at least one of its disjuncts is true. Hence, she may deduce from this

142

Ernest Sosa Encountering Chinese Philosophy

that a disjunctive proposition is justified if and only if at least one of its disjuncts is justified. But she will immediately figure out that neither of the disjuncts in b2 is justified. First, obviously the disjunct that “Brown is in Boston” is not justified, since according to the vignette there is no Smith-available information indicating that Brown is instantiating the norms related to “being in Boston.” Second, the disjunct that “Jones owns a Ford” is unjustified too, since Smith’s evidence for supporting this belief is not qualified enough, otherwise he would not mistake Jones as a Ford-owner. (In my view, the norms required by the predicate “a persistent Ford-owner” do involve some fairly hard satisfying conditions, such as the persistent availability of the perceptual evidence concerning the existence of such a Ford, and so on.) Critics may contend that if the foregoing analysis were acceptable, there would be no justified but untrue beliefs, or, in other words, this pattern of analysis would allow anyone to use their hindsight bias to say that any untrue belief is unjustified. But we do have the intuition that some previously justified beliefs can prove to be false eventually. Hence, the foregoing analysis cannot be correct. However, a Confucianism-friendly rejoinder may be like this: typical cases involving the so-called “justified but untrue beliefs” are mostly instantiated by inquiries into complicated issues, scientific studies especially. But in ordinary life, a subject-term’s being subject to norms imbedded in the predicate would suffice for the factual correlation between the entities designated by the subject-term and those designated by the predicate-term. Moreover, since norms imbedded in the predicate-terms are routinely endorsed by social authorities, the failure of justifying a belief only happens when norms lacking enough authority are mistakenly introduced, or when subjectterms are being compared with the wrong type of norms. But there would be no chance for a genuinely justified belief to be false, because genuine norms themselves cannot be wrong. Critics may still contend that we can figure out a possible world in which all of the current social-authority-endorsing norms are wrong. But a Confucian can easily dismiss this objection by retaking Sosa’s strategy. For instance, she can simply say that what the notion of the rectification of names needs is merely safety, namely, if p is true, the subject will believe p, rather than sensitivity, namely, if p were untrue, the subject would not believe p. Or, to put it in another way, Confucians and Sosa share the same intuition that speculations on possible worlds other than the actual one are not significantly beneficial to a virtue-based epistemological approach. Now I can easily explain how a Confucian point of view can enrich Sosa’s account of the “aptness” required by a proper treatment of Gettier cases. “Aptness,” in a Confucian narrative, just means the competence of correlating the subject-term of the target belief with the right type of norms embedded in the predicate part of the same belief. In this sense, both beliefs held by Smith are not apt in Cases I and II just because he lacks the competence in doing the required matching job, and the lack of such competence may be further explained in terms of his credulous to unreliable testimonies, and so on. As to why such an account still keeps Sosa’s AAA-model within the scope of animal knowledge, the main reason is: the Confucian model does not evoke the explanans–explanadum pair on the inter-belief level, but only evokes

Epistemic Virtues, Gettier Problem, Rectification of Names

143

relations between different belief-constituents within the same belief. Hence, mental processes handling these relations may be executed more efficiently, which results in processes of knowledge attribution nearer to the level of gut feelings. By contrast, experience/content-neutral executions of deductive rules have proven to be very psychologically demanding (as shown by the Wason selection task, cf. Wason 1968), and hence not animal-knowledge-related. Now I still owe the reader a further account of why the Confucian notion of justification is so different from its Western counterpart. My explanation is related to the features of Chinese language, through which the Confucian ideas were originally formulated.

4 “The rectification of names” in the light of six principles of composing Chinese characters In the preceding analysis, I have assumed that the Confucian notion of justification/ rectification of names can be reconstructed in English. But actually if Confucius had taken English as his working language, chances would be little for him to entertain the notion of the rectification of names. The reason is quite simple: from the point of view of Chinese language, norms imbedded in predicates are not entities encoding propositions of, say, the “if . . . then” structure, but something pictorially revealed by the very structure of Chinese characters. Hence, the Confucian notion of the rectification of names is at least indirectly related to composing principles of Chinese characters. The rules guiding the composing of characters, namely, Six Principles of Composing Chinese Characters (or “liùshū”, ‫ޝ‬ᴨ), were firstly formulated by Xu Shen, a Confucian scholar-officer and philologist. Six principles include (cf. Xu and Duan 2008: 2102– 2103): (a) Simple Indicatives (zhˇı shì ᤷһ), whose forms are iconic without being based on concrete objects. (b) Pictograms (xiàngxíng 䊑ᖒ), whose forms are pictorial representations of the concrete objects. (c) Form and Sound (xíngshēng ᖒ㚢), namely, phono-semantic compound characters. (d) Compound Indicatives (huìyì ᴳ᜿), meaning that the compound in question is given meaning through the graphic combination and interaction of both constituent elements. (e) Reciprocally Glossing (zhuanzhu 䕹⌘), which arose from a divergence of one character into two separate characters still linked in sound and meaning. (f) Loaning Characters (jiajie ‫)ُٷ‬, which essentially indicates one character whose meaning has been broadened to contain multiple definitions. It is worth noting that the six principles do not function on the same level. Simple Indicatives and Pictograms work on the most fundamental level, and the Form and

144

Ernest Sosa Encountering Chinese Philosophy

Sound and Compound Indicatives have taken the first two as their basis. And Reciprocally Glossing and Loaning Characters are derived from the preceding four. An interesting implication of the relationships among the six principles is that all Chinese characters are either iconic or pictorial, to some degree. Hence, semantic norms embedded in any character are either iconic or pictorial, to some degree. This can quickly explain why it is pointless to explicate these norms in verbal language, especially in propositional logic. Just two extra examples strengthen this point.

4.1 Fú ㅖ Fú means “correspondence” in English, hence, if someone needs to translate “the theory of correspondence” into Chinese, he needs to use this character to form the phrase fuˇhélùn ㅖਸ䄆. Due to the importance of the notion of “correspondence” in both Chinese and Western Philosophy, a comparative study of Fú and “correspondence” seems to be fairly useful for showing the difference between two traditions’ approaches to “correspondence.” As we know, in the Western tradition, the notion of “correspondence” in “the theory of correspondence” means a relationship between a proposition qua a truth-value bearer and a state-of-affairs or a fact qua non-linguistic entities. Due to the heterogeneity between the two relata in the correspondence relationship, it is always a troublesome problem in Western philosophy on how to make one of the two relata match one another without losing the minimal respect of the foregoing heterogeneity. By contrast, the notion of “correspondence” in Xu Shen’s text means something quite different. Here is his definition: Xu Shen on “fú”: ㅖˈؑҏDŽ╒ࡦԕㄩˈ䮧‫ޝ‬ረˈ࠶㘼⴨ਸ (Xu and Duan 2008: 530). My paraphrase: fú signifies the reliability of testimony. According to the custom in Han Dynasty, a correspondence is made of bamboo, which is as long as six cun (c. 126 cm), and a pair of correspondence is nothing but two slices of a bamboo tube divided from one piece of bamboo, and they are supposed to be able to match each other if they are really from the same piece of bamboo.

It is obvious that the original connotation of “correspondence” in Chinese, according to Xu Shen, is more relevant to the notion of “tally” in English, accordingly, derived Chinese phrase like huˇfú 㱾ㅖ have to be translated as “tiger-shaped tally.” The relevance between the reliability of testimony and fú could be illustrated via the functioning of a tiger-shaped tally: such a tally was routinely issued to generals as the imperial authorization for troop deployment in ancient China. The tally itself was carefully divided into two halves, and a troop-deploying document is justified when and only when the half in the hand of the document-deliverer perfectly fits the half in the hand of the commander of the relevant unit of troops. Hence, it is not hard to conclude that “correspondence” in Chinese is not construed as a relationship between a state of affairs and a proposition-encoding belief, but a relationship between two pictorially representable objects, which are supposed to be constituents of a onceintegrated object. This feature of “correspondence” naturally makes two relata of this

Epistemic Virtues, Gettier Problem, Rectification of Names

145

relation homogeneous to each other and hence makes one relatum’s tallying with another no longer metaphysically mysterious. This feature is also proceeded by the notion of the rectification of names, according to which the subject-term’s tallying with the predicate-term is not a relationship overarching mutually heterogeneous entities.

4.2 Hé ઼ and huò ᡆ If the two characters are construed in accordance with modern Chinese, hé means “and” and huò means “or” in English, hence, the two symbols here are critically important for formulating conjunctive/disjunctive reasoning in modern Chinese. And, as we have seen, according to the rules of Fregean propositional logic, two conjuncts/disjuncts in one conjunction/disjunction do not need to be semantically relevant to each other, otherwise no one could deduce “Brown is in Boston” from “Either Jones owns a Ford, or Brown is in Boston” in the second case of the Gettier problem. By contrast, in classical Chinese, both hé and huò do mean something different. What is provided here is Xu Shen’s definitions of the two terms: Xu Shen on “hé”: ઼ˈ⴨䅽ҏ. My paraphrase: hé, means mutually echoing. Xu Shen on “huò”: ᡆˈ䛖ҏDŽӾഇӾᠸˈԕᆸаDŽаˈൠҏDŽฏˈᡆ৸Ӿ൏. Xu and Duan 2008: 1760 My paraphrase: huò means small countries, and this character is semantically composed of koˇ (mouth) and gē (dagger-axe), the combination of which means the defense of one direction of land. Here the “oneness” means land. If huò is combined with another character, namely, tuˇ ൏ (earth), then a new character could thereby be formed, namely, yù ฏ (field).

Xu Shen’s interpretations of both characters are philosophically inspiring. Insofar as hé is concerned, since it refers to the process of echoing, and an echo is a reflection of an original sound, a conjunctive belief in Chinese has to involve some minimal similarity between two conjuncts. As to huò, though Xu’s interpretation of it looks a bit more complicated, the relevance between “huò” and the status of a small country under military protection is still quite visible here. According to Duan Yucai (⇥⦹㻱, 1735– 1815), an exegetist specializing in annotating Xu Shen’s works, the meaning of “doubt” could be derived from huò in the sense that mutual trust is hard to achieve among different defenders of different small countries (Xu and Duan 2008: 1761). Obviously the contemporary connotation of huò, namely, “or,” is further derived from the meaning of “doubt.” However, even this doubt-based meaning of “or” is itself based on the assumption that different country defenders are at least mutually competing with each other, and this assumption itself is based on a further assumption that different competitors are interested in the same type of resources, the quantity of which has to be limited. In this sense, the meaningfulness of having doubts on “whether A or B is the case” has to be based on the observation that A and B have to be minimally semantically relevant to each other as well. Hence, it would be very hard for a Chinese

146

Ernest Sosa Encountering Chinese Philosophy

mind to understand why “Brown is in Boston” can be deduced from “Either Jones owns a Ford, or Brown is in Boston,” given that the two disjuncts here are not relevant to each other at all. With the help of these illustrations, we could re-formulate the very notion of the rectification of names/justification in the following way. Given that both subject-terms and predicate terms have to be expressed in terms of characters in Chinese, the composition of each belief is actually a matter of how to negotiate different meanings of different constituting characters. Hence, if A and B represent two characters, the justifiedness of the belief “A is B” would be dependent on (1) whether the object depicted by A corresponds to the semantic norms imbedded in A; (2) whether the semantic norms imbedded in A correspond to the semantic norms imbedded in B. It is worth noting that both relations of correspondence have a direction from lower-level entities to higher-level entities (e.g. from subject to predicate, not vice versa), and, accordingly, norms embedded in subjects have to be subject to the norms embedded in predicates, otherwise the resultant belief would not be justified or rectified. The foregoing picture presupposes that all characters are norm-encapsulating. But why? Why are characters not just results of some accidental conventions? The Confucian response to this question could be found in Xu Shen’s postscript to his Explaining Graphs and Analyzing Characters, where legendary figures like Cangjie ‫ع‬乑 are listed as creators of characters. Since these figures are described as cultural heroes who could approach the holy meanings distributed in the heaven/sky, their deliverances, namely, characters, have to encode such sacred information as well (Xu and Duan 2008: 2100). Though his semi-religious explanation of the origin of the norms imbedded in characters may not satisfy today’s naturalism-oriented philosophers, it can precisely explain why Confucians are routinely labeled as traditionalists, namely, people highly respecting traditions created by some putative cultural heroes. Another epistemologically interesting implication of the Confucian respect of norms is that even the term “know” in ancient Chinese (zhīdào ⸕䚃) is norm-oriented rather than belief-oriented. To be more specific, although in modern Chinese, the combination of the two characters (zhi ⸕ and dao 䚃) means “(to) know” in English, in classical Chinese of the era of Confucius, this combination is actually semantically equivalent to two words in English: “(to) know” as a propositional attitude and the philosophical term “dao” as its object, whereas dao can also be interpreted as “a way of following norms.” Therefore, it makes no sense to add a further object clause following zhidao ⸕䚃 in classical Chinese. Incidentally, the lack of object clauses following zhi ⸕ may also explain why the notion of “truth” is so marginal in the ancient literature in China, since “truth” is usually used to make a higher-order evaluation of what is encoded in object clauses. This can additionally explain why truth-value-table-based justification appears to be so weird from a Chinese point of view.

5 Conclusion The main thesis of this chapter is that the Confucian notion of the rectification of names provides a better account of how an apt knowledge attributor can use his animal

Epistemic Virtues, Gettier Problem, Rectification of Names

147

knowledge alone to handle the Gettier cases within a bigger framework of Sosa’s virtue epistemology. Since this Confucian notion looks so natural when Chinese is taken as the working language, this research also shows the potential capacity of Chinese for handling contemporary philosophical issues.

Notes 1 This research is sponsored by the National Foundation of Social Sciences of China (No. 15ZDB020). 2 This is explicated in Confucian scholar Xu Shen 䁡᝾ (58?–147? AD) in his postscript to Explaining Graphs and Analyzing Characters (see Xu and Duan 2008). 3 Incidentally, Fan Zhen 㤳㕌 (450–510 AD) is a typical Confucian scholar of this type, but I simply have no space to discuss him here. 4 The English translation I am citing here is taken from Slingerland (2003). 5 A caveat here: in classical Chinese, the term “name” (míng ਽) refers to both proper names and general names rather than proper names alone, and it definitely refers to general names in the context of the rectification of names. Hence, when the relevant Confucian narrative is being reconstructed in the light of modern logic, it is more suitable to characterize Confucian “names” as propositional functions or predicates. 6 Another caveat here: the employment of the Fregean-Russellian notation in the preceding definition is just an expedient measure. Actually the Fregean assumption that a simplest proposition is composed of a name as a saturated element and a function as an unsaturated element may be rejected by a Confucian philosopher of language like Xu Shen, but for the purpose of an analytic-philosophy-friendly pedagogy, this deviation from the original meaning of Confucianism is temporarily tolerable. 7 I concede that such a reading is greatly inspired by Keith Donnellan’s (1966) distinction between the “attributive use” of a description and the “referential use” of it. Hence, in this context, “whoever will get the job,” if referentially used, is just a tool temporarily employed to pick out the target person in Smith’s mind.

References Black, T. 2002. “A Moorean Response to Brain-in-the-Vat Skepticism.” Australasian Journal of Philosophy 80(2002): 148–163. Donnellan, K. 1966. “Reference of Definite Descriptions.” Philosophical Review 75: 253–355. Gettier, E. L. 1963. “Is Justified True Belief Knowledge?” Analysis 23: 121–123. Pritchard, Duncan. 2005. Epistemic Luck. Oxford: Oxford University Press. Slingerland, E. 2003. Confucius Analects: with Selections from Traditional Commentaries. Indianapolis/Cambridge: Hackett Publishing Company. Sosa, E. 2007. A Virtue Epistemology: Apt Beliefs and Reflective Knowledge, vol. 1. Oxford: Oxford University Press. Wason, P. C. 1968. “Reasoning about a Rule.” Quarterly Journal of Experimental Psychology 20(3): 273–281. Xu, Shen and Duan, Yucai. 2008. Explaining Graphs and Analyzing Characters (with commentaries made by Duan, Chinese version). Beijing: The China Drama Publisher.

148

10

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way Out of the Ontological Predicament Yiu-ming Fung

1 Introduction Based on his interpretation of virtuous circularity in Descartes’ philosophy, Sosa provides a reliabilist reading of Davidson’s epistemology. He thinks that Davidson’s coherentism is not viciously circular, because a kind of virtuous circularity may be found below the surface in Davidson’s epistemological writings. One of the aims of this chapter is to demonstrate that this may not be a true account of Davidson’s theory. I argue that Davidson’s “triangulation,” together with his “correspondence without confutation,” does not commit to a reliabilist assumption, although Davidson’s externalism has its own problems, which are not addressed by Sosa. Sosa’s own theory of knowledge is based on a kind of epistemological competence together with a kind of foundational justification in terms of causal aetiology from experience to belief. I think his theory is a reasonable option which provides a middle way between classical foundationalism and coherentism. But, there is still a question for Sosa to answer. The question is: in what sense is experience as the foundation of knowledge objective or, as a starting point of the aetiology, coming from an identifiable source in the world? To borrow Johnston’s words, in what sense can the appearing be tracked by the appearance?1 If our experience is something which appears in our mind, it is personal and subjective. The question of the objective source or origin of our experience or the question of bridging the gap between objective reality and our experience is still unanswered. I think the trouble of searching for ontological source or ground of knowledge can be solved by Zhuangzi’s 㦺ᆀ proposal. That is: reality (or the external world in itself) does not play any epistemological role in forming our knowledge. Because reality is nothing. In this chapter, I also want to argue and suggest that Zhuangzi’s idea of “chaos” (hundun ⑮⊼), “undifferentiated oneness” (hunyi ⑮а), or “nothingness” (wu ❑) can offer a way out of the current epistemological predicament in searching for external objectivity and may be helpful for Sosa to answer or to switch the question that I think he has not yet addressed. 149

150

Ernest Sosa Encountering Chinese Philosophy

2 Sosa’s reliabilist reading of Davidson and the ontological predicament In contrast with the view of reducing external reality to subjective experience, Sosa thinks that there is an alternative option which “does not rely on any ontological reduction of the world to the mind.” But “it attempts to argue its way out from the internal to the external nevertheless, as did Descartes, for example, or in some other way” (Sosa 2009: 114). After rejecting the interpretation of Davidson as committing reduction (or deduction) from B to W (i.e. from the realm of one’s beliefs to the realm of the world) or W to B and a priori argument, Sosa continues: Two fascinating questions ensue. First of all, isn’t Davidson now drawing on externalist, and indeed reliabilist intuitions? It would seem to be the high level of reliability of our empirical beliefs, given his account of meaning and content, that now serves as the core of the special source of justification invoked to explain the high epistemic status of our empirical, and especially of our perceptual beliefs. The second interesting question concerns the status of Davidson’s theory and his “answer” to the skeptic. If the source of justification should now be viewed as distinct from any reasoning, from any invoking of a justifying argument, then it is no longer clear why it must be a priori. (Not that it was all that clear in any case.) It becomes positively opaque why the a priority of Davidson’s epistemologically effective reasoning should be an issue. Now it would seem to matter only that the reasoning establish the theory, for it is just the truth of the theory that has turned out to be epistemologically effective. What seems to matter is essentially that as things in fact stand in our contingent circumstances, content is set externalistically through causal linkages with our external environment. For it is through this fact that the reliability of our beliefs is assured. And it is from their assured reliability that their presumptive justification derives. Sosa 2009: 128–129

Based on this reliabilist reading, Sosa thinks that Davidson’s naturalized epistemology does not commit a vicious circularity. It is because: In Descartes’ case the effective fact would involve assent attendant on sufficiently clear and distinct perception, while favored by God’s epistemic benevolence. In Davidson’s case the effective fact would involve rather that we would not form beliefs having the contents of our empirical beliefs did we not interact appropriately with surroundings characterized generally as are our surroundings in this world. Sosa 2009: 129–130

Here, Sosa regards Davidson’s naturalized epistemology as one of the successful examples which “has swept away the priority of the radically subjective, and has moved epistemology beyond the inner, the internal, the foundational, towards a view of the subject as embedded in fundamental ways within his social and natural environment.

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

151

The mind is thus viewed as extended out into the world through its external contents” (Sosa 2009: 108). Nevertheless, I don’t think Sosa’s interpretation is a true story of Davidson’s theory. In his “The Problem of Objectivity,” Davidson does acknowledge the insight of Descartes for establishing his theory’s starting point. But he doesn’t follow Descartes to develop his theory or his theory doesn’t have, as Sosa claims, similar features to those in Descartes’s reliabilist epistemology. Let’s see Davidson’s summary at the end of the essay: Since Descartes, epistemology has been based on first person knowledge. We must begin, according to the usual story, with what is most certain: knowledge of our own sensations and thoughts. In one way or another we then progress, if we can, to knowledge of an objective external world. There is then the final, tenuous, step to knowledge of other minds. I argue for a total revision of this picture [my italics]. All propositional thought, whether positive or skeptical, whether of the inner or of the outer, requires possession of the concept of objective truth, and this concept is accessible only to those creatures that are in communication with others. Knowledge of other minds is thus basic to all thought. But such knowledge requires and assumes knowledge of a shared world of objects in a common time and space. Thus the acquisition of knowledge is not based on a progression from the subjective to the objective [my italics]; it emerges holistically, and is interpersonal from the start. Davidson 2004: 18

I think Davidson’s objectivity of knowledge or true beliefs is based on triangulation, which does not have and needs no reliabilist helper. It is not a project to treat “the mind as extended out into the world through its external contents” (Sosa 2009: 108). In other words, it is not “inside out,” nor “outside in.” Triangulation, for Davidson, is supposed to make our knowledge without massive errors; but it is not based on something reliable in the mind. So, Davidson distances from Descartes by saying that: My strategy, as I remarked at first, is in an important respect Cartesian. I have begun with the fact that I think, and I have asked what follows from that fact. . . . Aside from the starting point I share with Descartes, however, my epistemology, if that is the right word for what I am doing, is almost totally non-Cartesian [my italics], for I do not assume, as Descartes and endless idealists, empiricists, and rationalists have, that empirical knowledge depends on indubitable beliefs, or something given to the mind which is impervious to doubt, nor that the contents of our beliefs may in principle be independent of what lies outside us. In other words, I am an antifoundationalist, and I have left the door open for some form of externalism. Davidson 2004: 17

The non-Cartestian feature of Davidson’s theory lies in its different assumption from Descartes’. He does not assume something like reliable beliefs from reliable faculty.

152

Ernest Sosa Encountering Chinese Philosophy

Instead, his assumption is that at least two subjects (the speaker and the interpreter) engage in interaction with the external cause of the shared environment. Without the intersubjective interaction, there would be no common concepts. Without the external cause which stimulates their perceptual responses in interaction, there would be no convergence of common referent and no sense of objectivity. So, the three together in a holistic way contribute to the impossibility that our beliefs are massively false. In this way, the external cause plays an active role which does not, as claimed by Sosa, “rest on a subject-centered conception of epistemic justification as intellectual virtue” (Sosa 2009: 152). In his interpretation of Davidson’s epistemology, Sosa mentions a sort of a priori argument in Davidson’s theory: Despite the epistemologically coordinate status of the mind and the world, we are told, the content of the mind can be shown to entail how it is out in the world. According to Davidson, our beliefs could not possibly have the contents that they have, unless the world around us was pretty much the way we take it to be, at least in its general outline [my italics]. We are thus offered a way to argue, to all appearances a priori, from how it is in our minds to how it is in the world. The argument is a priori at least in being free of contingent premises or assumptions about the world around us or our relation to it. From premises about the contents of our propositional attitudes, the argument wends its way to a conclusion about how the world around us is structured and populated [my italics]. Sosa 2009: 109

It seems that the above description echoes Davidson’s own words in the following long citation: Consider how we discover what some simple sentence means, say “There’s a table” or “Here’s a piece of green paper”. Our basic evidence is that the speaker is caused to assent (not just on this occasion, but generally) to these sentences by the presence of tables or pieces of green paper, while the absence of these objects causes him (generally) to dissent from the same sentences. I do not think of assent and dissent as overt speech acts, but as attitudes towards sentences sometimes revealed in speech and sometimes in other ways. My main point is that our basic methodology for interpreting the words of others necessarily makes it the case that most of the time the simplest sentences that speakers hold true are true. It is not the speaker who must perform the impossible feat of comparing his belief with reality; it is the interpreter who must take into account the causal interaction between world and speaker in order to find out what the speaker means, and hence what that speaker believes. Each speaker can do no better than make his system of beliefs coherent, adjusting the system as rationally as he can as new beliefs are thrust on him. But there is no need to fear that these beliefs might be just a fairytale. For the sentences that express the beliefs, and the beliefs themselves, are correctly understood to be about the public things and events that, cause them, and so must be mainly veridical. Each individual knows this, since he knows the nature of speech and belief. This

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

153

does not, of course, tell him which of his beliefs and sentences are true, but it does assure him that his overall picture of the world around him is like the picture other people have, and is in its large features correct [my italics]. Davidson 2001: 174

To compare these two passages quoted above, I think Sosa’s description (i.e. “our beliefs could not possibly have the contents that they have, unless the world around us was pretty much the way we take it to be, at least in its general outline”) and Davidson’s own words (i.e. “his overall picture of the world around him is like the picture other people have, and is in its large features correct”) do not mean the same thing. For Sosa, the antecedent of the counterfactual is about “the world around us was pretty much the way we take it to be,” that is about the content we assign to the world. Moreover, he also regards this assignment or argument as a priori. But I think this is not Davidson’s transcendental or a priori argument. Instead, his own transcendental presupposition is:2 What stands in the way of global skepticism of the senses is, in my view, the fact that we must, in the plainest and methodologically most basic cases, take the objects of a belief to be the causes of that belief [my italics]. And what we, as interpreters, must take them to be is what they in fact are. Communication begins where causes converge: your utterance means what mine does if belief in its truth is systematically caused by the same events and objects. Davidson 2001: 151

For more details, I quote from Davidson: Given these three patterns of response [i.e. the child finds tables similar; we find tables similar; and we find the child’s responses in the presence of tables similar] we can assign a location [my italics] to the stimuli that elicit the child’s responses. The relevant stimuli are the objects or events we naturally find similar (tables) which are correlated with responses of the child we find similar. It is a form of triangulation: one line goes from the child in the direction of the table, one line goes from us in the direction of the table, and the third line goes between us and the child. Where the lines from child to table and us to table converge, ‘the’ stimulus is located [my italics]. Given our view of child and world, we can pick out “the” cause of the child’s responses. It is the common cause of our response and the child’s response. Davidson 2001: 119

As I understand, Davidson’s transcendental argument is from the objects of a belief to the causes of that belief, or from the convergence of three patterns of response in the presence of tables to the location of “the” stimulus. So, what is assigned or located is not the mind-independent content or, in Sosa’s words, what is “structured or populated” in the world. Nevertheless, it seems to me that Davidson’s own problem or trouble is that: it is difficult, if not impossible, for Davidson to identify “the” stimulus and thus ask the

154

Ernest Sosa Encountering Chinese Philosophy

skepticism of the external world to get lost. Because the objects of a belief in one’s mind may not be identical with the external causes or may not be a true reflection, in some sense, of the external objects, though based on social interaction, we may identify the external causes or objects (C) as the objects of a belief (O of B) or to take the objects of a belief (O of B) to be the external causes or objects (C). In Kantian or some other sense, these external something out there are transcendent of our mind and thus may not be included (by virtue of the act of intentionality) in the furniture of our experience. Based on social interaction or triangulation, to take O of B to be C cannot exclude the possibility that O of B is not C or O of B is not a true reflection, in some sense, of C, though it may be helpful to form a world of intersubjectivity. Moreover, Davidson seems also treating his meta-belief of taking O of B to be C as evidence to justify objectivity. In other words, the objective base is not the external cause per se in triangulation, but the meta-belief of taking the objects of a belief to be the external causes. Even though triangulation may be helpful for forming a common cause (or a seeming common cause) in collective intentionality, the meta-belief of taking the collectively intended object to be the external something out there is coming from a transcendental argument. To escape from this trouble or predicament, I think, one of the strategies adopted by philosophers is Husserl’s idea of giving up the natural attitude by bracketing or suspending (epokhē) this something in terms of phenomenological reduction. I think to identify a real cause of our experience in the external world is not an easy job; for example, as we are told by biologists that the color or shape sensibility of birds is different from that of humans. Suppose that, in an owl’s body, it has different sensory structure in its nervous system from ours; and suppose further that something in the external world (say, a target seen by an owl which is called by humans red apple) causes an owl’s sensory system and thus makes it receive a particular experience of the kind in the owl’s mind. Based on its experience, the owl forms a belief that there is a red apple of the kind in the owl’s perception. Here my question is: if it is possible that the object in the owl’s perceptual belief (say, a red apple of the kind in the owl’s mind) is not similar to the object in a human’s perceptual belief (say, a red apple of the kind in a human’s mind), how could we justify that we can take the object of our perceptual belief and the object of the owl’s perceptual belief to be the same cause (i.e. that very something in the external world)? If there is no one-to-one mapping, how can we make sure that both kinds of perceptual object are coming from the same external cause in their respective causal chains? That is the reason why I am not satisfied with Davidson’s transcendental argument, which says that: “What stands in the way of global skepticism of the senses is, in my view, the fact that we must, in the plainest and methodologically most basic cases, take the objects of a belief to be the causes of that belief ” (Davidson 2001: 201). Davidson’s triangulation may be helpful to build a world of intersubjectivity, but it does not guarantee the rejection of skepticism, especially the type of “brains in a vat.” For brains in a vat, it is possible that the super-scientific computer is able to give a systemic stimulation which leads them to experience a similar kind of convergence of responses. It seems to Davidson that the origin or source of objectivity is the external cause though it cannot play the role of reason. But, the identification or assignment of

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

155

the external cause by Davidson’s transcendental argument is not better than Kant’s or Santayana’s. They both think that what appears in our mind or what is perceived in our mind is causally related to or affected by something unknown (Kantian “Thing-inItself ” or Santayana’s “Efficacious Reality” or “God”). Davidson’s argument, like Kant’s and Santayana’s, is not sound, because all their arguments go beyond the empirical to the non-empirical or trans-empirical realm. For Kant and Santayana, that is a kind of noumena; for Davidson, that is a kind of area in which the entities in themselves are without concepts or language. In other words, the objects of one’s belief are minddependent while the external causes are mind-independent. The mind-independent causes are without content otherwise they can be conceptualized and expressed in propositions before using Davidson’s transcendental argument to identify them. However, Davidson is confident to believe that we can be successful to identify the common cause. His reason is that: Social interaction, triangulation, also gives us the only account of how experience gives a specific content to our thoughts. Without other people with whom to share responses to a mutual environment, there is no answer to the question what it is in the world to which we are responding. The reason has to do with the ambiguity of the concept of cause. It is essential to resolve these ambiguities, since it is, in the simplest cases, what causes a belief that gives it its content. In the present case, the cause is doubly indeterminate: with respect to width, and with respect to distance. The first ambiguity concerns how much of the total cause of a belief is relevant to content. The brief answer is that it is the part or aspect of the total cause that typically causes relevantly similar responses. What makes the responses relevantly similar in turn is the fact that others find those responses similar; once more it is the social sharing of reactions that makes the objectivity of content available. The second problem has to do with the ambiguity of the relevant stimulus, whether it is proximal (at the skin, say) or distal. What makes the distal stimulus the relevant determiner of content is again its social character; it is the cause that is shared. The stimulus is thus triangulated; it is where causes converge in the world. Davidson 2001: 129–130

Nevertheless, Davidson’s “the” stimulus located by triangulation is proximal, not really distal. Experimental psychologists have demonstrated that indefinitely many distal stimuli can give rise to the same proximal stimuli.3 So, whether we can identify the common cause through triangulation is not an easy question to answer. In the next section, let’s examine Sosa’s middle way.

3 Sosa’s middle way between foundationalism and coherentism Sosa is not satisfied with both classical foundationalism and coherentism. He thinks that, on the one side, philosophers such as Schlick, Hempel, C. I. Lewis, and Chisholm who hold the view of empirical justification fail to find foundations contentful enough to found our rich knowledge of an external world; on the other side,

156

Ernest Sosa Encountering Chinese Philosophy

Neurath, Sellars, Rorty, and Davidson, among others, claim that experience bears causally on our beliefs without the power of justification (Sosa 2009: 88). Here, Sosa finds that this controversy leaves the middle ground untouched. Because both classical foundationalism and coherentism have a blind spot covering by their sharing a problematic assumption (A): (A) Experience can bear epistemically on the justification of belief only by presenting itself to the believer in such a way that the believer directly and noninferentially believes it to be present, and can then use this belief as a premise from which to reach conclusions about the world beyond experience. Sosa 2009: 88–89

Sosa believes that we can go beyond the traditional controversy by rejecting assumption (A). One of the reasons is that: “Experience can bear epistemically on the justification of a perceptual belief by appropriately causing that belief.” For example, “while viewing a snowball in sunlight I may have visual experience as if I see something white and round, which may prompt the corresponding perceptual belief. In that case it will be an important part of what makes my perceptual belief epistemically justified—and indeed of what makes it a perceptual belief—that it is caused by such experience” (Sosa 2009: 89). But is this sufficient to provide foundational justification for our perceptual beliefs? Sosa continues to explain his view in the following: Take a perceptual belief prompted appropriately by a corresponding experience. Take a belief that “this is white and round,” one prompted by visual experience of a sunlit snowball in plain view. Is that perceptual belief foundationally justified simply in virtue of its causal aetiology? When Sellars inveighs against the Myth of the Given, he targets not only the radical version of the myth involving direct apprehensions of given experience. He objects also to the more moderate version that postulates foundational knowledge through perception. Indeed the key passage that encapsulates his opposition to a foundational epistemology targets not a foundation of introspective direct apprehension but a foundation of perception. Sosa 2009: 89

The last sentence clearly indicates that Sosa’s foundation is different from Sellars’ in the sense that Sosa does not treat knowledge as grounding on any foundation of perception, but defines knowledge as true belief held “out of intellectual virtue.” His epistemological theory is carefully constructed. He also examines a lot of examples counter to his theory and provides convincing arguments to explain them away. So, I regard his project as promising. Nevertheless, as mentioned in the previous sections of this chapter, I think the question of objectivity or the objective source or origin of knowledge is not well addressed by Sosa. According to Sosa, there is a necessary condition for knowledge of either sort, animal or reflective:

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

157

(VR) A belief amounts to knowledge only if it is true and its correctness derives from its manifesting certain cognitive virtues of the subject, where nothing is a cognitive virtue unless it is a truth-conducive disposition. Sosa 2009: 135

In addition to (VR), which is more closely adequate as an account of animal knowledge than of a reflective one, reflective knowledge requires a specific further condition, namely perspectival endorsement of the reliability of one’s sources. To sum up, he says: Virtue Reliabilism is true both for animal and for reflective knowledge. Prominent among values of the higher, reflective level is that of understanding. It is in part because one understands how one knows that one’s knowing reaches the higher level. A belief constitutive of reflective knowledge is a higher epistemic accomplishment if it coheres properly with the believer’s understanding of why it is true (and, for that matter, apt, or true because competent), and of how the way in which it is sustained is reliably truth conducive. Cohering thus within the believer’s perspective is, moreover, not irrelevant to a belief ’s being deeply attributable to the believer’s epistemic agency. Guiding one’s thinking with sensitivity to the truth would seem to involve some perspective on how one is forming and sustaining one’s beliefs. Of course one knows plenty through one’s animal nature, sans rational agency; which is how we know some of the things we know best. Even when one could take charge, finally, as a deliberative rational agent, it may be best to proceed on automatic pilot. But we do often take pride in grasping the truth through its deliberate pursuit, which hence is also valued as a positive accomplishment. Sosa 2009: 138

It seems that Sosa doesn’t touch the ontological problem about the objective source or origin of our knowledge. As I understand it, there are at least five kinds of ontological view or attitude in relation to epistemology: (A) To be is to be instantiated or manifested in phenomena which can be perceived or known. (Metaphysical Realism: Platonic Ideas) (B) To be is to be represented in terms of its correspondence to our perceptual belief. (Rorty’s Mirrored Reality or Davidson’s Unorganized Content) (C) To be is to be, in a mystical way, causally related to what is perceived or what appears in our mind. (Kantian “Things-in-Themselves” or Santayana’s “God”) (D) To be is to be perceived. (Subjective Idealism: Berkeleian ideas) (E) To be is to be suspended or bracketed. (Husserl’s Phenomenological Reduction: Naturalistic Objects vs. Intentional Objects) Views (A), (B), and (C) presuppose two levels of world: a realm transcendent of empirical knowledge (say, Platonic noumena) or a realm without language and thought (say, Buddhist nirvana) and a realm of our empirical knowledge. I think all views embedded with two levels of world would be subject to challenge from skepticism: if

158

Ernest Sosa Encountering Chinese Philosophy

the transcendent realm (such as the realm constituted of Kant’s things-in-themselves or the realm of Santayana’s efficacious reality or God) is without language and thought, its relation to the empirical realm cannot be explained in a sensible way;4 If the transcendent realm can be expressed with a very special language (such as the Zen language in Buddhism), which is incomparable or incommensurable to ours, it still cannot be understood by us intelligibly. Even though the Platonic noumena or intellectual world can be explained with discursive language, its relation to the phenomena (such as instantiation or participation) is still not explained in a sensible way. In contrast, (D) and (E) can be interpreted in some sense as free from the challenge of skepticism or as an escape from the challenge. The view of (D) entails that there is no non-subjective external world while the strategy of (E) gives no epistemic role for the external world to play in intentionality or empirical knowledge, by giving up the “natural attitude” whose ontological commitment to an external world is naturalistic. All these views have their own problems, either their explanation of the relation between mind and reality is speculative or the mind-independent objects in question are suspended without explanation. Sosa is right to say that: “Knowledge requires not only internal justification or coherence or rationality, but also external warrant or aptness. We must be both in good internal order and in appropriate relation to the external world” (Sosa 2009: 204). Nevertheless, it seems to me that Sosa and his allies who hold the view that there are basic or foundational beliefs in our perceptual knowledge do not make clear their ontological commitment in relation to empirical knowledge. To use the example mentioned above, Sosa says: [W]hile viewing a snowball in sunlight I may have visual experience as if I see something white and round, which may prompt the corresponding perceptual belief. In that case it will be an important part of what makes my perceptual belief epistemically justified—and indeed of what makes it a perceptual belief—that it is caused by such experience. Sosa 2009: 89

Here, the front end of the causal aetiology is experience and the back end is belief. But it doesn’t touch the problem, like, where and how our experience comes from or how we correctly identify the external cause of our experience. That is an ontological problem, not an epistemological one. The perceptual experience by itself does not guarantee its objectivity. The sentence, “I see a cat on the mat,” may be a true description of my personal experience and thus leads me to conclude that, “There is a cat on the mat.” But the former sentence is not the factor which causes me to make the latter belief. The perceptual experience is grounded in my retinal stimulus from the external world. There is no guarantee of sensory certainty in terms of the relation between the internal and the external only based on comprehensive coherence and virtue perspectivism. It is because the channel to the external (or distal) factor (or factors) is still dark and foggy. As described by Rescher, is it an account which offers us, in a way, “Hamlet without the ghost?” (Rescher 2004: 162). In other words, it is a case of “the knowable effect without knowable cause.”

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

159

When I say “I see a white snowball over there,” literally, it means that, through my eyes, I see an object called “white snowball” in the external environment. But it is not an accurate description of my perceptual story. More accurately, it is about a process of perception traveling from the distal factors in the three-dimensional visual world to the two-dimensional surface of light receptors in my eyes whose retinas transduce light stimulus to form signals in my nervous system. Our perceptual experience is formed from the effect of this causal process, the last stop of the journey; but we never know the cause in reality, the starting point of the travel. We just use our experience to make beliefs about the external world. We think that there is something out there which causes our vision, but we don’t directly know the constituents of that cause and in what way each of them plays the causal role in making our visual perception. This is exactly the situation as described by Minsky, that: “Just as physicists cannot see the atoms they talk about, psychologists can’t watch the processes they try to examine. We only ‘know’ such things through their effects [my italics]” (Minsky 1985: 60). Of course, we use many similar causal stories to form a Great Story, i.e. a theory with a series of coherent stories and, based on the Great Story to speculate or conceptualize the cause in reality. This indirect tracking is helpful for fitting our intersubjective world, but there is still no direct contact with reality.5 It seems to me that most epistemologists just speculate a cause in the external world by a transcendental argument from experience in the Great Story to the unknown objects, a necessary elf of their speculation or presupposition. I think it is the reason why Kant claims that, without intuition, objects in themselves are not known to us at all though one can think an object through categories. Here, I think, if Sosa and his allies were to suspend this ontological problem, it would show no significant difference from Husserl’s strategy of phenomenological reduction. However, if they regard objectivity as only about empirical knowledge without origin or source from something transcendent of our empirical knowledge or from something without language and thought, it may be similar to Zhuangzi’s view, which I will interpret and discuss by the end of this chapter. I also think that it may be a way out of the ontological predicament mentioned above.

4 Is the skepticism of “brains in a vat” self-refuting? As we know, one of the serious epistemological challenges from skepticism is Putnam’s thought experiment of “brains in a vat” (hereafter, “brains in a vat” is simplified as “BIV”). But, according to Putnam, it is not a real challenge, because, he argues, to say “we are BIV” is self-refuting. His reason is: [W]hen the brain in a vat (in the world where every sentient being is and always was a brain in a vat) thinks “There is a tree in front of me,” his thought does not refer to actual trees. On some theories that we shall discuss it might refer to trees in the image, or to the electronic impulses that cause tree experiences, or to the features of the program that are responsible for those electronic impulses. These theories are not ruled out by what was just said, for there is a close causal connection

160

Ernest Sosa Encountering Chinese Philosophy

between the use of the word “tree” in vat-English and the presence of trees in the image, the presence of electronic impulses of a certain kind, and the presence of certain features in the machine’s program. On these theories the brain is right, not wrong in thinking “There is a tree in front of me.” Given what “tree” refers to in vat-English and what “in front of ” refers to, assuming one of these theories is correct, then the truth conditions for “There is a tree in front of me” when it occurs in vat-English are simply that a tree in the image be “in front of ” the “me” in question—in the image—or, perhaps, that the kind of electronic impulse that normally produces this experience be coming from the automatic machinery, or, perhaps, that the feature of the machinery that is supposed to produce the “tree in front of one” experience be operating. And these truth-conditions are certainly fulfilled. Putnam 1981: 14

So, he concludes that: “if we are brains in a vat, then the sentence ‘We are brains in a vat’ says something false (if it says anything). In short, if we are brains in a vat, then ‘We are brains in a vat’ is false. So it is (necessarily) false” (Putnam 1981: 15). Putnam just gives a general outline of his argument against the skeptical view that “we are BIV.” But he doesn’t provide a full-fledged argument. Elsewhere I have argued that most of the arguments elaborated from Putnam’s outline are not successful in rejecting the skeptic (Fung 1995: 177–195). Let’s see one of the examples of the elaboration of Putnam’s argument offered by Brueckner: (A) If we are BIV, then our tokens of “We are BIV” are true iff we are BIV in the image. [causal-theoretic semantics] (B) If we are BIV, then we are not BIV in the image. (C) If we are BIV, then our tokens of “We are BIV” are false. [(A), (B)] (D) If we are not BIV, then our tokens of “We are BIV” are true iff we are BIV. (E) If we are not BIV, then our tokens of “We are BIV” are false. [(D)] (F) Either we are BIV or we are not BIV. (G) Our tokens of “We are BIV” are false. [(C), (E), (F)] (H) Our tokens of “We are not BIV” are true. [(G)] Brueckner 1992: 126 This is not a successful argument with the form of dilemma, because (C) is derived from (A) and (B) described in Vat-English and (E) is derived from (D) described in Human-English. For BIV, the consequent of (C) “our tokens of ‘We are BIV’ are false” means that we are not BIV in the image; while for humans or non-BIV, the consequent of (E) means that we are not BIV. So, the conclusion (G) from this seeming dilemmatic argument is ambiguous or equivocal. Another example provided by Dell’Utri is also not successful: (1) We are BIV. [by hypothesis] (2) If we are BIV, then we are not BIV in the image. [explanation of the hypothesis] (3) We are not BIV in the image. [(1), (2)]

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

161

(4) If we are BIV, then every word we utter refers to objects in the image. [1st and 2nd premise] (5) Every word we utter refers to objects in the image. [(1), (4)] (6) Every uttered sentence A is true iff A is true in the image. [(5)] (7) The sentence ‘We are BIV’ is true iff it is true in the image. [instantiation of (6)] (8) The sentence “We are BIV” is true in the image iff we are BIV in the image. [thesis of equivalence] (9) The sentence “We are BIV” is true iff we are BIV in the image. [(7), (8)] (10) The sentence “We are BIV” is false. [(3), (9)] (11) The sentence “We are BIV” is false iff the sentence “We are not BIV” is true. [definition of falsity] (12) The sentence “We are not BIV” is true. [(10), (11)] (13) The sentence “We are not BIV” is true iff we are not BIV. [thesis of equivalence] (14) We are not BIV. [(12), (13)] Dell’Utri 1990: 87 The above argument looks fine. But there are some missing words without which the argument is incomplete. Let’s find the missing words: (1 ) (2 ) (3 ) (4 )

We are BIV [in the real world]. If we are BIV [in the real world], then we are not BIV in the image. We are not BIV in the image. If we are BIV [in the real world], then every word we utter refers to objects in the image. (5 ) Every word we utter refers to objects in the image. (6 ) Every uttered sentence A [in Vat-English] is true iff A is true in the image. (7 ) The sentence “We are BIV” [in Vat-English] is true iff it is true in the image. (8 ) The sentence “We are BIV” is true in the image iff we are BIV in the image. (9 ) The sentence “We are BIV” [in Vat-English] is true iff we are BIV in the image. (10 ) The sentence “We are BIV” [in Vat-English] is false. (11 ) The sentence “We are BIV” [in Vat-English] is false iff the sentence “We are not BIV” [in Vat-English] is true. (12 ) The sentence “We are not BIV” [in Vat-English] is true. (13 ) The sentence “We are not BIV” [in Vat-English] is true iff we are not BIV [in the image]. (14 ) We are not BIV [in the image]. If I am right, the above argument cannot be regarded as self-refuting. Because the correct expression of premise (1) is not merely “We are BIV,” but that of (1 ) “We are BIV [in the real world];” and the correct expression of (14) is not “We are not BIV,” but “We are not BIV [in the image].” So, the conclusion (14 ) is not the negation of the premise (1 ). One of the main points used by some other elaborations of Putnam’s argument against the skeptic’s view is that: there is no reason to use both the principle of disquotation (non-BIV’s version of thesis of equivalence) and the principle of non-

162

Ernest Sosa Encountering Chinese Philosophy

disquotation (BIV’s version of thesis of equivalence) in their arguments. (For example, the consequent of (A) and that of (D) in Brueckner’s argument and (8) and (13) in Dell’Utri’s argument.) However, for the sake of argument, if there is no problem about using both the principle of disquotation and the principle of non-disquotation and if there is also no problem about the ambiguity of “in Vat-English” and “in HumanEnglish” or about the ambiguity of “we are BIV” and “we are BIV in the image,” I think we still cannot make a valid argument to prove that we are not BIV. It is because there is still a problem about the interpretation of any argument against the case of BIV. In other words, any valid argument can be accepted by both non-BIV (humans) and BIV. They share the same type of argument though they have different tokens of argument. My argument is like that in the BIV-story. According to Putnam, a real man cannot make a true statement that “I am a BIV” because he is not a BIV. Nevertheless, for a BIV, it also cannot make a true statement that “I am a BIV” because it is not a BIV in the image. In other words, a real man cannot use Human-English to make the statement that “I am a BIV” true and a BIV also cannot use its Vat-English to make a similar statement true. From God’s eye, for the former, “BIV” means BIV; for the latter, “BIV” means BIV in the image or some specific electrical impulses at the BIV’s nerve ending. Although the two tokens of “I am a BIV” have different meanings for a non-BIV and a BIV, we are still unable to prove that we are not BIV. For example, suppose we have the following self-refuting argument form: (1) We are BIV. .......... .......... .......... (n) We are not BIV.

Suppose further that the argument from (1) “We are BIV” to (n) “We are not BIV” is proved to be valid. If we are really humans, the concluding sentence token is used to refer to the fact that we are really not BIV. But, unfortunately, if we are BIV, the concluding sentence token is used to refer to the fact that we are really not BIV in the image. Since we do not know we are BIV or not BIV from the very beginning, the concluding sentence can be interpreted as referring to the fact that we are not BIV or as referring to the fact that we are not BIV in the image. But we have no evidence to support which interpretation is right for us before we know our real status. So, skepticism is still there! Is there a way out of the skeptic challenge? Sosa finds a way out from Descartes’ best circles. Based on this idea, for example, he reconstructs G. E. Moore’s argument in the following way: Compare now how Moore might have proceeded: 1. 2.

Datum: I know with a high degree of certainty that here is a hand. I can see and feel that here is a hand, and that is the only, or anyhow the best account of the source of my knowledge that here is a hand.

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way 3. 4. 5.

163

So my perception that here is a hand is what explains why or how it is that I know (with certainty) that here is a hand. But my perception could not serve as a source of that degree of justified certainty if it were not a reliable faculty. So, finally, my perception must be a reliable faculty.

Moore could of course go on to say more about the nature of the perception that assures him about the hand. He might still say that such perception involves an implicit inference from what is known immediately and introspectively, perhaps an inductive or analogical inference of some sort. And that might make his view more comprehensively coherent, but we have already seen reasons why postulating such an inference is questionable. So we focus rather on a second alternative: Moore might well take perceiving to involve no inference at all, not even implicit inference, but only transfer of light, nerve impulses, etc., in such a way that the character of one’s surroundings has a distinctive impact on oneself and occasions corresponding and reliable beliefs. This might also amount eventually to a comprehensively coherent view of one’s knowledge of the external world. And its epistemologically significant features would not distinguish it in any fundamental respect from the procedure followed by Descartes. Sosa 2009: 184–185

This is one of the arguments with the spirit of virtue perspectivism (or virtue, externalist or generic reliabilism). Here, if Sosa is right, it seems that he can explain away the BIV type of skepticism. His argument is this: Take the brain in a vat, for example, or the victim of the evil demon. An adult recently envatted, or victimized by the demon, can be indistinguishable from the best of us in respect of the comprehensiveness and coherence of their beliefs and experiences. Even when right about environing objects and events, such a victim’s beliefs are far from being knowledge. Might such comprehensive internal coherence exhaust all cognitive or intellectual competence, at least when it comprises not only beliefs but also experiences? Not if it is part of competence to get it right non-accidentally. For such mental coherence might conceivably be detached from the environing world of the thinker, so as to deprive him of reliable access to the truth. And this would leave him short of a kind of cognitive or intellectual competence, that which consists at least in part in being so constituted and so situated as to enjoy such reliable access. Sosa 2009: 190

Nevertheless, I have two questions: one is about the question whether BIV’s mental coherence is conceivably detached from the environing world at all as claimed by Sosa. Another is about whether Sosa’s externalism is down-to-earth. My answer to the first question is that BIV is not detached from the external world though the way of its connection with the external world is different from ours. We (non-BIV) contact the external world through our perception while BIV receives

164

Ernest Sosa Encountering Chinese Philosophy

images or electronic impulses from the super-scientific computer whose mechanism of input-output is transformed or reflected in some sense from the external world. If the super-scientific computer is not God, it cannot do anything unless it has some connection with the external world. So, BIV is not totally detached from the external world; it does connect with the external world indirectly. In other words, the BIV story presupposes the existence of an external world and, more importantly, its possibility requires a connection between BIV and the external world is some special way. In regard to the second question, my answer is that Sosa’s reliable access to the truth presupposes our connection with the external world. But the connection is not clearly explained in Sosa’s theory because the relatum on the side of the external world has not yet been identified in his theory. Is the relatum the proximal or distal stimulus? If it is the former, it may not be the true reflection of the real causal factor or factors in the external world. If it is the latter, what is the mechanism in Sosa’s theory to identify the real causal factor or factors? In this regard, I don’t think the non-BIV in Sosa’s theory has a slight edge over the BIV.

5 How to ask the skepticism of “brains in a vat” to get lost? Let’s look at a story about Galileo’s “invisible crystalline substance”: Famous in the history of science is the argument ad ignorantiam given in criticism of Galileo, when he showed leading astronomers of his time the mountains and valleys on the moon that could be seen through his telescope. Some scholars of that age, absolutely convinced that the moon was a perfect sphere, as theology and Aristotelian science had long taught, argued against Galileo that, although we see what appear to be mountains and valleys, the moon is in fact a perfect sphere, because all its apparent irregularities are filled in by an invisible crystalline substance. And this hypothesis, which saves the perfection of the heavenly bodies, Galileo could not prove false! Legend has it that Galileo, to expose the argument ad ignorantiam, offered another of the same kind as a caricature. Unable to prove the nonexistence of the transparent crystal supposedly filling the valleys, he put forward the equally probable hypothesis that there were, rearing up from the invisible crystalline envelope on the moon, even greater mountain peaks—but made of crystal and thus invisible! And this hypothesis his critics could not prove false. Copi and Cohen 2009: 93

In this story, the skeptic challenges that we are unable to prove the non-existence of the transparent crystal supposedly filling the mountains and valleys on the moon though Galileo believes that there are mountains and valleys on the moon. To rebut this ad hoc hypothesis, Galileo’s ingenious response is that: the skeptic is unable to prove the nonexistence of the mountains and valleys, which are made of invisible crystalline substance, rearing up from the invisible crystalline envelope on the moon. Following Galileo’s way of thinking, I think, it may be a good strategy to defeat skepticism. Suppose that there is a mini-brain parasitic on a BIV. In this regard, the

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

165

BIV skeptic has no better position than ours. Because, just like us, s/he also suffers from a second-order skepticism, a challenge from the mini-brain. In this story, the objects of reference are different for a real man, a BIV, and a mini-brain, respectively: (A) A real man’s term “tree” refers to a real tree. (B) A BIV’s term “tree” refers to the image of a tree (or some specific electrical impulses at the BIV’s nerve endings). (C) A mini-brain’s term “tree” refers to an inversed image of a BIV’s image. Here, the message of the story is that: all skepticism can be challenged by a secondorder skepticism and, similarly, a second-order skepticism can be challenged by a third-order skepticism, and so on and so forth. So, it would be an infinite regress. Besides, in comparison, the skeptic’s world is poor, narrow, and shallow while the nonskeptic’s world is rich, broad, and deep in the sense that the former, unlike the latter, does not have the following things or events: (1) To plant an apple tree, to eat an apple, to give an apple to a friend, . . . (2) To have an intention to give an apple to a friend, to be angry with a friend who rejects your giving, . . . The events in (1) are mainly operated by an agent’s body. Those in (2) are mainly operated by an agent’s will. In comparison with the non-skeptic’s world, an agent in the skeptic’s world cannot do both things in (1) and (2), because, for example, as a BIV or a mini-brain parasitic on a BIV, the agent in the skeptic’s world does not have agency or active (mental) power to do anything; its seeming agency is nothing but a kind of input image. In addition to lacking agency, the skeptic’s world is also poor in the sense that the BIV or the mini-brain only has bi-angulation with the terminal which makes input-output for the BIV or the mini-brain; it does not have tri-angulation with other entities. I think the fatal weakness of the BIV type skepticism is that its possibility is based on a trivial or swollen assumption, which is not difficult to reject. For the case of BIV, the unnecessary assumption is the super-scientific computer which can be rejected by another assumption of a supra-scientific computer which invades the super-scientific computer with a kind of virus and thus makes the super-scientific computer construct a fictional story of BIV. But, in reality, there is really no BIV and we, as humans, are not BIV. Here, why is the BIV type skepticism refutable? One of the reasons, I think, is that the assumption of BIV-makers, including the super-scientific computer, is not only not necessary, but also without philosophical significance. If it is necessary and significant, the assumption of a ghost or God, a soul or lemures, is also. So, assuming that there is a BIV-maker is similar to assuming that there is a God or ghost; assuming there is a supra-maker of a seeming story of BIV is similar to assuming that there is a supra-God or supra-ghost. All these assumptions are not only challenged by Kant’s antinomies, but also trivial and swollen. I think all the ways of forming skepticism have to suppose that: our beliefs of the external world are not the true report or reflection of the cases or facts in reality. This

166

Ernest Sosa Encountering Chinese Philosophy

supposition entails that there are cases or facts in reality which in principle could be described as other than what we do with our belief-sentences. So, from the eye of the skeptic of this kind, our descriptions are false while their descriptions, if any, are true. Since to assert that there are cases or facts in reality which in principle could be described as other than what we do with our belief-sentences entails that there is a conceptual scheme or language which could do the job which our conceptual scheme or language fails to do, the skeptic of this kind has to assert that reality is describable, in principle, in their conceptual scheme or language, but, for us, it is a mission impossible. Here, my argument is that, unless the skeptic asserts that, in reporting or reflecting reality, her/his conceptual scheme or language is different from (or even incomparable to) ours, s/he cannot claim that her/his reports are right and ours are wrong in knowing reality. But, I think, there is a similar problem for the skeptic. That is: how can they use their conceptual scheme or language to describe reality rightly? What is the relation between their language and reality? Is it correspondence, representation, stimulation, or causation in any sense? I think both we and the skeptic are in the same boat in facing this problem. So, if the skeptic cannot justify their possibility, they have no ground to challenge our understanding of the external world. In this regard, I think the only option for skepticism is that: it can survive in God’s mind. In general, I think we cannot ask skepticism to get lost if God’s eye or mind is different from ours. So, we cannot kick out the skepticism of God’s kind. But, I think we can ask the skepticism of the kind speculated from our skeptical anxiety. The main reason is that: If our descriptions of the external world are coherent and also workable in the sense that these descriptions or beliefs make contribution to our survival as a necessity for the well-being of our life, even though, by speculation from our anxiety, skepticism is possible, it is a burden for the skeptic to prove their descriptions or beliefs could also make the same or more contribution. But I don’t think any language-users, including the skeptic, could do the job better than we have done. Previously I have made some arguments to weaken skepticism, but, honestly speaking, none is a knocking-down argument. Now, I think Zhuangzi’s insight may help us to make a good argument to do the job. His lesson given to us is that there is no entity in reality which can play a role of the base or ground to make any epistemological relation, including correspondence, representation, stimulation, or causation, possible. Because there is nothing in reality (or reality is nothing), to say there is something in reality, we have to use concepts or linguistic entities to identify what it is or what they are. To borrow Quine’s slogan “no entity without identity,” we cannot have ontological commitment to any entity in reality without the mechanism of identification. Similarly, for Zhuangzi, before using language and concepts to describe or recognize something in reality, there is nothing. It is a logical thesis which says that there is only one world in which we live and of which we are also a part. There is no other world or other level of world which is different from the world we know or recognize. Before we use language and concepts to describe or recognize something in this world, it is nothing to be identified or discerned. When we use language and concepts to identify something in the world, there will be something transformed from nothing. In other words, with

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

167

a Husserlian style, the room of reality is open to us to come in when we turn on the light. Metaphorically speaking, there is only one room which has two states: one can be seen with light and the other is in darkness.

6 Zhuangzi’s proposal Zhuangzi does not accept a view of two worlds or two levels of world. Instead, he accepts a view of two states of one world. But, his two-state view is not the kind like that of the waking state and the dreaming state in our life. He thinks that, before we use language and concepts to express the external world or even when no people use their linguistic and conceptual tools to express the external world, it is still there. But, in this moment or situation, metaphorically speaking, just like a raw material, the world as it is has not yet been “cut” or “processed.” So, it has no form or characteristic. For Zhuangzi, without language and concepts, we cannot know it or treat it as something which is similar to or different from others. In other words, in this natural state, the world is nothing or it is in a chaotic state. So he uses a fable of hundun ⑮⊼ (chaos) to express this point:6 The Ruler of the Southern Ocean was Shu ‫ݥ‬, the Ruler of the Northern Ocean was Hu ᘭ, and the Ruler of the Centre was Chaos. Shu and Hu were continually meeting in the land of Chaos, who treated them very well. They consulted together how they might repay his kindness, and said, “Men all have seven orifices for the purpose of seeing, hearing, eating, and breathing, while this (poor) Ruler alone has not one. Let us try and make them for him.” Accordingly they dug one orifice in him every day; and at the end of seven days Chaos died. Zhuangzi 7: 7 Legge 1891 ই⎧ѻᑍ⛪‫⎧ेˈݥ‬ѻᑍ⛪ᘭˈѝཞѻᑍ⛪⑮⊼DŽ‫ݥ‬㠷ᘭᱲ⴨㠷䙷ᯬ⑮⊼ ѻൠˈ⑮⊼ᖵѻ⭊ழDŽ‫ݥ‬㠷ᘭ䄰๡⑮⊼ѻᗧˈᴠ˖NjӪⲶᴹгヵˈԕ㿆㚭 伏᚟ˈ↔⦘❑ᴹˈే䂖䪯ѻDŽnjᰕ䪯аヵˈгᰕ㘼⑮⊼↫DŽ㦺ᆀ៹ᑍ⦻)

I think Zhuangzi’s view of the natural or original state of the world or reality (which is called “hundun” ⑮⊼) can be regarded as a way out of Davidson’s and other philosophers’ ontological predicament. Zhuangzi does not treat reality as a correspondence base for objective truth; nor does he treat it as a kind of “thing-initself ” which causes our perception. It seems to me that the epistemological role of reality is suspended in a sense similar to phenomenological epoche. But it is not a phenomenological reduction, because there is nothing to suspend and what is suspended is the epistemological role of reality (not reality itself). Zhuangzi just yearns for entertaining a spiritual vision of harmonic state through his idea of aesthetic (not ontological) mysticism.7 Sometimes, Zhuangzi borrows the voice of a wise man in ancient times [i.e. “ancient man” (gu zhi ren ਔѻӪ), “true man” (zhenren ⵏӪ), “perfect man” (zhiren 㠣Ӫ), “sage” (shengren 㚆Ӫ), or “spirit-like man” (shenren ⾎Ӫ)] to express his view:

168

Ernest Sosa Encountering Chinese Philosophy

Among the men of old their knowledge reached the extreme point. What was that extreme point? Some held that at first there was not anything. This is the extreme point, the utmost point to which nothing can be added. A second class held that there was something, but without any responsive recognition [conceptualization] of it (on the part of men). A third class held that there was such recognition, but there had not begun to be any expression of different opinions about it. It was through the definite expression of different opinions about it that there ensued injury to Dao. It was this injury to Dao which led to the formation of (partial) preferences. Zhuangzi 2: 7 Legge 1891 ਔѻӪˈަ⸕ᴹᡰ㠣⸓DŽᜑѾ㠣˛ᴹԕ⛪ᵚ࿻ᴹ⢙㘵ˈ㠣⸓ⴑ⸓ˈнਟԕ ࣐⸓DŽަ⅑ԕ⛪ᴹ⢙⸓ˈ㘼ᵚ࿻ᴹሱҏDŽަ⅑ԕ⛪ᴹሱ✹ˈ㘼ᵚ࿻ᴹᱟ䶎 ҏDŽᱟ䶎ѻᖠҏˈ䚃ѻᡰԕ㲗ҏDŽ䚃ѻᡰԕ㲗ˈᝋѻᡰԕᡀDŽ㦺ᆀ啺⢙䄆)

For Zhuangzi’s ancient perfect man, the natural or original state of the world is that there has not yet been anything (wei shi you wu ᵚ࿻ᴹ⢙). Later, people cannot reach this perfect vision and think that there is a reality as an entity which has not yet been bound by linguistic and conceptual tools (wei shi you feng ᵚ࿻ᴹሱ). People think that the reality is bound by linguistic and conceptual tools and there has not yet been conflict in truth and falsity (wei shi you shifei ᵚ࿻ᴹᱟ䶎). When there is conflict in truth and falsity, dao as the original state of this world (i.e. the ultimate or original reality) would be lost (just like the destiny of Chaos) and prejudice as personal love would be formed. The above view can be summarized in a pair of sentences: “A path is formed by walking. A thing is to be the case by saying” (Zhuangzi 2: 6) (䚃㹼ѻ㘼ᡀˈ⢙䄲ѻ㘼 ❦). This means that, in reality (or in the natural or original state of the world), there is no path or, more correctly speaking, there is no such a thing as a path before people’s walking.8 Similarly, there is also no such an object or event before people’s saying. In other words, all things are reality’s being conceptually or linguistically “processed” or “polluted.” If the original state of nature were to be carved by a linguistic and conceptual scheme, we would enter into a world of sameness and difference, truth and falsity, love and hate, a world in which “a path is to be formed by walking; a thing is to be the case by saying.” In this situation, dao, or the undifferentiated vision of “wei-yi” ⛪а (being one), would be lost. Of course, Zhuangzi does not reject the existence of reality; but it is not a reality recognized as a base of correspondence or representation, or as a starting point of stimulation or causation. What he means by “reality” is an undifferentiated oneness, an unknown what it is, an unsayable dao, or a kind of huntun, which has not yet been differentiated by conceptualization into individual things. Let’s look at his idea of “wei-yi” (being one): Heaven, Earth, and I were produced together, and all things and I are one. Since they are one, can there be speech about them? But since they are spoken of as one, must there not be room for speech? One and Speech are two; two and one are

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

169

three. Going on from this (in our enumeration), the most skillful reckoner cannot reach (the end of the necessary numbers), and how much less can ordinary people do so! Therefore from non-existence we proceed to existence till we arrive at three; proceeding from existence to existence, to how many should we reach? Let us abjure such procedure, and simply rest here. Zhuangzi 2: 9 Legge 1891 ཙൠ㠷ᡁі⭏ˈ㘼㩜⢙㠷ᡁ⛪аDŽᰒᐢ⛪а⸓ˈфᗇᴹ䀰Ѿ˛ᰒᐢ䄲ѻа ⸓ˈфᗇ❑䀰Ѿ˛а㠷䀰⛪ҼˈҼ㠷а⛪йDŽ㠚↔ԕᖰˈᐗᳶн㜭ᗇˈ㘼 ⋱ަࠑѾʽ᭵㠚❑䚙ᴹˈԕ㠣ᯬйˈ㘼⋱㠚ᴹ䚙ᴹѾʽ❑䚙✹ˈഐᱟ ᐢDŽ㦺ᆀ啺⢙䄆)

Zhuangzi uses “dao” 䚃, “nothing” (wu ❑), “chaos” (huntun ⑮⊼), and “natural” (ziran 㠚❦) to express an undifferentiated oneness (hunyi ⑮а) or being one (wei-yi ⛪а). Zhuangzi claims that the heaven, earth, and I can be seen as being produced together and all things and I as being one if we can transcend the relative attitude and conceptually carving mentality. In other words, there is no temporal priority between heaven, earth, and me and there is also no spatial difference between all the things and me, including no individuality of objects and no distinctions among them. When one goes beyond the relativity of language and thinking, one can entertain the spiritual vision of undifferentiated and harmonic horizon. Why Zhuangzi’s “wei-yi” ⛪а (being one: all as one, or one as an undifferentiated entity) or “tai-yi” ཚа (great one) is not “yi-ti” а億 (one body: all in one, or one as a whole) as claimed by Hui Shi ᜐᯭ, a friend of Zhuangzi and a sophist?9 One of the reasons is as follows: Therefore his liking was one and his not liking was one. His treating one was one and his not treating one was one. In being one, he was acting as a companion of Heaven. In not being one, he was acting as a companion of man. When man and Heaven do not defeat each other, then we may be said to have the True Man. Zhuangzi 6: 1 Legge 1891 with minor modification ᭵ަྭѻҏаˈަᕇྭѻҏаDŽަаҏаˈަнаҏаDŽަаˈ㠷ཙ⛪ ᗂ˗ަнаˈ㠷Ӫ⛪ᗂDŽཙ㠷Ӫн⴨ऍҏˈᱟѻ䄲ⵏӪDŽ㦺ᆀབྷᇇᑛ)

In other words, there is no distinction between sameness and difference in the undifferentiated oneness though, in Hui Shi’s “one body” or “one whole,” the distinction between sameness and difference is a necessity to identify a whole and its members. Furthermore, Zhuangzi thinks that the ideal (the natural) is just inside the actual (the artificial). That is: “If you could hide the world in the world, so that there was nowhere to which it could be removed, this would be the grand reality of the everduring thing” (Zhuangzi 6: 2) (㤕ཛ㯿ཙлᯬཙл㘼нᗇᡰ䚟ˈᱟᙶ⢙ѻབྷᛵҏ). Just like the flour is inside the bread, we cannot separate them into two realms of reality. They are nothing but two states of one thing. To go beyond this man-made side to see this stuff, Zhuangzi believes that, through self-cultivation, such as sitting for

170

Ernest Sosa Encountering Chinese Philosophy

forgetting all things (zuo-wang ඀ᘈ) and fasting or cleaning of the mind (xin-zhai ᗳ 啻), our naturalistic view can be transformed into a spiritual vision without being restricted in the artificial side and suffering from the conceptual “pollution.” For Zhuangzi, living in a conceptually carved world is not an ideal life; but we cannot escape from this cage. If we want to grasp the ideal life, “we should not exclude what is right and wrong (in thinking) and thus we might live in peace with the prevalent views” (Zhuangzi 33: 6) (н䆤ᱟ䶎ԕ㠷ц؇㲅). It means that if one excludes or rejects any view of truth and falsity, s/he would also commit a view of truth and falsity. In other words, Zhuangzi does not recognize all the views as true or false in terms of correspondence or representation; he just transcends all these disputations and thus enters into a realm of undifferentiated harmony, just like a happy fish (though he does not reject the ordinary use of truth and falsity in daily life). But how to live in peace with the prevalent views that there are truth and falsity in our conceptual world? Zhuangzi’s answer, I think, is that beliefs accepted by people are mainly based on their function or usefulness in practical life. This is Zhuangzi’s idea of “gong” ࣏ (function) or “yong” ᓨ (usefulness). So he says: From the point of view of function, if we regard a thing as useful because there is a certain usefulness to it, then among all the ten thousand things there are none that are not useful. If we regard a thing as useless because there is a certain uselessness to it, then among the ten thousand things there are none that are not useless. If we know that east and west are mutually opposed but that one cannot do without the other, then we can estimate the degree of function. Zhuangzi 17: 5 Legge 1891 ԕ࣏㿰ѻˈഐަᡰᴹ㘼ᴹѻˈࡷ㩜⢙㧛нᴹ˗ഐަᡰ❑㘼❑ѻˈࡷ㩜⢙㧛 н❑DŽ⸕ᶡ㾯ѻ⴨৽ˈ㘼нਟԕ⴨❑ˈࡷ࣏࠶ᇊ⸓DŽ㦺ᆀ⿻≤)

In regard to different views, his attitude is that: “This being so, let us give up our devotion to our own views, and occupy ourselves with the views’ function” (2: 6) (⛪ᱟ н⭘㘼ሃ䄨ᓨ). It seems that, for Zhuangzi, there is no objective truth in an absolute sense. If there is any truth, there is only objective truth in a relative sense, i.e. relative to the function of a belief which is with collective recognition or agreement.

7 Concluding remarks I think the skepticism of the external world cannot survive in Zhuangzi’s world-view, because the so-called “external world” or “reality” is nothing as demonstrated by Zhuangzi. We often think that the external world or reality includes what is outside our skin and what is inside our skin. What is inside is a sensing, thinking, and saying machine which can sense something, think of something, and say about something of what is outside and inside. Whether reality is the same as what is sensed, thought of, and talked about is an ontological problem which has puzzled many philosophers for centuries. If it means that the relation of sameness between reality and what is

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

171

experienced, I think it is a view of insolvable mystery. Nevertheless, if we follow Zhuangzi to treat reality and what is experienced as the same world of different states: the natural state without or before epistemic invasion and the artificial state which is full of epistemic coloring, I think there is no skepticism that can survive and the objectivity of knowledge can probably be living in the world of the artificial state only. In summary, I think Zhuangzi’s view of two states of one world is a logical thesis:10 In regard to his idea of reality or his view of the natural state of the world, we can ask him two questions. One is “What reality is not?” The other is “How to get out of skepticism or the ontological predicament?” According to the text, his answer to the first question can be interpreted with the following points: (1) Reality (i.e. the original, natural, or ultimate state of this world) is nothing (not a thing or things), because things or entities are individuated and distinguished by our thinking machine. (2) Reality is also ineffable or unsayable, because it is the state before any procedure of sensing and has not yet been conceptualized in our thinking and saying. (3) Moreover, it is also not a base of correspondence or representation, not a starting point of stimulation or causation, because there is only one world which has two different states: one is the state without any epistemic access; the other is transformed into what is known to us through our sensing, thinking, and saying. Zhuangzi’s answer to the second question is: 1. Zhuangzi’s reality is not a base of correspondence or representation, a starting point of stimulation or causation, so there is no conceptual relativism and global skepticism. 2. There is no external cause which is outside our experience and conceptual scheme; no causal efficacy of objects and events (such as Kant’s “thing-in-itself ” or Santayana’s “efficacious reality,” i.e. “God”) can survive outside this conceptual world. 3. There is only one world of two states: the natural or original state which does not participate in epistemology and the actual or artificial state which is knowable in our experience. 4. Zhuangzi’s aesthetic mysticism is not an ontological mysticism. His dao is ineffable because it is nothing. So, there is no contradiction. But the ineffability of the ultimate in ontological mysticism is self-refuting.

Notes 1

Mark Johnston of Princeton University is currently working on a book entitled The Manifest, which explains how the world of lived experience can be very much as it appears to be, despite the discoveries of the physical and biological sciences. In chapter 7 of the book, he makes a distinction between “appearing” and “appearance.” He thinks that: “The causing [of the perceptual process] does not come before the appearing, but rather constitutes it. The appearing is not to be broken down into an

172

Ernest Sosa Encountering Chinese Philosophy

appearance on the end of a causal chain. The causal argument does not get a foothold” (see the draft chapter on the website: http://www.nyu.edu/gsas/dept/philo/courses/ consciousness97/papers/johnston/chap7.html). 2 Similar words from Davidson are: “If nothing is systematically causing the experiences, there is no content to be mistaken about. To quote myself: ‘What stands in the way of global skepticism of the senses is, in my view, the fact that we must, in the plainest and methodologically most basic cases, take the objects of a belief to be the causes of that belief ’ ” (Davidson 2001: 201). Here, for Davidson, the correctness is not about the conclusion “how the world around us is structured and populated,” but about our pictures of the world, by virtue of triangulation, are not mistaken. 3 One of the popular examples is the phenomena of metamerism. In colorimetry, metamerism is a perceived matching of colors with different (nonmatching) spectral power distributions. See details in entry on “metamerism” in Luo (2016): 928–932. 4 For example, “God creates things in the world” is senseless if “creation” is defined (partly) in terms of “causation.” The question is: How can a non-physical entity cause a physical entity? 5 The usual causal stories in our empirical world are not of this kind. For example, when we saw the burning of the house and later we also saw the roasting of the pig, we usually thought with a causal statement that “the burning of the house caused the roasting of the pig.” Here, both the cause and the effect are observable and can be known to us. But, our perception of the burning of the house is an effect caused by an event in the external world which, as a cause of the perceptual experience, is not observable or cannot be known in a Kantian sense. So, the first kind of causality is only found in our empirical world while the second kind is branching out from the empirical to the non-empirical. 6 Hereafter, all the quotations from the Zhuangzi in (Legge 1891) are from the Chinese Text Project, http://ctext.org/zhuangzi. 7 Details of the arguments for Zhuangzi’s idea of aesthetic (but not ontological) mysticism can be found in Fung (forthcoming). 8 In our experience of hiking or maintain-climbing, we know that sometimes there is no path/road to climb and some of the paths/roads are formed by consistent walking or climbing in the same area. Similarly, things (wu ⢙) are formed by linguistic or conceptual carving. After the linguistic activity and epistemic attempt, according to Zhuangzi, dao or the undifferentiated vision of “being one” (wei yi ⛪а) would be lost. 9 I have explained the difference between wei-yi and yi-ti and argued against A. C. Graham’s view, which identifies these two as the same concept (Fung forthcoming). 10 Why is this a logical thesis? It is because reality as the natural or original state of the world is by definition or is logically equivalent to the world before conceptualization or linguistic carving. Metaphorically speaking, it is what Zhuangzi describes as “huntun” ⑮⊼ (chaos) “wu” ❑ (nothingness) or “wei-yi” ⛪а (being one) without distinctions.

References Brueckner, Antony. 1992. “If I Am a Brain in a Vat, Then I Am Not a Brain in a Vat.” Mind, New Series 101(401): 123–128. Copi, Irving M. and Cohen, Carl. 2009. Introduction to Logic. New York: Macmillan.

Sosa’s Reliabilist Reading of Davidson and Zhuangzi’s Way

173

Davidson, Donald. 2001. Subjective, Intersubjective, Objective. Oxford: Clarendon Press. Davidson, Donald. 2004. Problems of Rationality. Oxford: Clarendon Press. Dell’Utri, Massimo. 1990. “Choosing Conceptions of Realism: The Case of the Brains in a Vat.” Mind, New Series, 99(393): 79–90. Fung, Yiu-ming. 1995. “The Argument against ‘We Are Brains in a Vat’ and the Problem of Skepticism” (in Chinese). Collected Essays of the Fourth Seminar on American Literature and Philosophy. Taipei: Institute of European and American Studies, Academia Sinica: 177–195. Fung, Yiu-ming (forthcoming). “Zhuangzi’s Idea of Wei-yi (Being One): With Special Reference to the Story of Happy Fish.” In Yiu-ming Fung, Language, Truth and Logic in Ancient China. (expected to publish in 2022). Legge, James. 1891. The Writings of Chuang Tzu, Oxford: Oxford University Press. Refer to Chinese Text Project, http://ctext.org/zhuangzi, edited by Donald Sturgeon. Johnston, Mark (forthcoming). The Manifest. Draft chapter 7 from the website: http:// www.nyu.edu/gsas/dept/philo/courses/consciousness97/papers/johnston/chap7.html Luo, Ming Ronnier. ed. 2016. Encyclopedia of Color Science and Technology. New York: Springer. Minsky, B. M. 1985. The Society of Mind. New York: Simon & Schuster. Putnam, Hilary. 1981. Reason, Truth and History. Cambridge: Cambridge University Press. Rescher, Nicholas. 2004. “Sosa and Epistemic Justification.” In John Greco (ed.), Ernest Sosa and His Critic. Oxford and MA: Blackwell Publishing, pp. 145–164. Sosa, Ernest. 2009. Reflective Knowledge: Apt Belief and Reflective Knowledge. Oxford: Oxford University Press. Stroud, Barry. 1984. The Significance of Philosophical Scepticism. Oxford: Oxford University Press.

174

11

Resurrecting Daoist Virtuosity Epistemology Chad Hansen

1 Introduction This chapter is an attempt to align Classical Chinese philosophy alongside modern naturalistic philosophy, then orient it for a productive encounter with Ernest Sosa’s virtue epistemology. Sosa’s theory is a wide-ranging rearrangement of the concepts of the rationalist tradition of Western epistemology from Plato to Descartes, Kant and Gettier. My hypothesis is that a wide-angle view of Chinese philosophy will motivate reflection on more parallel insights than appear on the surface. The surface appearance, particularly since I will be concentrating on the classical period, will seem radically detached from the long history of concerns of the Western tradition. We certainly should not expect easily recognizable treatment of issues arising in the form of early modern Cartesian evil demon reflections or from twentieth-century Gettier puzzles, let alone their meta-versions. In this sense, it may come to seem that an encounter with Sosa’s philosophy, deeply concerned with improving on traditional and rival modern ways of dealing with the fine details of the Western tradition, is likely to be unproductive. What I will try to do instead is explain the core differences in the distant, Classical Chinese conceptual schemes in which their discussion of knowing emerges. Alignment seems more optimistic when we look at the title theme of Sosa’s approach: Virtue Epistemology. Virtue ethics has attempted to naturalize normativity by resurrecting a Classical Greek, Aristotelian theme. This has drawn interest and enthusiasm among many interpreters of Chinese ethics. It allows us to think of how ethics is possible without the supernatural moral law-giver, the God of Middle-Eastern and Western ethics. Classical Chinese are naturalists so could share the motivation to understand knowing without supernatural, imaginary, or abstract items (the god-like soul-mind). Using Aristotle’s model in epistemology is a by-product of the modern movement of naturalizing epistemology. That movement was arguably triggered by Quine’s reflection that much of the work of traditional epistemology (e.g. “How do we know?”) should be handed over to the empirical science of the central nervous system, neuropsychology, and cognitive science. The balance of what was traditionally viewed as a priori reflection on the nature of the soul-mind’s knowing was reformulated as broadly 175

176

Ernest Sosa Encountering Chinese Philosophy

normative. Epistemologists began to read and borrow strategies from ethics. Virtue epistemology, as a strategy to naturalize normativity, is one way of using ethical lines of reasoning to approach this modern twist on the traditional Western epistemic project. My task here is broadly comparable to reaching back to Aristotle for conceptual materials, but instead it reaches back and around to recruit conceptual strategies that Classical Chinese Masters used. They viewed all language and concepts as normative and knowing as being able to navigate along the paths of guidance information: 䚃 dàopaths. Being a good person entailed being a good knower. And, they were rather more inclined than the Greeks to accept the biconditional—being a good knower was being a good person.1 The envisioned alignment is apt enough in temporal terms—Greek and Chinese Classical periods almost exactly overlap. Both traditions trace their orientations back to these classical periods which are comparably influential even in their modern forms.2 Aligning both does involve extensive, imaginative adaptation or reinterpretation in our modern scientific conception of the natural world (relativity and quantum mechanics) and our understanding of how living forms change (evolution) and how we guide our behavior (the central nervous system). Current versions of virtue epistemology, especially Sosa’s, have put a lot of thought into bringing the Greek rationalist insights up to date. Updating should be interestingly different for the two Classical models, but, as I suggested, the “title” pivots are in place. The Chinese classical philosophical tradition has what appear to be comparable and closely related concepts of ⸕zhī know and ᗧdévirtuosity.3 Beyond those two, however, the Chinese conceptual territory is populated with a quite different collection of related philosophical concepts—concepts that structure their theories of nature (the world), mind (their “heart-mind”), and language. So, the two pivotal terms in their respective languages will differ not only in phonetics, orthography, and syntax, but importantly in the conceptual roles each plays in their respective, vastly different, conceptual neighborhoods. If we can align these conceptual schemes enough, however, I then hope to use the Chinese model to address and address briefly some other ways Chinese reflections converge with Sosa’s virtue epistemology proposal—e.g. the nature of knowledge and how to relate natural human knowing to natural animal knowing.

2 The nature of knowledge An avenue of progress toward this productive encounter was opened by Prof. Sosa’s recent presentation of his Accurate, Adroit, and Apt (AAA) triadic structure of epistemic virtue. In his first lecture at Suzhou University he treats that analysis as integral to studying the nature of knowledge. This question as to the nature of knowledge has been central to epistemology in recent decades, as it was for Plato. Sosa 2011: 2; italics mine

Resurrecting Daoist Virtuosity Epistemology

177

The account of epistemic normativity as a sort of performance normativity helps explain the nature of knowledge, which amounts to belief that is apt, belief that is an apt epistemic performance, one that manifests the relevant competence of the believer in attaining the truth. Sosa 2011: 4; italics mine

However, the alignment will stumble over the italicized continuation. The Classical Chinese concept was different from Plato’s, and, as I will explain below, the Western epistemological tradition and modern conceptual analysis focuses less on the natural concept of knowing than on a socially constructed, Western conception: knowingthat, which is importantly different from the concept used in Classical Chinese normative theorizing. What Chinese reflections on knowing involved was a kind of knowing-how, a modern insight into varieties of knowing popularized by Gilbert Ryle (Ryle 1949).4 Ryle’s concept of knowing aligns much more easily with the non-Platonic, Classical Chinese world and, accordingly will ground our query into the nature of ⸕zhīknowing. Importantly, it still invites the virtue ethicist’s naturalistic focus, but it is considerably easier to see the Classical Chinese focus on knowing-how as bringing us closer to the core of what knowledge is, its nature. It makes it easier to see knowing as a natural kind that plays a role in nature here, in ཙлtiānxiàsky:nature-below (the world of living things). What opens the door to comparison is Sosa’s 2011 use of the AAA model of virtue epistemology to address performance normativity, the exercise of our capacity to know. That provides a more common ground; knowing how to and knowing-to clearly addresses one’s performance. However, Professor Sosa’s formulations still tie us into the intellectual, Greek Rationalist conceptual neighborhood of knowing as believing+. We don’t typically speak of believing-how or believing-to. Believing plays its role only in the Western social construct: knowing-that. While I can try to imagine ways we might treat believing that P as a performance, I would find it more natural to treat performances as utterances (including sub-vocalized, imaginative rehearsing of potential utterances), as drawing inferences, assigning terms to things etc. What I propose to do is extrapolate what I take to be the Classical Chinese version of its rough counterpart, a conception of a belief-like performance that is a special, distinctly Chinese kind of knowing-how—one that is pivotal in the dialectic between Confucianism and Daoism. It is the Daoist concept of ⛪wéideem:do (to guide behavior with a social construct).5 And that partly will help explain why some extreme versions of Daoist skepticism (❑⛪wúwéi lack deeming-as) take the form of advising us to abandon knowledge, rather than simply doubting it.6 That conception of a belief-performance will ground a different account of epistemic virtue, one more resonant with so-called alternative (Turri, Alfano, and Greco 2018),7 rather than conventionalist (Platonic, belief-focused) accounts of epistemic virtue. The true belief component of the conventional Platonic analysis will be a counterpart of knowing-(how)-to express what one knows in a declarative sentence of a language— knowing to deem a thing as X in guiding one’s behavior toward it. That allows us to skirt any a priori commitment to treating the object of knowledge as a sentential object. It makes the notion of an epistemic performance more natural.

178

Ernest Sosa Encountering Chinese Philosophy

Conventional virtue epistemology does not deal with the nature of knowledge per se but with a peculiar, grammatical construction used among the inheritors of the Platonic tradition: knowing-that. To understand knowing itself, I would advise we look more broadly at constructions in Classical Chinese and their closely related structural counterparts in English. That way of opening out the discussion resonates with other proposals for naturalizing our understanding of knowing. One of these is Edward Craig’s suggestion in Knowledge and The State of Nature that we can understand the natural normative role of knowing using strategies from political philosophy (Craig 1999). I have written extensively about Hobbes-Rawls-like State of Nature reflections which were present at the birth of Chinese naturalistic philosophy. These “just-so” historical archeologies are not attempts to salvage the Middle-Eastern notion of a political authority to pronounce laws (general imperative sentences) and enforce them through morally legitimate, retributive, punishment of actions that violate these laws. Mozi’s original form was offered instead as a naturalistic way to motivate reforming traditional 㗙 yìmorality, transmitted and inherited social norms. It didn’t target supernatural political authority but the Confucian historicist social 䚃 dàopath consisting of inherited ⿞ lˇı rituals/ceremonies. Both sides of this ancient Debate, Confucians and Mohists, understood the norms of language use to be integral to this social dào. Mozi, using his state-of-nature reflections, tried to motivate reforming the entire structure including our ways of using the terms in our shared language for guiding and coordinating social behavior. Edward Craig concluded that the point of the concept of knowing is to identify reliable informants, but his was still a traditional focus on knowledge as a species of propositional belief, de dicto belief, and its truth value. Rather than seeking the nature of that parochial philosophical construction using the word “know,” I will urge that we look at all the ways of using the concept that is nearly universally shared in all human languages.8 The concepts of belief and truth are far less widely shared. Of special interest in this encounter with Chinese philosophy is that they are not found in Classical Chinese. We find a straightforward Chinese concept of ⸕zhīknow but not of belief or truth. Chinese normative epistemology resonates with that of another rival virtue ethics theorist, Linda Zagzebski. Like her, Classical China treated epistemology as part of “the normative” in general, and, like Zagzebski and Craig, leads in the direction of an exemplarist moral theory. The naturalist normative stance shared across schools in the Classical Chinese tradition is a version of metaethical naturalism that encompasses language use and knowing. Its distinctive twist on these Alternatives is that it is more intuitively naturalistic. I argue that is because its focus is on the notion of 䚃 dàopaths rather than on normative propositions (commands, propositions containing “should” or “ought”). The shared lexical notion of a “law” in the ancient Western, Middle Eastern, and South Asian normative political theory invites elaboration using supernatural theology. Finding normative guidance in information structures in nature, in path-like structures rather than command-like sentences, directs some attention to prior animal and human behavior (forming the path by prior walking [Zhuang: 4/2/33]) but relies

Resurrecting Daoist Virtuosity Epistemology

179

ultimately on the natural “lay of the land” (Laozi: ch. 8). It thus provides a more natural explanation of the human cognitive access to normative guidance that make human knowing continuous with animal knowing. This pivot of my alternative account of the Chinese focus, 䚃 dàopath, turns out to be another of the terms found in nearly all human languages. A state of nature reconstruction of this conception of guidance is readily at hand. Bipedal creatures emerging from the trees to seek life on the savanna could hardly avoid having a concept of paths, path finding, path choosing, and path following. By contrast, not all languages contain concepts of a law or a conception of norms as commands from some de jure authority, someone who by their verbal performance creates obligations and justifies punishment. The Platonic core of Western monotheistic supernatural theology, by contrast, includes this concept of law and links it to its distinctive focus, the faculty of reason. Around that core, we locate the other elements of the parochial Platonic tradition’s focus on belief, truth, fact, and intention and action. I hope by laying out the contrasts in the conceptual scheme surrounding the Chinese focus on dàos rather than sentences, it can trigger a new perspective on the intramural dispute of conventionalist vs. alternative virtue epistemologists and those between intellectualist (Platonic idealist) and pragmatic, naturalist accounts of the nature of knowing.

3 The contrasts The radical separation of the two conceptual neighborhoods is a product of the interaction of geography and history of the two “worlds.” The ancient, so-called, “West” encompassed the region from the Himalayas to Northern Africa. The mountain chain and deserts separating these ancient centers of Western (of the Himalayas) civilizations sufficiently explain why the East Asian dawn of civilization would not have been a player in the “Old World” drama. That drama played out in migrations, enslavements, dislocation, and mixing of populations throughout the region, which provided a (perhaps unwelcome) context for blending their religions and concepts into a broadly shared, proto-religious ancient Western background worldview. It is that broad mixture that we are conventionally invited to imaginatively use as the “beginning of human thought.” It includes the familiar creation stories, flood myths, God-kings, their laws, commands, and punishments. They shared other parts of the basic conceptual apparatus for talking about human minds and languages—psychology and meaning.9 These notably also include phonetic alphabets and complex inflected grammars, a contrast of thought and feeling, belief and desire, reason and emotion. That ancient history of sometimes coercive, sometimes trade-inspired, cultural mixing could explain how these core components of a proto-religious worldview might have come to be shared across these Sumerian, Persian, Indian, Egyptian, and the Hellenic worlds by the dawn of history. We needn’t assume that they are the natural basic starting points of human thought. This self-recognition that “our world history” might not be a neutral, scientific construction of natural starting points of human conceptual thought puts the

180

Ernest Sosa Encountering Chinese Philosophy

marginalization of Chinese philosophy in a different light. And it invites us not to place undue emphasis on Greek rationalist metaphysics and idealism, the contrast of appearance and reality as the natural place to start talking about knowledge. If modern animal science, anthropology, and neuroscience all pose problems for this implicitly idealist or supernaturalist Western starting point, we should welcome rather than resist the challenge. History, particularly political history, contributes to the dominance of the structuring metaphor implicit in the conceptual network of Indo-European/Middle Eastern normative political theory—that of the authoritative command, of law. This metaphor structures how Western culture has traditionally thought, and still thinks, about normativity, ethics, behavior guidance, and morality.10 It is the idea of a de jure normative authority (a rational being) whose desires, expressed as general commands, orders, rules, performatively creates obligations—the paradigmatic norm. What gives that human such authority? It must be the command of some “higher” (more rational?) authority. We see the natural effect of this regress implicit in the law metaphor in its earliest historical formulation: the law code of Hammurabi. We learn to treat normative discussion as centered on declarative versions of such imperatives, sentences containing a “should” or an “ought.” An act of promulgation by some authority (perhaps of reason itself) renders certain human actions required or prohibited. The relation between actions and the law is a broadly logical-syntactic one (general to particular) rooted in the conception of the sentence and its functional (inflection marked) components. (Nietzsche 1899, “Reason in Language”). The concept of law-giver becomes part of our concept of a king and a kingdom and we learn to treat it as a natural evolutionary development from the smaller prehistorical tribal scale social-political structures (see Hart 1961 as a paradigm). It’s routine to explain written history of political civilization as being a by-product of achieving such a large-scale organization—the kingdom is the natural successor of the tribe. The king is a sovereign defined by this authority to pronounce and enforce law on the expanded social group. I will argue that Classical Chinese culture should obviously count as an expanded social group, but is an exception in that it does not draw its authority from the conceptual apparatus of the law-giving sovereign. It starts from the more natural conception of the authority of the path-finder— expert authority as contrasted with performative authority.11 The political leader is one we follow because we judge that he knows the way to food, water, safety, home, the way to choose a stone to flake to create a scraping knife, how to shape it, etc. Classical China independently evolved a large-scale human political organization that did not draw on this Western paradigm concept of normative guidance, of rule by law. Let me start by visually presenting the contrasts in the conceptual apparatus of the two philosophical traditions for representing human normative guidance and the human capacity for finding and following it. Mind maps are a modern tool. Here, they offer an admittedly simplified shortcut through a lot of (still controversial) theorizing using “radical translation.” I trust the maps can deliver a visual impact of the claims about what it could mean to have different conceptual schemes. The key is how core central linking concepts structure other terms which may be found in different maps. Notice that we need not expect such large differences in our empirical, natural kind

Resurrecting Daoist Virtuosity Epistemology

181

concepts like sun, moon, mountains etc.12 These, as in Quine’s “web of belief ” metaphor, would lie at the periphery of the web with the abstract philosophical concepts illustrated here nearer the center (Quine 1978: vi, 147).13 Consider first the structure (Figure 11.1) an advanced undergraduate philosophy major might be expected to have mastered—as Simon Blackburn put it metaphorically, to know her way around these concepts (Blackburn 1984). By contrast, a student of, e.g., Zhuangzi or Xunzi would be similarly able to find their way around the concepts in Figure 11.2. My strategy in my early career was to use radical translation methodology to identify concepts from the Western analytic structure that were arguably missing or radically displaced in the Chinese concept map.14 I diagram the outcome of that strategy here (Figure 11.3) to remind us of what we are entitled to suspect might be parochial Western conceptual commitments (i.e. those we need not assume are natural human conceptual concerns). There are lessons in the pattern of absent and dislocated concepts. They pivot around the black spot that is logic and the syntactical sentence or proposition. This key quartile of the traditional fields of Western philosophy shapes the inferential context that structures the Western notion of Reason. It creates a conceptual space in which we arrange metaphysical, epistemological, and ethical concepts which give pride of place Reason with all its links to meaning, truth, belief, experience, mind, and thought. Other important contrasts in emphasis accompany these differences in conceptual structure. We are inclined to view the Western mind map as a map of thought. And we use the word “thought” in two ways; one to describe a mysterious process taking place hidden within an individual’s psyche, soul or mind, and the other to refer to the sum of religion + philosophy. As a graduate student in Hong Kong in 1969, I noticed that I could not find the Chinese philosophical works I was looking for under “thought” or “mind,” or next to “psychology.” I would find those books in the education or literature section. The content we call thought (second sense) is the product of transmission between generations by teaching, learning, and practice. The origin is a long (perhaps now lost) historical story, not a case of a teenager sitting under the stars at night, wondering “why are we here?” or “where did this all come from?” We learn our religious dogmas in childhood from parents, teachers, siblings, and peers. We do not individually reinvent them in some rite of passage in a moment of transcendental insight. Confucius viewed knowing as tied to teaching, learning, and practice: “To learn and routinely practice, do we not value this?” (Confucius: 1:1). Thought content is something transmitted, taught, and learned. It is a shared social content coded in behavior, including language behavior—a social 䚃 dàopath. Western translators sometimes misleadingly render the term ཙлtiānxiàsky:nature-below in the Chinese concept map as “the world.” However, it contrasts with ཙൠtiāndìthe cosmos the way natural kind contrasts with social construct. ཙлtiānxiàsky:nature-below refers to the social world (sometimes rendered “the empire”), the world of living things to which the leader is to provide guidance. In Classical Chinese thought the counterpart of the Indo-European (IE) contrast of the world of appearance vs. reality is that between the natural and the socially constructed. And the battle between them is played

182

Figure 11.1 Western thought-world map.

183

Figure 11.2 Chinese thought-world map.

184

Figure 11.3 Map of the Chinese thought world via negativa

Resurrecting Daoist Virtuosity Epistemology

185

out on the field of what should we be teaching and learning, not what should be believed or doubted. Like the Western dialectic structured by the reality–appearance dualism, the Chinese contrast is one of a constant, reliable, external world-guidance vs. an inherited, transmitted changeable social system of behaviors which differ from region to region and time to time. Ancient Chinese masters mapped their concepts around the contrast of natural courses of behavior vs. socially constructed courses of behavior (ཙ䚃tiāndàosky:nature guide Ӫ䚃réndào human guide). Broadly, the conservative Confucians champion the inherited social 䚃 dàopath (hence humanists) while other schools, particularly the primitivist Daoists, attempt to give priority to natural dàos. The contrasting Western focus, the picture metaphor, shapes the Indo-European mind–body dualism—the mind as a container of sense-data experience, ideas and structures of these, beliefs, desires, which interact in an intellectual or mental space—Plato’s world of the forms or the Cartesian realm of individual mind, deploying its defining faculty—reason. The picture in the Chinese case would be the map—a two-dimensional guide to a three-dimensional, real-world information structure. Chinese epistemology and skepticism focus more on how the transmission of social guidance might have been distorted in transmission—as in the Chinese whispers game. Their conception of the internal mechanisms guiding behavior does not separate thought vs. feeling/desire as contraries—hence the famous, sometimes controversial, translation of ᗳxīnheart:guiding organ of the body as “heart-mind.” Motivation and learned content are not distinct—because what is learned is ⿞lˇı ritual_ceremony patterns of behavior, these learned social dàos are the source of normative, behavioral guidance. We acquire moral character (ᗧdévirtuosity) through such learning and disciplined practice. Normativity pervades the entire context of Chinese epistemology and skepticism from the earliest texts. Skepticism and knowledge within the “hundred schools” of Classical Chinese thought, accordingly, address the key natural glue binding a social dào—language. Language, for them, consists of strings of ਽míngwords:names. Language behavior is normatively guided by the same approximate process—teaching, learning, and practicing. As the dialectic develops in Ancient China, issues of knowledge vs skepticism start to pivot more and more around the norms of language (word) use— whether objects, those we mark with names, are socially constructed or natural, worldguided, entities independent of our conceptual schemes. Desires are regarded as directed to objects—of either sort. So, desires too can similarly be natural or socially constructed. This contrast of how dominant a role sociology, history, and culture play in China vs. the individual psychological role, the contrast of subjectivity and objectivity, the notion of the soul/mind/ego in the West is vaguely reflected in the contrast of Chinese ideograph and the Western idea. This notoriously fascinating contrast between Chinese and Western conceptions of language reconfigures our sense of how languages relate to writing and pictures. The characters have a social broadly evolutionary history. They are public, social objects governed by social practices of use. We give Western “ideas” a personal, experiential history. They are subjectively, publicly inaccessible objects.

186

Ernest Sosa Encountering Chinese Philosophy

The gradual adaptation of the written character forms, their modification, the different styles and the popular stories infuse and enliven these fascinating characters in their social history as they are intertwined with events and changing mores and conventions. Like ideas, they are popularly portrayed as originally representational (pictographic or ideographic) items created (like other conventions) by historical sagekings (cultural pioneers),15 then passed down with the social dào encoded in them through teacher to student. These social objects, the written characters, bridge the natural world and the social world. Fingarette noted this contrast of the Confucian social vs. the Western psychological focus and advocated avoiding any psychology to Chinese writers in translation and interpretation (Fingarette 1972). However, it is not that individuals do not have a psychology, but that the psychology is broadly behavioral. ᗧDévirtuosity is predominantly ᗧdévirtuosity 㹼xíngwalk:behavior, excellence in practical interpretation, following paths correctly. We practice and learn these until they become ritualized behavior (including language behavior). Psychology viewed as a mental propositionforming capacity of individuals would be treated as shaped by learning a language. It’s not an inherent structure of a human mind (brain). The function of language is to improve the efficiency of storage, transmission, and access to a social repository of guidance information (䚃 dàopaths) in the world. Humans are distinctive in being able to accumulate, preserve, and transmit that guidance from generation to generation. In doing so, we contribute to children’s and students’ ᗧ dévirtuosity—the ability of the whole person to find, choose, and follow behavioral paths in the real, natural world.

3.1 Early Mohist epistemic reflection So, although we do not find a simple equivalent of “truth” and “belief ” in the Chinese philosophers’ character scheme, we do find a counterpart of that which may be known vs. what is subject to philosophical doubt. The social practice of linking a term to an external, natural object may be a mere, possibly revisable, social convention or it may be guided essentially by the natural world.16 One of the clearest early statements of this different view of what knowing is can be found in The Mozi. A blind person may say “bright white, dark black.” Even the keen-sighted would not correct him. But if we combine black and white natural objects and invite the blind person to pick them out, they couldn’t know how to do it. Hence in saying the blind do not know white-black, it’s not in their use of the words (mastery of a social dào), but in their inability to choose (distinguish particulars in the natural world). Mozi: 31/19/5–6

Further, the difference in the pivot of cognition theories, natural vs. social rather than appearance vs. reality, inspires the Daoist epistemic quietism mentioned above. Theirs is an opposition to imposing social-constructions, learned artifice, on natural guiding structures and natural human capacities for finding and following them. The focus of

Resurrecting Daoist Virtuosity Epistemology

187

their skeptical attitude is on whether there are any naturally correct language uses, not on the vagaries of sense-perception. A corollary of Daoist skeptical lines of thought, “Abandon knowing,” most typically found in Daoist Primitivism, was viewed as entailing that we should avoid the use of language. A cardinal difference lies in the link of the virtue-like notion and its implicit target, 䚃 dàopath. Guidance by a path has three phases where truth of a sentence is a simpler metaphysical or normative relation. Being good at following paths involves finding (knowing-of) them, choosing (ᱟ䶎 shì-fēithis-not that) and following them, not straying. Virtuosity consists in all three. So knowing-of, being able to distinguish the path from what is not the path, knowing-to “this” and “not that,” and knowing how to execute correctly are all included. Paths of natural objects are causal paths. Causation does not operate between sentential events but to a thing’s course of life. Things are composed of their proper parts, not substance (subject) with properties (predicates). The guiding organ takes the information (discrimination) from the senses and makes the this/not that choices not an inner realm, a subjective space populated with picture-like copies of things, facts, and events. Normativity is akin to navigation, not obedience. Ambitions are plans of future unfoldings, not sentential intentions. 䚃 Dàopaths are guidance structures for all living things—not merely humans. The distribution of stuff, things, and parts of things in space creates natural structures of probability which creatures, especially with central nervous systems, can find, commit to, and follow through on. They are structures of causal probability that are relative to things, physical structures that determine possible histories of those things, of how the future might unfold for them if they have the relevant ᗧdévirtuosity. We don’t, for good reason, think of those structures as being chosen, for example, by light rays, but they are otherwise similar—paths of probable unfolding histories of that (kind of) thing made probable by the distribution of mass in the universe. It’s not unusual to find popular accounts of physics and biology that succumb to language that projects intentionality on inanimate objects, and we usually read these charitably. Potential causally probable paths for lots of inanimate objects, famously billiard balls and atoms, are in the layout of stuff in its environment. For living species, we can even read the theory of evolution as broadly intentional, of a species exploiting an environmental niche (a 䚃 dàopath) and discoveries about how even trees release defensive chemicals that alert other trees to dangers like giraffes, parasites, and diseases easily tempt us to something like taking an intentional stance toward them. While personification may be natural to explain trees behavior, it is more naturally natural(!) to speak of them having a way, an inner mechanism, a ᗧdévirtuosity, of exploiting—choosing and effecting for them and their neighbors that future path. We can adopt minimal degrees of animation in describing the behavior without adopting personification in the more implausible sense of their having beliefs and desires and a mental process of pursuing the logic of their law-like beliefs and desires to derive an intention to act in a way that causally has an event as effect. I would claim that so-called primitive personification (animism) is a broadly naturalistic theory—not science, but science-like in being naturalistic and objectively wrong about the kind of mechanism in objects. But there is little that would need

188

Ernest Sosa Encountering Chinese Philosophy

revising in the 䚃 dàopath ᗧdévirtuosity account of this mechanism—only the detail to be added. Even to speak of the dào of the universe from the Big Bang to its cold death is more naturalistic than to postulating a cause which issues mental commands to objects to move in ways that separate light and darkness. Naturalistic versions of the behavior of other things on the face of earth (i.e. the space between sky and earth) would still have normative implications—e.g. to cooperate as far as possible and find and choose among our paths to pursue those which reduce conflict and clashes with others— harmony and cooperation with other natural things. The Western, law metaphor, by contrast, gives us the strict deontology of obedience to the plans and projects of this imaginary being outside of nature. In moral theory, virtue ethics is typically motivated as an alternative normative stance to a law-based deontological model. Virtue ethics thus has an obvious naturalistic motivation. However, without some account of a target, autonomous, pure virtue theory seems unmotivated and notoriously vulnerable to sheer traditionalism— merely listing well-entrenched, thick, descriptive predicates of our language which we typically use in talking about moral exemplars. They may come to seem arbitrary or boringly conventional. In epistemology, virtue is presented as an alternative to process reliabilism. The Chinese metaphor links virtues strictly to processes, to dàos, in general. A thing’s ᗧdévirtuosity includes that thing’s mechanism for finding processes. However, dàos can iterate. There can be several ways of finding ways to do things, and for us, they should be ways humans can find ways. There can be ways of choosing among known ways of knowing, maybe faster, maybe more reliable, more or less easy to execute, more or less cooperative with others, more or less clearly marked, etc. And there can be ways of following, executing, a commitment and ways of finding, choosing, and following those. So ways give creatures a kind of autonomy. Humans can be different from animals in how much more a reflective iteration of dàos we are capable of, especially socially, given our use of language, accumulation of culture, ability to coordinate cooperation etc. But otherwise, human knowing is continuous with the knowing of animals. We can see the point of knowing of, knowinghow, and knowing-to from an imagined state of nature. Knowing of a way of doing the thing (getting water, making a stone tool), executing the plan or project, interpreting it in our performance are natural capacities worth recognizing. The Classical Chinese normative apparatus is deeply naturalistic. Its commitment to natural causal probabilities for the unfolding of the history of various things can be embellished with more science, but needs little core modification to fold easily into our current view of scientific naturalism. The West’s metaphysical commitments, even God aside, to truth, facts, actions, and events as the building blocks of reality, to beliefs, ideas, meanings as units in the brain or central nervous system are, as things now stand with physics and neurophysiology, relatively more risky commitments. Whereas, a commitment to neural paths along which chemical and electrical signals pass in the course of finding, choosing, and following externally marked (epistemically available) paths is an incomplete promise of an explanation, but not the same risk from what we presently know. Most centrally, that alternative way of understanding humans as normative creatures, seeing them as cooperative language users in a shared process of finding

Resurrecting Daoist Virtuosity Epistemology

189

and exploiting natural possibilities, seems a straightforwardly, intuitively natural insight into our normativity. Obviously, it places less of a burden on the isolated individual, the psychology itself, and more on the social construction of paths of cooperative behavior, allowing the group to survive and thrive together in a natural world of actual paths to future outcomes, 䚃 dàopaths for our 㹼xíngwalking:behavior that are cooperative and bring the world of living beings ཙлtiān-xiàthe world of living-things into harmony with each other and with the natural cosmos ཙൠ tiāndì the cosmos:sky+earth.

3.2 Mozi’s state of nature theory of normative knowing That, at least, is what I take Mozi’s state of nature story at the dawn of the classical Chinese philosophical dialectic to have presupposed. It so blatantly resembles Hobbes’ legal-political version that it is almost universally regarded as a complete counterpart. Further, Mozi opposes the more traditional and famous conception of the ཙભ tiānmìng sky fated-role “The Mandate of Heaven,” widely treated as a counterpart to the Divine Right of Kings in the West. That earliest political doctrine is presented in one of the earliest extant texts from the tradition, the Book of History. Tradition held that the Duke of Zhou concocted the narrative to legitimize the Zhou Dynasty overthrow of the Shang empire that preceded it. Both were kingdoms in the sense of being relatively large-scale political systems with a monopoly on coercion, but neither presented anything resembling the legal conception of a sovereign, i.e. of a performative law-pronouncer with de jure authority to create obligations by issuing general commands. And the transfer of the role of ruler was not a grant of such authority. In fact, the Zhou theory developed a different contrast—seemingly between: 1. The natural human, family, authority structure, parent–child authority chains which, combined with traditional ancestor-worship led inductively to the authority of the “ཚ⾆tàizuˇ great ancestor” or “кᑍshàngdìabove-emperor (of the afterlife hierarchical society which contains my grandparents). 2. The ability (ᗧ dévirtuosity) of a leader to find and guide the people along a path, navigating between the opportunities and dangers in the natural world. It was more the expert, non-performative, authority of the skilled pathfinder, finding a way and leading humans to live in harmony with ཙ tiānsky-natural ൠdìearth forces of destruction (famine, flood) and creation, order, and disorder. The roughly Confucian continuation of the ཙભtiānmìng natural role theory was filled out with an account of “Sage Kings,” who first found and led the culture along various paths of social practice. These neither inherited statuses nor law-giving roles. Each ᗧdévirtuoso way-leader picked his most dé aide to take up the leader/pathfinder role, and the social practices established by them are thus transmitted through history using the natural teacher–student model. They, no doubt, may have thought of humans as the only learning creatures, but treating the nature of knowing as tied to teaching and learning is clearly a viable naturalistic approach. Knowledge is taught and learned,

190

Ernest Sosa Encountering Chinese Philosophy

stored and accessed in historically enduring social practices—Confucian ⿞lˇı ritual_ceremony is a Ӫ䚃réndào human guide. Mozi’s state of nature theory challenged the implicit conservatism of that conception of normativity—creating a social process, a social 䚃 dàopath of reforming it, improving it by constructing a better social dào. The guide for that construction was an early version of ཙ䚃tiāndào natural guide based on a distinction that (Mozi alleged) is not a social construction, benefit vs. harm. Mozi’s historically imagined process of constructing the scheme of normative knowledge was not, contrary to some accounts, an authoritarian, top-down process of judgment by the “wisest and best.” It did draw on, but was a dramatic amendment to, the Confucian Sage King model. First, the social world as a whole selected the “wisest and best” and titled him (presumably) ཙᆀtiānzˇı natural philosopher.17 He immediately displayed his wisdom by being epistemically modest. He could not, by himself, say how to guide, to say what was moral for, everyone everywhere. So, as in the prior traditional account, he selected the “wisest and best” ministers not to succeed him, but to help him in the construction project. Each in turn similarly argued that they did not know how to guide everyone, everywhere in their regions. So, they selected deputies, who selected deputies—always choosing the next wisest and best who similarly confessed their lack of overall knowledge of how everyone under them was to live. Finally, getting down to the individual, the process required everyone to “report up” what they knew to be good or bad behavior/practice. Then (by a method left frustratingly vague given the absence of a conception of logic) each ascending layer of wise and good leaders would transmit, report up, until it reached the top judge, the natural (School)-Master who would, Supreme Court-like, decide, thus setting a moral precedent to be followed by all the hierarchical layers below. There the authority was like that in a moral judicial system— absent the notion of law. The precedent setting choice was a precedent of choosing ᱟ䶎 shì-fēithis-not that. Everyone was, in effect, coordinating their situational, contextual judgments according to the precedent established by this ongoing social construction project. The line of discovery was bottom up, and the line of transmission of decision was from the top down—always from the wiser. The process of norm construction started from everyone’s practical experience and was the basic normative know-how. Knowledge was stored in social practices, in behaviors taught, learned, and practiced in the process of making normative ᱟ䶎 shì-fēi this-not that decisions. The joint Mohist accounts illustrate the conception of the nature of knowing in a way that could engage with Sosa’s focus on performance—the performance of distinguishing one thing from another in context: sorting black from white pebbles. Second, the behavioral repertoires can be taught, learned, thus inherited, stored, accumulated, and assimilated to those of a larger group by the natural socialization process of acquiring practices that help us coordinate our behavior with others— starting implicitly with ways of making the distinctions that go into the words of a shared language. Confucius’ social 䚃 dàopath application of the metaphor treated society and human behavior as a structure of ⿞lˇı ritual_ceremony which consists of performances,

Resurrecting Daoist Virtuosity Epistemology

191

behaviors, of named role players. (਽ míngname and ભ mìngstatus_role are semantically as well as phonetically and graphically related terms.) Naming right is assigning the person to the role and right behavior to the one playing that role in the shared social performance. Confucius had a broadly aesthetic appreciation for the elegance and harmony of the resulting social ceremony—which he also likened to music performances and instrument performers. We learn the roles from ᗧdévirtuoso teachers, exemplars, leaders, those who guide us along the performance path as we learn and practice. This social knowledge is both normative and intertwined with the normativity of language/name use—which in The Analects translated to Confucius’ commitment to rectifying names—correctly selecting and titling performers and behaviors (and ritual objects). The aesthetic feature of Confucius’ conception of the information stored in social dàos was one target of Mozi’s skeptical challenge. He argued that the distinction between benefit and harm was not a social construct, but a natural, objective basis which should serve as the ultimate ground of the process of constructing a language to help store guidance information, knowledge. Reforming social practices (including named role players and roles played) would be guided by choosing to bring about general benefit to all. His “state of nature” was a Chinese example of the thought experiment Craig envisioned.

3.3 Xunzi: summary of the classical naturalistic concept of know-to and know-how A relatively succinct statement of this alternative conceptual perspective on knowing comes from a later Confucian, Xunzi. The Mozi and the later Mohists observe a range of ways we use the concept of knowing—as a capacity (one that we are in paradigmatically when alive and awake vs. the opposites) and contact or encounter. As they put it, when awake, the knowing knows. This one comes closest to the Western preoccupation with entry transitions. We know things, concrete particulars, by distinguishing them from their surrounding context using our senses. We come across a path—pretty clearly using the process we saw in the metaphorical structure of the character ᗧ dévirtuosity. Xunzi here fills this line of reasoning out in ways that underline the differences between this Ancient Chinese and the Greek conceptions of perception and the nature of natural practical knowing. In view of this, following what are we able to group similar and different? The answer: following the natural entrances. All things of similar types and essential emotions have natural entrances, way of imaging things that are also similar. So, comparing with the points of doubt in the other similarities is communication. With this process, that which can be shared becomes a conventional name, mutually convenient to the time. The shapes of parts, patterns of color are distinguished by the eye. The sound, the tone, the clarity and muddiness, the tunes of a flute, and other strange noises are differentiated by the ear. Sweet, bitter, salty, and sharp, sour and strange tastes are distinguished by the mouth. Fragrant, smelly, flowery, rank, rancid, putrid, offal and other strange odors are distinguished by

192

Ernest Sosa Encountering Chinese Philosophy

the nose. Damage, itch, cold, heat, smooth, rough, light, heavy, are distinguished by the body parts. Approval, consequence, happiness, anger, sadness, pleasure, love, disgust, desire, dispute by the guiding organ (the ᗳ xīnheart-mind). The guiding organ has guiding knowledge. Having guiding knowledge, it can follow the ear and know a sound is acceptable; following the eye it can know the image is acceptable. In this way guiding knowledge must await the natural entrances hitting on and classifying their types and only after that, can it determine acceptability. The five entrances hitting on something and not knowing it, the guiding organ’s guiding knowledge has nothing to say or approve; in that case humans, without exception, will call this not knowing. Xunzi: 83/22/15–21

And finally, Xunzi’s account of the similarity and difference between animal and human knowing: Fire and water have ≓qìenergy:stuff but lack ⭏shēnglife; grass and trees have life but lack ⸕zhīknowing. Birds and beasts have knowing but lack 㗙 yìmorality. Humans have energy, life, knowledge, and in addition morality. Accordingly, they are the most exalted in the realm of living things . . . . Why is this? Because they are able to form communities, and [animals] are not. Why can humans form communities and animals not? Because they can draw divisions, and how can these divisions unfold in behavior? In morality. Accordingly, being moral in drawing divisions we are in harmony, and in being harmonious, are one and as one our power is greater. We can control other things and bring overall benefit to the Biosphere. There is no other basis for this than in making moral divisions. Xunzi: 28/9/69–73

3.4 Traditional intellectualism vs. its modern naturalist alternatives Mozi points to the broadly pragmatic thrust of Chinese philosophy. Accordingly, it would count as another alternative to the conventional, intellectualist direction of rationalism. It is not obvious how a comparative perspective can contribute to the debate between these traditions. From Dewey’s Reconstruction in Philosophy (Dewey 1920) to Stephen Stich’s From Folk Psychology to Cognitive Science (Stich 1983) to Alvin Goldman’s A Theory of Human Action (Goldman 1970) to the modern proponents of alternative and anti-intellectualist revisions of this tradition, e.g. Wittgenstein, Sellars/ Brandom, continue to make the case from within the tradition, but the tradition survives. I have noticed three defensive strategies for surviving calls to naturalize Western philosophical tradition in the light of modern science and analysis. One is illustrated by Davidson’s observation that trying to “make sense of much of our common talk” is just what philosophers do (Davidson 1969: 162). This rationalist structure permeates not only philosophy writings, but much of Western religion and common sense, indeed our whole sense of our place in the natural world. Straightforward elimination often makes nonsense of ordinary talk.

Resurrecting Daoist Virtuosity Epistemology

193

A second strategy is one we saw in our discussion on Sosa about knowledge constructions like knowing-how and knowing-to. He promptly showed how to translate such constructions into knowing-that constructions using the word “ought.” And doing so, we can observe, has the obvious value that we can reduce the problem of understanding a so-far un-analyzed construction into one we have analyzed for millennia and feel we understand well. A third defense can be found in Christian List’s recent defense of free will (List 2019: 101–102).18 It shares with the other two the advantage that our concepts are inferentially related to the whole network of our discourse about ourselves, e.g. our implicit deontology, and our familiar account of our exalted status in our world. If we are to talk with each other at all, especially about morality, desert, blame etc., we are committed to the broad outlines of that conceptual structure. The practice of philosophy, constructing arguments, usually contained within an article or book very much favors a focus on revising or eliminating some particular part of that structure. Any piecemeal proposal for elimination or radical reconstruction is vulnerable to these defenses. So alternative, non-idealist, naturalistic, pragmatic approaches can usually be fought to a draw in any attempt at piecemeal reform of our conceptual structure. The philosophers within the Western academy do not need (and probably could not rationally draw) help and support from Chinese thinkers. These are highly creative and disciplined thinkers on both sides, and it is natural that the battle will be fought out on their home ground. I do not normally side with partisans for appealing to non-Western philosophy to settle such disputes or solve such problems. My own view is that exposure to other systems of philosophy is beneficial mainly in informing and enriching the familiar process of philosophical thought experiments, philosophical curiosity, and openmindedness. Without fully engaging in experimental philosophy, looking at other traditions gives us a way to ground thought experiments in the real world—of thinking in the context of alternative philosophical traditions with alternative conceptual schemes. Given the stand-off these defensive strategies provide against piecemeal attempts at elimination or restructuring of our traditional conceptual structure, conceptual scheme revisionists may try to construct a whole new, scientifically respectable conceptual structure, one more in tune with the explosion in our scientific understanding of our minds and the world (with some progress in sociology thrown in). A version of this thought, presumably, helped motivate the abortive “ideal language” movement of the twentieth century.19

3.5 The argumentative role of comparative philosophy I do not propose to revisit my arguments for these conceptual differences in this chapter. I also do not intend to argue for the elimination of idealist concepts from our normative theories in general, hence from our normative epistemology in particular. I accept the argumentative force of the conservative strategies in favor of working with our own philosophical texts and the tradition growing out of them. It is the work of

194

Ernest Sosa Encountering Chinese Philosophy

philosophers to try to bring more coherence to our conceptual apparatus and that the enterprise is, in that way at least, internal to that conceptual scheme. Philosophical metaethics is still different from religious traditions of internal reasoning, such as Rabbinical interpretive disputes. The difference is not that we do not have a canonical textual tradition, e.g. the one inheriting and reacting to the Platonic structural conception of knowledge/belief/propositions. It is rather, as Gibbard has suggested (Gibbard 2003: 187–188),20 that internal conceptual structure of our tradition commits us to reject the inference from: P. Our tradition (our social 䚃 dàopath) implies that we ought to X to C. I know to X.

So, while I accept appealing to the conservative strategies to try to rescue and resuscitate the concepts we find natural to use in our philosophical context, we should be careful when relying on them not to rule out the possibility of other ways of talking, of other social dàos. The philosophical goal of clarifying, making sense of “our common talk” should proceed within our understanding our social dào to be itself a product of that kind of internal evolution of an Ancient Greek tradition of thought. Philosophical dispute still differs from the disputes of religious scholars in that we do not take the fact that something follows from our canonical ways of talking (our social 䚃 dàopath) to count in favor of adopting our way of talking. Crucially, we understand our texts, our “way of talking” about normative issues, as including a norm that bars our treating this as a valid inference. We are justified in our conclusion by our reasoning from our norms to the conclusion, but that does not imply we are justified in relying on our norms—with the indexical included. I have argued that the Chinese tradition has a similar kind of built-in objectivism, at least in the Mohist and Daoist tradition.21 It starts with Mozi’s rejection of the inference from the fact that something accords with a traditional social 䚃 dàopath that social agreement makes it ᱟshìthis:right. Still, thinkers from Chinese tradition are as warranted as Western philosophers are in appealing to the three conservative lines of reasoning. They should, like Zhuangzi, allow for doubt that our social dào, our tradition, is the naturally right one. Nature does not speak; does not tell us what to do. We “this” and “not that” our way through available guiding information structures found in nature—always with a degree of epistemic risk. The way in which comparative Chinese philosophy can contribute to the discussion between Western traditionalists and pragmatic, more naturalistic alternatives, is that it is not burdened with trying to argue internally for piecemeal reductions, elimination, or replacements of that conceptual scheme. It rather starts from a whole different, more deeply pragmatic and normative framework. While in this way, more holistic, the strategy does not purport to offer an ideal language derived from science. It is rather a holistic, actual example of an alternative Classical conceptual philosophical tradition, one which actually spawned a philosophical history of comparable length, comparable richness of textual development and dispute, and comparable impact on its culture.

Resurrecting Daoist Virtuosity Epistemology

195

While being naturalistic, and having its own science-like components (Needham and Wang 1956),22 it too should be updated to incorporate what we now understand about the natural world from modern science. As we have seen, the Chinese scheme is not supernaturalist, not only in lacking the author of nature, the law-pronouncing intellectual being, but also in lacking commitment to a host of other abstract intellectual sentence-like atoms of the structure of nature (beliefs, facts, events etc.) or natural kinds (substance + property). So, it offers an interesting, real-world grounded opportunity for a more holistic thought experiment. How could/should the Chinese conceptual system for understanding normativity in general and the normativity of knowing in particular, adapt to accommodate what modern science tells us? Would modern views about the cosmos, about language, about biology, evolution, and the brain accommodate internally into Chinese (Daoist) reasoning using natural elaborations and evolution of its classical conceptual structure in the ways Western thought has with Platonism? First, it should accept (and has accepted) that the norms of language use are not limited to what concepts we are to use in our social, co-operative dào.23 They include norms of grammatical structure that classical thinkers largely ignored. The sentence is indeed a fundamental unit of language. And logic is a way of making the inferential relations among sentences containing shared concepts and their norms of use explicit. A normative conception of modern logic, grammar, and semantics as developed in the West should accommodate itself easily as an embellished linguistic pragmatism. There is more to the 䚃 dàopath of language than the traditional Chinese masters realized. Second, naturalist thinkers in that tradition need embellish, but not abandon, their broadly naturalistic, causal, historical conventionalist view of language normativity to incorporate logic and syntax. They would continue to understand it as part of a social dào that includes all other normativity of socially cooperative behaviors and as something taught and learned, transmitted through generations. They need not treat correctly produced sentences of language as a translation of inner thoughts (mental states), but as participation in the cooperative sharing of guidance information that facilitates social coordination and harmony. Third, like other behaviors, linguistic behavior is learned through example and practiced in interaction with our community. Humans are special in having a ᗧdévirtuosity at carrying on and continuing the productions of sentence expressions in language following the norms, the social dào of their community. They are not special in being in the image of some supernatural thinker or having some normative dignity that elevates them out of the realm of nature. Xunzi’s self-congratulatory view of humans as naturally dominant over other creatures was contentious even when formulated. This important addition to the normative, social-historical view of language would inform the evolved conceptual scheme emerging from the historical dialectic of Confucians, Mohists, Legalists, Daoists, and Buddhists in philosophical debate. Introducing it had already started with Buddhism but did not compel traditional thinkers to adopt the alien view of mind. They still took ᗳxīn heart:guiding organ to be a guidance mechanism, detecting things in nature and responding to them in naming

196

Ernest Sosa Encountering Chinese Philosophy

and social behavior. The learning of language, like the learning of other aspects of a social dào, includes mimicking, emulating, practicing and rehearsing, sub-vocalizing, performances in giving and asking for guidance information to navigate together real avenues of causal probability. Those avenues are actual, not merely logical or imagined, structures of causal possibility and degrees of probability. Navigating them remains a matter of the ᗧdévirtuosity of individuals, schools, and societies. It starts from our capacity to discriminate distinct, concrete particulars in the world and use language in sharing the knowledge we acquire by, as Mozi says, coming across it by following those norms of information sharing. It follows that they would promote the social dào of the scientific community. They would accept the norms of scientific discussion, teach and promote them, and accordingly have a normative conception of truth, validity, evidence, etc., in participation in this social dào. Science, indeed, would fit into their concept of human cooperation as a deeply normative enterprise particularly helpful in discovering, storing, and transmitting natural dàos. It is a way of finding possible ways. Logic and mathematics, axiomatic structures, contribute to science’s ability to store this guidance information, this know-how, efficiently. It makes teaching and learning the knowledge acquired easier, and leaves use of the information more autonomous, less dependent on warranting authority, and ultimately more widely available to humans. The Classical Chinese view of authority, of expert vs. juridical authority, would smoothly integrate with science as a structure. We would, as I must, defer to experts in our linguistic use. I may understand only part of the case for using the term “dwarfplanet” or “Kuiper belt object” rather than “planet” but I know to say there are eight planets and not to count Pluto among them. My epistemic commitment is to the dào of science, not to a body of beliefs. I seek to learn it, to teach it to my students and my son. I needn’t be a front-line theorist to exhibit ᗧdévirtuosity in my role in a social community in which we accept the norms of science. Commitment is understood as a dynamic state of having embarked, having an inertial trajectory, along path-like normative structures. These may be internal to our brain, the learned paths of linked neuronal firings that constitute memories and habits, to learned paths to cooperating with siblings, playmates, teachers, or external trajectory down the road in my car or flying. Commitments can be changed but there is a cost, energy expended in changing them. All sentential commitments could be viewed as potential structures in the dào of our language. This way of adapting science to a Classical Chinese conceptual scheme would naturally incline toward instrumental conceptions of our participation in the social dào of science, but could conceivably reach to accepting instrumental commitments that lead us to know-to treat, e.g., fermions and bosons, dark matter and energy, as real. If doing so makes me better able to participate in this shared enterprise, then I would come to know to regard the world as 96 percent mysterious and perhaps ultimately not accessible to our shared capacities to know. We could develop the Chinese conceptual scheme to function very smoothly with modern science without being required to treat sentential structure as being inherent

Resurrecting Daoist Virtuosity Epistemology

197

in the metaphysical structure of nature outside of and independent of our shared languages and social practices. The social practices must unfold in the context of realworld probable dàos. We need not say the fundamental constitution of the world is as a conglomeration of facts, that causation is a relation among events, that human behavior is an aggregate of actions, that cognition consists in beliefs. The constituents of the world would be particles or strings or . . . causation would be how those particles combine and interact in the probabilities of the unfolding of cosmic history, we would have plans and projects that require us to understand, accommodate, and sometimes guide those probabilities as they unfold.

4 Conclusion: a truce So, the opening to an encounter with Sosa’s account of epistemic virtue remains available given his willingness to consider belief as a quasi-linguistic performance in an normative epistemic social dào in which I and others may (or may not) exhibit degrees of virtuosity relative to our roles in that social dào. It would be manifest as embarking on or maintaining a trajectory down a behavioral path of de-re ⛪wéideeming-treating_as in our cooperative social dào, our normative process of sharing guidance information. In other words, it need not be independent and separable from our language. We could introduce Tarski’s truth-in-L(anguage) but need not commit to an abstract, metaphysical truth beyond the world-guided causal paths of possible–probable unfolding of things. We would commit to expressing truth in our language—following our community’s linguistic dào in knowing behavior as we navigate together the world consisting of natural dàos. Our social interactions and our constructions of social dàos can and should fully exploit the resources and ways of interacting that flow from these insights into how languages function and can be utilized by humans. But we need not ⛪wéideemstreats_as laws as written in the fabric of nature itself—only the dàos are there. And we may or may not be good at finding (constructing), choosing, and following them. The outcome of this thought experiment, I submit, is not an argument that should convince Western philosophers to abandon their traditional commitment to speaking of knowledge as true belief plus something or other. It rather proposes a truce. Chinese epistemology is as entitled to appeal to the three conservative strategies of accommodating their traditional conceptual structure as are Western thinkers. The navigational concepts of 䚃 dàopath and ᗧdévirtuosity work well enough without commitment to natural law and events, to psychological laws of beliefs or normative laws and obedience in intention and action. As neither tradition rules out the other, the main insight from this world-guided thought experiment with an actual alternative conceptual scheme is how it is possible to be an ethical (and epistemological) naturalist naturally. One needn’t go through the complicated process of thinking one’s way out of the Platonic intellectual world from within it. Metaethical naturalism is, to repurpose Simon Blackburn’s metaphor, “harder to earn” for a Western philosopher than it would be for Daoists.

198

Ernest Sosa Encountering Chinese Philosophy

Notes 1

This is because theirs is not a Greek rationalist, intellectual account of cognitive content, of knowing. It was a more embodied account—it is the person who knows, knows to ᱟ(this) and 䶎(not that). Performing involves both the senses and a guiding mechanism or organ which the Chinese identified as ᗳxīn heart:guiding organ. 2 A significant difference is the intellectual “Western” invasion from Buddhism after the Chinese philosophical tradition was about seven to eight centuries old and four centuries after the zenith of its Classical development. The end of that creative burst, like that in the West, had been followed by a relatively (philosophically) “dark” age. 3 This term is usually translated as virtue, and usually with the caveat that it is in Aristotle’s sense of virtue as excellence. My notation attempts to incorporate that caveat into the familiar translation so readers can identify it more easily in translations. It also reminds us that the concept operates only as a paired companion with 䚃 dàopath—one exhibits virtuosity in finding, choosing, and following a possible path of behavior. The Chinese compound for ethics is 䚃 dàopath ᗧdévirtuosity. 4 Although it is common to trace the distinction of knowing-how and knowing-that to Ryle’s development of the theme, some Western classicists find it in the Greek rationalist tradition. See e.g. Gould (1955). 5 This concept operates in a kind of de re belief construct in Classical Chinese. Attitudes are dispositions to treat objects as classified using social constructs, words. 6 These are also implausible in my view, although many take them to be Canonical Daoist commitments. As a professed Daoist, I deny this and treat mature Daoism as the position outlined in the Zhuangzi. 7 Since I will be arguing that the focus of Chinese epistemology is on knowing-how, we might also think of this alternative conception as a version of anti-intellectualism in the context of discussion of knowing-how. 8 The methodology of identifying counterparts is complicated and tentative. See Hyejin Youn et al. (2016). 9 The outlines of our introductory narrative of the history of thought are familiar. When humans had progressed to the point of satisfying their basic needs, they began to wonder “where did the world come from?” and “why are we here?” We construct these stories this way because we know where we want the story to end—with creation myths and supernatural God’s design plans. But why should a natural creature, a branch of the great apes, one of many hominid types wandering the Savanna, hunting, gathering, surviving, have this as their first leisure question? Might we not construct a narrative that makes the first leisure philosophical question that of a Chinese thinker at the dawn of their history as well as some Ionian Pre-Socratic at the dawn of ours? Why must the natural start of philosophy be one in the line of descent from Hammurabi rather than King Wen and the Duke of Zhou? 10 Virtue ethics recognizes this fixation in orthodox systems of Western ethics and seeks to replace it. One of the central objections to virtue ethics is that it is implausible without recognizing some way of distinguishing real virtues from sham ones. However, if we say real virtues are those that generate actions which conform to moral laws, then virtue ethics collapses into law-based ethics. The present, Chinese suggestion, is to link ᗧdévirtuosity to the capacity to find, choose, and follow guidance in natural and historical dàos and remains naturalistic because dàos are natural objects, not verbal productions of beings whose legitimacy is itself poses an ethical issue. The implicit regress of authority is what grounds the moral argument for God—we need

Resurrecting Daoist Virtuosity Epistemology

11

12

13 14

15

16 17

18

19

199

some ultimate, purely rational embodiment of the performative authority to make one’s desires or their utterances normatively real. The creativity of Western philosophy has been required in re-formulating different embodiments of rationality itself, whether God or Kant’s Good Will or Rawls’ unanimous rational choice in the original position. Sociologists (e.g. Max Weber) sometimes characterize this as charismatic authority. Path-finder authority is a less obscure explanation of why one follows the leader. Notice that the concept of leader as authority invites the Western orientation—the notion of placing normative trust in a knowing being is treating him (or her) as authoring normative guidance. However, it’s easy enough to distinguish this version as the authority of the scientific, experienced knower—expert vs. performative or de jure authority. This is the basis of Donald Davidson’s challenge to what I’m doing here—to the very idea of a conceptual scheme (Davidson 1973–74). I don’t argue that Chinese is untranslatable. I agree that it is because we are causally tied to the same physical world and psychically interact with it in species-like ways. However, the concept structure underlying philosophy need not have the same degree of structure (and even the physical term structure might be more different than we naturally suppose—as Quine noted). Both Quine and Davidson, of course, work within the same conceptual scheme—one with truth and belief as coherence pivots. I’m obviously illustrating my conclusions by degree of darkening—totally missing is black, degrees of difference in grays etc. The use of a mind-map to represent the conceptual territory of a language or philosophical tradition is, of course, anticipating one of the larger themes in this chapter—how we can think of inference relations among concepts (meanings) as other than a body of theory consisting of law-like “definitions.” It plays on Simon Blackburn’s metaphor for avoiding Chomsky’s attribution of knowledge of the rules of grammar to children. We know our ways around the language as monkeys know their way through trees—not by knowing laws, but by knowing the paths/ways to construct sentences using one concept to sentences using another. The easiest Western comparison to this Chinese theory of language is the causalhistorical theory of language coiners and the realist inheritance of reference (Kripke 1980). As a modern realist would put it, carving reality at the joints. This way of describing the title is, of course, one that would be treated as contentious by devotees of traditional Western accounts (translations), which fit this in the divine authority model of kingship by rendering it as “the Son of Heaven.” The -zi here, however, is the same as in Kongfu-zi (Confucius) Mozi, Meng-zi (Mencius), Laozi, and Zhuangzi. Further, as my argument shows, the appointee does not even pretend to divine wisdom but protests his ordinary ignorance and inability to construct morality on his own. Despite its appeal to the familiar scientific realist slogan, I suspect this strategy is a variation on the first—with the advantage of being able to say that, e.g. neuroscience doesn’t yet have satisfactory explanations for familiar claims we want to make in ordinary language. We shouldn’t prematurely rule out versions of the claim that, e.g. the universe speaks to us in mathematics. However, we need not confuse that program in philosophy of science with Daoism. The latter is comfortable (I contend) with admitting the universe

200

20

21

22

23

Ernest Sosa Encountering Chinese Philosophy ཙ does not speak, while still claiming to find guidance in nature in the form of dàos which we, with our natural ᗧdévirtuosity, can access to extract information, real-time guidance for behaviors. This is a refinement of Simon Blackburn’s Quasi-realism, which notes that it is an internal part of my system of norms that its being right to do something is independent of the fact that doing so follows from my system of norms. Its role is to remove the indexical from our normative judgments. I take this quasi-objectivity to be implicit in the Classical Chinese account of 䚃 dàopath and ᗧdévirtuosity, and to be the starting point for Zhuangzi’s Daoist skepticism. Adopting a know-to expression of his insight here, I trust not distorting too much the thrust of his study of “Thinking how to live.” Confucianism may be more like a religion in this sense—more likely to take the tradition to be self-justifying. Most philosophical defenders (and Neo-Confucianism) tend to, like Daoism, reject bald historicism in favor of a more naturalistic grounding of normativity. The Chinese tradition as a whole rejects the inference, following Mozi’s argument above. I remind readers of many caveats, including many I have expressed, about Needham’s calling the technological achievements and implicit naturalism of Chinese tradition “Science.” The important continuity is the consistent naturalism internal to that Chinese tradition. It is informed by Mohist appeal to measurement-like standards but lacks the insights from mathematical logic and the ways it informed and structured scientific experimentalism. Although the story here is long and deeply interesting (see Kurtz 2011). China was exposed to systems of Western logic both in Buddhism (second century) and from Jesuits (seventeenth century) and failed to take it seriously, in part because it seemed like a series of verbal tricks designed to get naturalistic Chinese scholars to accept the existence of God or Buddhist absurdism. That changed when logic’s role in science (hence wealth and power) was appreciated in the nineteenth and twentieth centuries.

References Blackburn, Simon. 1984. Spreading the Word: Groundings in the Philosophy of Language. Oxford: Oxford University Press. Confucius. Fifth century BC. The Alalects. Craig, E. 1999. Knowledge and the State of Nature. Oxford: Clarendon Press. Davidson, Donald. 1969. “The Individuation of Events.” In N. Rescher (ed.), Essays in Honor of Carl G. Hempel. Synthese Library (Monographs on Epistemology, Logic, Methodology, Philosophy of Science, Sociology of Science and of Knowledge, and on the Mathematical Methods of Social and Behavioral Sciences), vol. 24. Dordrecht: Springer. Davidson, Donald. 1973–74. “On the Very Idea of a Conceptual Scheme.” Proceedings and Addresses of the American Philosophical Association 47: 5–20. Dewey, John. 1920. Reconstruction in Philosophy. New York: Holt and Co. Gibbard, Alan. 2003. Thinking How to Live. Cambridge, MA: Harvard University Press. Goldman, A. 1970. A Theory of Human Action. New York: Prentice Hall. Gould, John. 1955. The Development of Plato’s Ethics. Cambridge: Cambridge University Press. Hart, H. L. A. 1961. The Concept of Law. Oxford: Oxford University Press.

Resurrecting Daoist Virtuosity Epistemology

201

Kripke, Saul. 1980. Naming and Necessity. Cambridge, MA: Harvard University Press. Kurtz, J. 2011. The Discovery of Chinese Logic. Leiden & Boston: E. J. Brill. Laozi. Third century BC. The Daode Jing. List, C. 2019. Why Free Will Is Real. Cambridge, MA: Harvard University Press. Mohists, L. Third century BC. Canon I. Mozi. Fourth century BC. Harvard-Yenching Mozi Yinde. Cambridge, MA: Harvard University Press. Needham, Joseph and Wang Ling. 1956. Science and Civilization in China Vol. II History of Scientific Thought. Cambridge: Cambridge University Press. Nietzsche, Friedrich. 1899. The Twilight of the Idols; Or How to Philosophize with a Hammer. London: T. Fisher Unwin. Quine, W. Van Orman. 1978. The Web of Belief. New York: Random House. Ryle, Gilbert. 1946. “Knowing How and Knowing That.” Proceedings of the Aristotelian Society Suppl. 20: 1–16. Ryle, Gilbert. 1949. The Concept of Mind. London: Hutchinson. Sosa, Ernest. 2011. Knowing Full Well. Princeton, NJ: Princeton University Press. Stich, S. 1983. From Folk Psychology to Cognitive Science: The Case against Belief. Cambridge: MIT Press. Turri, J., Alfano, M., and Greco, J. (2018, Summer). “Virtue Epistemology.” Retrieved from Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Summer 2018 Edition) https://plato.stanford.edu/archives/sum2018/entries/epistemology-virtue/ Xunzi. Third century BC. Harvard-Yenching Xunzi YinDe. Harvard-Yenching. Youn, Hyejin, Sutton, Logan, Smith, Eric, Moore, Christopher, Wilkins, Jon, Maddieson, Ian, Croft, William and Bhattacharya, Tanmoy. 2016. “On the Universal Structure of Human Lexical Semantics.” Proceedings of the National Academy of Sciences of the United States of America 113(7): 1766–1771. Zhuangzi. Fourth century BC. Harvard-Yenching Zhuangzi Yinde. Harvard-Yenching.

202

12

A Third Platonic Problem for Sosa? Or How Wang Yangming Can Know Better than Full Well? Yong Huang

1 Introduction In his Knowing Full Well (2011), Ernest Sosa aims to provide contemporary solutions to the two Platonic problems involving the constitution of knowledge and value of knowledge, which he deals with in chapters 1 and 3 respectively. In this chapter, I shall introduce a third Platonic problem involving the impact of knowledge, not because the two Platonic problems Sosa deals with are unimportant, but because this third one is not only important in its own right but can also shed light to the first two problems. In the attempt to provide a solution to this third Platonic problem, I shall draw on the philosophical insights from the neo-Confucian philosopher Wang Yangming, according to whom genuine knowledge is one that motivates one to act, and thus one who knows in Sosa’s sense cannot be said to know full well, if the person is not motivated to act according to his knowledge, or if he can be said to know full well, then a person who knows in Wang Yangming’s sense knows better than full well in Sosa’s sense.

2 A third Platonic problem for Sosa The first Platonic problem, in Theaetetus, concerns the nature of knowledge. To Socrates’ question “what is knowledge?” Theaetetus first responds by offering examples of knowledge, which Socrates says are neither sufficient nor necessary for a definition of knowledge. After his response to Socrates’ question by defining knowledge as perception is also rejected by Socrates for many reasons (such as: animals also have perception; no one would then be wiser than anyone else since everyone has the same ability to perceive; anyone who perceives the utterance of a language would be said to know the language; and there is part of knowledge that does not rely upon perception, etc.), Theaetetus comes up with the response that knowledge is true belief or judgment (Theaetetus 200e). To this, Socrates raises the objection: “Then suppose a jury has been justly persuaded of some matter which only an eye-witness could know, and which 203

204

Ernest Sosa Encountering Chinese Philosophy

cannot otherwise be known; suppose they come to their decision upon hearsay, forming a true judgment: then they have decided the case without knowledge, but, granted they did their job well, being correctly persuaded” (Theaetetus 201b–c). So knowledge is not merely a true judgment, and Theaetetus adds that “it is true judgment with an account (logos) that is knowledge; true judgment without an account falls outside of knowledge” (Theaetetus 201d). Then the question is what this account is. Three proposals are discussed but none of them is found satisfactory: “The first would be, making one’s thought apparent vocally by means of words and verbal expressions—when a man impresses an image of his judgment upon the stream of speech, like reflections upon water or in a mirror” (Theaetetus 206d). This however is found not sufficient, as it is “a thing that everyone is able to do more or less readily . . . if he is not deaf or dumb to begin with” (206d). In other words, anyone who has the correct judgment can do it and thus the correct judgment would be knowledge itself. The second proposal is that by account it is meant “being able, when questioned about what a thing is, to give an answer by reference to its elements” (207a). This answer is considered ridiculous, because it is ridiculous to answer what something is by just listing all the parts that make up the thing, especially when a part of one thing may be exactly the same as a part of another thing, as the letter “t” is both a part of “Theaetetus” and a part of “Theodorus.” The last possibility of what an account means discussed is “being able to tell some mark by which the object you are asked about differs from all other things” (208c). This possibility of account is excluded as redundant, as a true judgement of something must already be a judgement that tells this thing apart from all other things. The conclusion of Theaetetus is entirely negative regarding the nature of knowledge: “Knowledge is neither perception nor true judgement, nor an account added to a true judgment” (210a–b). Still, it is reasonable to think that knowledge must be a true judgment or belief, even though a true judgment or belief may not be knowledge yet. So, Sosa states that the post-Gettier form of this Platonic question is “what further condition, added to, or in place of, being competently held, must a true belief satisfy in order to constitute knowledge?” (Sosa 2011: 2). To solve this Platonic problem, Sosa develops his idea of fully apt belief in three steps. First, a belief is apt as animal knowledge when it succeeds in its aim, being true, not by luck but through exercising the first order competences, the competences “that reliably enough yield the correctness of the beliefs produced” (Sosa 2011: 11); second, a belief is then meta-apt as reflective knowledge when it is formed not by luck but through exercising the meta-competence, the competence that “governs whether or not one should form a belief at all on the question at issue, or should rather withhold belief altogether” (Sosa 2011: 12); and third, a belief is fully apt and one knows full well when one’s apt belief, the animal knowledge formed through exercising the first order competence, is guided by the meta-apt belief, reflective knowledge formed through exercising the meta-competence (Sosa 2011: 9–13). The second, related, Platonic problem in Meno that Sosa deals with concerns the value of knowledge in comparison with mere true belief. In other words, what does knowledge add to true belief? As Sosa states, “Plato wondered how knowledge can be more valuable than its corresponding true belief, if a true belief will serve you equally

A Third Platonic Problem for Sosa?

205

well. True beliefs will guide you to your objectives no less efficiently than would the corresponding knowledge. In line with this, we ask: How can knowledge as such always improve on the corresponding merely true belief?” (Sosa 2011: 56). To illustrate it, in the dialogue, Socrates says that “a man who knew the way to Larissa . . . and went there and guided others would surely lead them well and correctly”; here Socrates is talking about a person with knowledge, and he compares the person with someone with merely true belief: “someone had had a correct opinion as to which was the way but had not gone there nor indeed had knowledge of it,” but would also lead correctly (Meno 97a–b). Socrates further says that “as long as he has the right opinion about that of which the other has knowledge, he will not be a worse guide than the one who knows, as he has true opinion, though not knowledge.—In no way worse” (Meno 97b). If this is the case, then it seems that knowledge does not have any additional value in comparison with true opinion or belief, and this being the case, Meno wonders, “why knowledge is prized far more highly than right opinion” (97d). Socrates’ own solution to this problem is that true opinion is not stable and can easily get lost: true opinions “are not willing to remain long, and they escape from a man’s mind, so that they are not worth much until one ties them down by (giving) an account of the reason why” (Meno 97e–98a). Apparently not satisfied with Socrates’ own solution to this problem about the value of knowledge, as one way to solve it himself Sosa explains the value of knowledge through what he calls “a knowledge norm of assertion,” where knowledge, but not merely true belief, can function as a norm of assertation, and in this sense, and to that extent, knowledge is better than merely true belief or has a value that merely true belief lacks. To explain it, Sosa states that sincerity is an epistemic norm of assertion in the sense that “an assertion falls short epistemically if it is insincere” (Sosa 2011: 47). Now one can be insincere and thus acting improperly not only in asserting something that one knows is not true but also in asserting something true without knowing it. Instead, “[i]f an assertion (in one’s own person) that p is not to fall short epistemically it must be sincere, and a sincere assertion that p will be apt only if the subject knows that p” (Sosa 2011: 48). In this chapter, however, I shall not endeavour to show the merits, or demerits for that matter, of Sosa’s solutions to these two Platonic problems. Instead, I shall focus on a third Platonic problem, one discussed primarily in Protagoras. This is not simply because it is also a central issue involving knowledge to which Sosa has not paid sufficient attention; but it is also because, without this Platonic problem adequately solved, the two Platonic problems that Sosa pays attention to cannot be adequately solved, at least in some cases. This is a problem about whether it is possible for one to act against one’s own best knowledge or whether one’s knowledge is inherently motivating. Although this is often considered to involve the issue of weakness of the will, which I discuss elsewhere (Huang 2008 and Huang 2014: ch. 3), in connection with Sosa’s first two Platonic problems regarding the nature and value of knowledge respectively, we may tentatively see this third Platonic problem as one regarding the impact of knowledge. Indeed, this is not an entirely separate problem. A proper answer to this third problem may require us to revise our answer to the first question and help us better answer the second question. This is a question deserving our consideration in

206

Ernest Sosa Encountering Chinese Philosophy

this context, because, for example, if one has a fully apt belief, in Sosa’s sense, that all things considered it is better to not smoke and yet continues to smoke, or that all things considered it is better to tell the truth and yet continues to lie, can we really say that this person knows full well that it is better to not smoke or better to tell the truth? In raising (and indeed solving) this problem, Socrates starts with a common opinion that knowledge “is not a powerful thing, neither a leader nor a ruler . . . [W]hile knowledge is often present in a man, what rules him is not knowledge but rather anything else—sometimes anger, sometimes pleasure, sometimes pain, at other times love, often fear; they think of his knowledge as being utterly dragged around by all these things as if it were a slave” (Protagoras 352b). In other words, most people think that one may act against one’s best knowledge, as knowledge is not a powerful thing. Then Socrates asks Protagoras the question of whether “knowledge is a fine thing capable of ruling a person, and if someone were to know what is good and bad, then he would not be forced by anything to act otherwise than knowledge dictates, and intelligence would be sufficient to save a person” (Protagoras 352c), to which Protagoras responds that “it would be shameful indeed for me above all people to say that wisdom and knowledge are anything but the most powerful forces in human activities” (352d). Socrates’ own view is that “No one who knows or believes there is something else better than what he is doing, something possible, will go on doing what he had been doing when he could be doing what is better. To give in to oneself is nothing other than ignorance, and to control oneself is nothing other than wisdom” (Protagoras 358c). In this passage, Socrates makes two parallel claims: (1) A person who truly knows must have the desire to act according to it. So, for Socrates, it is an absurd position that “frequently a man, knowing the bad to be bad, nevertheless does that very thing, when he is able not to do it, having been driven and overwhelmed by pleasure, and . . . that a man knowing the good is not willing to do it, on account of immediate pleasure, having been overcome by it” (355b), since “No one goes willingly toward the bad or what he believes to be bad; neither is it in human nature . . . to want to go toward what one believes to be bad instead of to the good” (Protagoras 358d); (2) A person who does something worse than he or she “knows” and “believes” does not have genuine knowledge. In other words, this person is ignorant: “those who make mistakes with regard to the choice of . . . good and bad, do so because of a lack of knowledge . . . . And the mistaken act done without knowledge you must know is one done from the ignorance” (357d–e). To explain the apparent fact that a person who “knows,” for example, that smoking is bad and continues to smoke is not ignorant in the same way as a person who does not “know” smoking is bad and continues to smoke is ignorant, Socrates uses an analogy: “things of the same size appear to you larger when seen near at hand and smaller when seen from a distance . . . . And equal sounds seem louder when near at hand and softer when further away” (Protagoras 356c). So, in our example, such a person somewhat does believe that smoking is bad, but also believes that the pleasure from smoking is good, and that the latter is more important than the former, and thus this person does not really know that, all things considered, smoking is bad or he or she ought to not smoke. Put together, these two claims amount to saying that knowledge that does not motivate a person to act accordingly is not genuine knowledge, and thus a person who

A Third Platonic Problem for Sosa?

207

knows that one ought to not smoke and yet continues to smoke cannot be said to have known it full well. In other words, if Plato is right, then what Sosa calls fully apt belief is still short of knowledge. So the solution to this third Platonic problem has the dividend to help us solve both the first Platonic problem about the nature of knowledge and the second Platonic problem about the value of knowledge, the two problems that Sosa aims to solve.

3 Normative knowledge When I introduce the third Platonic problem above, I emphasize “in some cases,” because the third Platonic problem does not seem to apply to knowledge in general. For example, with my fully apt belief that “this is a barn,” this problem does not arise, as there is no specific motivation/action this knowledges implicates. Rather, it only involves a specific type of knowledge, normative knowledge, to which moral knowledge belongs, as in our examples discussed toward the end of the last section. This may explain why Sosa does not discuss this problem, as he writes mostly as an epistemologist and not as an ethicist. Still, we can learn something that Sosa says about ethics, since his epistemology is a virtue epistemology, and in developing his virtue epistemology he often draws its parallels from virtue ethics. For example, in his paper, “Knowledge in Action,” he quotes Aristotle: “Human good turns out to be activity of soul in accordance with virtue, and if there are more than one virtue, in accordance with the best and most complete” (Aristotle EN I 7, 1098a16f). Interpreting this passage using the other passage of Aristotle in which the analogy of a grammarian who has both done something grammatical and done it grammatically, i.e. does it “in accordance with the grammatical knowledge in himself ” rather than “by chance and at the suggestion of another” (Aristotle EN II 4, 1105a22ff ), Sosa states that: Just as the grammaticality of an utterance can be in accordance with the grammatical knowledge or competence in the agent, so the good quality of an action or an activity can be in accordance with the virtue seated in the agent. In both cases, the good quality of the performance must be sufficiently attributable to the competence and not just to chance. Sosa 2016a: 3

Then he starts to develop his virtue epistemology, saying that “Virtue epistemology is analogous to such Aristotelian virtue ethics, and is indeed a special case, as will now be argued” (Sosa 2016a: 4). As a historical example of this epistemological analogue of Aristotle’s virtue ethics, Sosa cites Descartes: If I simply refrain from making a judgment in cases where I do not perceive the truth with sufficient clarity and distinctness, then it is clear that I am behaving correctly and avoiding error. But if in such cases I either affirm or deny, then I am not using my free will correctly. If I go for the alternative which is false, then

208

Ernest Sosa Encountering Chinese Philosophy

obviously I shall be in error; if I take the other side, then it is by pure chance that I arrive at the truth, and I shall still be at fault [. . .]. In this incorrect use of free will may be found the privation which constitutes the essence of error. Descartes [1641] 1996, Meditation IV, par. 104; emphasis added by Sosa

The reason is that, for Sosa, both epistemology and ethics are branches of a general theory of performance, which aims at certain goals. The performance that epistemology is concerned with aims at true belief, while the performance that ethics aims at is right action. Virtue is relevant competence manifested in the performance. Virtue ethics is about the performance acquiring right actions by exercising moral competence, and virtue epistemology is about the performance acquiring true beliefs by exercising cognitive competence. Thus Sosa states that “A virtuous performance, whether a correct belief due to intellectual virtue or a right action due to practical virtue, will involve both the agent’s constitution and his situation” (Sosa 2007: 81). This contrast between virtue epistemology as performance of aiming at true belief through exercising epistemic virtue or competence and virtue ethics as performance of producing right action through exercising moral virtue or competence is not quite precise, however. For one thing, while one may say that the main concern of deontological or consequentialist ethics is performance to produce right actions, virtue ethics is often said to be not focused on the action but on the agent. So we may say that virtue ethics is concerned about performance aiming at cultivating virtuous persons or virtuous characters in persons. For another, while a person’s morally virtuous character or, as Sosa prefers, competence, can indeed be manifested in one’s right action, Aristotle says that it is also manifested in one’s emotion.1 This is so especially in cases of irresolvable moral dilemmas when no actions are right, but a virtuous person always exhibits appropriate emotions. The third and most serious deficiency in Sosa’s contrast between virtue epistemology as concerned with knowledge and virtue ethics as concerned with action, however, is that it misleads one to think that knowledge is absent in virtue ethics, as if epistemology is all about belief, while ethics is all about actions. Yet knowledge, particularly moral knowledge, is indispensable in virtue ethics, and, as I shall argue in the next section, moral knowledge (in contrast to moral knowledge) is inherently action-motivating. Aristotle, for example, emphasizes that “actions done in accordance with virtues are done in a just or temperate way not merely by having some quality of their own, but rather if the agent acts in a certain state, namely, first, with knowledge, secondly, from rational choice, and rational choice of the actions for their own sake, and, thirdly, from a firm and unshakeable character” (Aristotle NE II.4; 1105a). Here, out of the three conditions that have to be met in order for an action to be regarded as virtuous, the first one is knowledge. Now the type of knowledge involved in moral virtue or competence is normative knowledge. For example the knowledge involved in the action of helping a needy person is not merely knowledge about the most efficient way to help the person but more importantly is the knowledge that “I ought to help the needy person.” Although Sosa does not spend much time discussing this type of knowledge, unlike moral noncognitivists and subjectivists, Sosa clearly holds a realist view of such knowledge. In

A Third Platonic Problem for Sosa?

209

other words, he is a moral realist. For example, in his paper, “Moral Relativism, Cognitivism, and Defeasible Rules,” he defends moral beliefs against anti-realists’ challenges, including both non-cognitivism such as prescriptivism and emotivism and subjectivism such as moral relativism (Sosa 1994). In an earlier paper, Sosa makes it clear that our moral knowledge is different from social convention: whether we should drive on the right or, on the contrary, should drive on the left, depends simply on what rule the community adopts. But whether one should kill someone at random and in cold blood just for the fun of it does not seem the same sort of question. Refraining from such activity is not a requirement that derives its force merely from the arbitrary conventions of one’s tribe. Sosa 1987: 7172

So we can assume that Sosa’s virtue epistemology can also provide an account of normative beliefs in general and moral beliefs in particular. According to such an account, a belief is knowledge when it is not only true but is also fully apt, i.e. not only when its truth is acquired through exercising one’s first order competence, and so it is apt, and through exercising one’s meta-competence, and so it is meta-apt, but also when its exercising of the first order competence is guided by its exercising of the second order competence, and so it is fully apt. Now can a person with true belief on moral matters acquired in such a way be said to have full knowledge of it or know it full well? According to Sosa’s solution to the first two Platonic problems, particularly the first one, the answer is an obvious “yes”! In the context of the third Platonic problem we examined in the last section of this chapter, however, the answer seems to be equally obvious, but it is “no”! If I have a fully apt belief, in the sense Sosa defines it, that I ought to help the needy person when I can and yet do not do it, how can I say that I know, to say nothing of knowing full well, that I ought to help the needy person when I can? What is crucial here is that, for Sosa, knowledge and action (corresponding to knowledge) are separate. This may sound surprising, as Sosa often connects knowledge with action. Indeed, the titles of two of his papers published in 2016 are “Knowledge as Action” and “Knowledge in Action” respectively (Sosa 2016a and Sosa 2016b), and one central chapter, chapter 6, of his recent book, Judgment and Agency, has the title “Knowledge and Action” (Sosa 2015). What Sosa means, however, is something different. This is made clear when he says that “Knowledge is a form of action, to know is to act, and knowledge is hence subject to a normativity distinctive of action, including intentional action” (Sosa 2017: 207), and when he says that “the aim of judgmental knowledge is like that of intentional action. This is because judgment  is  a kind of action, with judgmental belief the corresponding intention” (Sosa 2015: 25). So when Sosa says that knowledge is action, he means the epistemic performance intended to form the true belief. This is in one piece with his claim that knowledge is just one class of performance. This must also be true of normative beliefs; that knowledge is action for Sosa here simply means that knowledge is the epistemically competent performance to form a true normative belief, “one ought to help the needy person,” for example, and not that one actually performs the action of helping the needy person. Indeed, since Sosa understands epistemic virtue as competence and not

210

Ernest Sosa Encountering Chinese Philosophy

as character traits, an epistemically virtuous, i.e. competent person, is merely a person who is competent to perform the epistemic action, and not necessarily one who loves or is motivated or has a disposition to do such a performance. Thus, Sosa says that “[c]ompetences in general are dispositions to succeed when one aims to attain certain given objectives. A competence to believe correctly is a special case of that” (Sosa 2019: 22; emphasis added). Notice that here competence is only a disposition to succeed when one tries to attain a certain given objective, not a disposition to attain such an objective in the first place. This is because, for Sosa, “epistemology is not a department of ethics” (Sosa 2017: 149). Here what Sosa means by ethics is what he calls “intellectual ethics” in another place (2007: 89–91).3 It is not yet about whether one is motivated to act corresponding to one’s belief, for example, that one ought to help the needy, but about whether one is motivated to form the belief that one ought to help the needy. Epistemology is not related to the latter, to say nothing of the former. Thus, he states that “there is a distinctive dimension of epistemic assessment isolated from such broadly ethical (or prudential) concerns” as “an admirable love of truth (on a certain matter) and willingness to pursue it with persistent toil and sacrifice” (Sosa 2017: 150) and as “open-mindedness and intellectual courage” (Sosa 2019: 21). It might be said that, even in the case of moral beliefs, as a virtue epistemologist, Sosa here is only concerned about our epistemological competence or virtue, about how true beliefs are formed on moral matters. The question about how a morally right action is produced is the concern of virtue ethics. In this picture, an ethically virtuous person needs to perform an epistemic action to acquire a true moral belief through his/her purely epistemic virtue or competence first, knowing full well what is the right thing to do, and then perform an additional action to bring about this right action through his/her moral or practical virtues. However, since Sosa regards moral action, just like epistemic action, as a special case of the general performance, and he understands moral virtue, just like epistemic virtue, as a special case of competence in general, then the motivation issue is still not solved. Why? For Sosa, to say that I have competence to do something is not saying that I have motivation to do it but only that when and if I’m motivated to do it, I’m disposed to do it well. This is most clear in the competence of driving that Sosa uses as an analogy to explain epistemic competence in his defence of virtue ethics from situationist critics. Sosa defines “driving competence as a disposition to produce driving that is safe when one is at the wheel and efficient in routing one to one’s destination upon getting directions” (Sosa 2017: 184). Notice here that driving competence is not a disposition to drive but a disposition to drive safely and efficiently when one actually drives. By analogy, moral competence would be the disposition to produce the action that is morally right when one is already motivated to do it in the first place. On the one hand, a person who is motivated to do the morally right thing but lacks the moral competence may either not be able to produce the morally right action or produce it merely by luck; and on the other hand, a person, with moral competence or not, who lacks the motivation to do the morally right thing, will not produce the morally right thing.4 Of course, in our context, whether moral competence is morally motivating or not really does not matter. It is really troublesome that the fully apt moral belief produced

A Third Platonic Problem for Sosa?

211

by the epistemic competence, for example, the belief that I ought to help the needy, is not morally motivating, i.e. does not motivate anyone with the belief to actually do things to help the needy. So a person with such a moral belief certainly cannot be said to know full well that he ought to help the needy when he does not do so. This is precisely what is implied in the third Platonic problem that I mentioned above.

4 Wang Yangming: how to know better than Full Well It is here that I think we can solicit help from the Confucian tradition, particularly the neo-Confucian philosopher Wang Yangming (1472–1529) in the Ming Dynasty. The type of moral knowledge we have been talking about here is called liang zhi 㢟⸕ in Wang Yangming, where zhi ⸕ literally means “knowledge,” liang 㢟 literally means “good,” and so the whole phrase literally means “good knowledge.” Since here the term “good” is used in the moral/ethical sense, it is “moral knowledge”; but since for Wang Yangming, liang zhi is not merely knowledge about moral matters but is knowledge that itself is moral or good, in a sense to be explained below, I shall use “moral knowledge” to refer to it in contrast to “moral knowledge.” In this connection, I shall discuss three unique features of Wang’s moral knowledge. While the first indicates Wang’s solution to the third Platonic problem, the other two explain why it can provide such a solution. The first feature of Wang’s moral knowledge is that it inclines one to act: “there has never been one who knows and yet does not act. To know without acting is really without knowing” (Wang 1992: 4). Clearly he is talking about normative knowledge, as no specific action is precisely corresponding to my non-normative knowledge such as my knowledge that there is a barn on the roadside. In contrast, a specific action is always called for by my normative knowledge such as my knowledge that I ought to love my parents. Wang explains the relationship between knowledge and action with the analogy of water and its tendency to flow downward: knowledge is like water, and knowledge’s disposition to produce action is like water’s disposition to flow downward. Just as there is no water that does not (have no disposition to) flow downward, there is no knowledge that does not incline one to act. It is in this sense that he claims that knowledge and action are one in their original state, and thus the effort to know is also the effort to act and the effort to act is also the effort to know (Wang 1992: 42); in other words, knowledge and action are two aspects of the same thing: “knowledge is the intention of the action, while action is the effort of knowledge” (Wang 1992: 4); and “the vivid and solid aspect of knowledge is action, while the intelligent and perceptive aspect of action is knowledge” (Wang 1992: 42). With such a knowledge, Wang Yangming says, “one naturally knows to be filial when seeing father, naturally knows to be loving when seeing old brothers, and naturally knows to commiserate when seeing a child about to fall into a well. All these are the manifestations of the moral knowledge without being obstructed by selfish desires” (Wang 1992: 6). Here what I translate as “knows to be filial,” “knows to be loving,” and “knows to be commiserating” in Chinese are zhi xiao ⸕ᆍ, zhi ti ⸕ᚼ, and zhi ceyin ⸕ᜫ䳡 respectively. In each case, after the same character for the verb “know” (zhi ⸕) is

212

Ernest Sosa Encountering Chinese Philosophy

another verb, xiao ᆍ (be filial), ti ᚼ (be [brotherly] loving), and ceyin ᜫ䳡 (to commiserate) respectively; and so the best translation of such a way of expression in English is to use “know to [act],” as in both this English translation and the Chinese original, what is conveyed is not merely that one knows that she ought to act (knowingthat) or knows how to act (knowing-how), but that the person is inclined or disposed to act. Notice that, in contrast to what we have said above about Sosa’s idea of competence, which is a disposition to act well when one acts (and so the competence is useless if the competent person is not motivated to act in the first place), here Wang’s moral knowledge is a disposition to act (with the content of the act being the same as the content of the corresponding knowledge). It is in this sense that, in a previous paper, I characterize Wang Yangming’s moral knowing as a kind of knowing-to (see Huang 2017). As far as I know, Cline M. Beck is the first person who uses the term “knowing-to” in the sense I use it here: “ ‘he knows to’ [is] used to imply that, given the appropriate circumstances, he does behave in the manner in question, at least most of the time” (Beck 1968: 176). Later, Chad Hansen, when using “knowing-to” to characterize the uniqueness of knowledge discussed among ancient Chinese philosophy, provides similar explanations of this type of knowing: “know-to is prescriptive knowledge . . . . The knowledge is expressed in actions,” and “X knows to do action A” means that “X is disposed to do action A” (Hansen 1992: 253); know-to “is not merely intellectual but includes an attitudinal ingredients: desires and aversions” (Hansen 1992: 252); and, he uses an example of everyday life to illustrate such a meaning: “knowing to come in out of the rain,” emphasizing the motivating aspect of such a knowing. I consider knowing-to in this sense to be a third type of knowing in addition to Gilbert Ryle’s knowing-that and knowing-how (leaving open whether there are additional types of knowing). The fully apt belief that Sosa focuses on is clearly what Ryle regards as the intellectualist knowing-that that traditional epistemology deals with (and contemporary epistemology is still largely devoted to). Surprisingly, Sosa’s competence, with which he understands virtue, is a prime example of Ryle’s “knowinghow.” Let me explain.5 Knowing-how is knowing how to perform certain tasks such as to “make and appreciate jokes, to talk grammatically, to play chess, to fish, or to argue” (Ryle 2009: 17). Indeed, such tasks also include forming belief, and knowing-how in performing such a task is people’s “capacities to find out truth for themselves and their ability to organize and exploit them, when discovered” (Ryle 2009: 17). When people with such knowing-how “perform these operations, they tend to perform them well, i.e., correctly or efficiently or successfully” (Ryle 2009: 17; emphasis added). As we can see, Ryle’s characterization of knowing-how is very similar to, if not exactly the same as, Sosa’s characterization of competence: “What is required for possession of a given competence is that one be disposed to succeed if one tries to ø, which requires not only that one have an ability to ø but also that one would employ that ability (or some other in a restricted ability range) when one tried to ø” (Sosa 2017: 205; emphasis added). Just as Ryle’s person with knowing-how tends to do things well if she does it, Sosa’s person with competence is disposed to succeed if he tries to do it in the first place. Here, just like knowing-that, knowing-how does not motivate a person to act. It

A Third Platonic Problem for Sosa?

213

is precisely in this sense that knowing-to is unique, as it alone can motivate the knower to act accordingly. So a person who knows to help the needy is disposed or inclined to help the person. Only with such knowledge can we solve the third Platonic problem, and a person with knowing-to can be regarded as knowing better than “full well”! Wang Yangming’s solution to the third Platonic problem may appear to be superficial, however, if we don’t explore the second feature of his moral knowing as knowing-to. Indeed, the author of the best-selling textbook in logic, Irving Copi, uses Wang Yangming’s concept of moral knowing as an example of the fallacy of begging the question. He quotes the following passage from Wang Yangming: “There is no such thing as knowledge which cannot be carried into practice, for such knowledge is really no knowledge at all,” as one of the exercise questions under the group heading “Identify any fallacies of relevance in the following passages and explain how the argument in question involves that fallacy,” and then in the answer key to the above exercise question, it is marked “Begging the question (petitio principii). The argument is obviously circular” (Copi and Cohen 1994: 142, 138, 638). We can avoid this misunderstanding only if we examine the constitution of Wang Yangming’s moral knowing. Why can moral knowing be motivating? In Wang’s view, this is because it is not merely belief-like but also desire-like. As he states, “moral knowing is nothing but the heart-mind of discriminating between right and wrong, and discriminating between right and right is nothing but to love [doing the right thing] and hating [doing the wrong thing]” (Wang 1992: 111). So to have moral knowing is, on the one hand, to believe that something is right and something is wrong and, on the other hand, is the desire to do the right thing and to hate the wrong thing. Here the belief and desire are not two separate mental states but two aspects of one single mental state, which is both belief-like and desire-like. To illustrate this, Wang uses the example of our seeing a beautiful flower and smelling a bad odor: seeing the beautiful flower belongs to knowledge, while loving it belongs to [desire to] act. When you see the beautiful flower, you are already loving it; it is not the case that you first see it as beautiful and only later decide to love it. Smelling the bad odor belongs to knowledge, while hating it belongs to [desire to] act. When you smell the odor as bad, you are already hating it; it is not the case that you first smell it as bad and then decide to hate it. Wang 1992: 4

Here, when you see a beautiful flower, you form a belief that it is beautiful and a desire to love it. However, these two are not two separate processes. They are instead two aspects of the same process; and what this single process results in is not two separate mental states, belief and desire, but one single mental state which is both belief-like and desire-like. If you don’t believe that it is beautiful, you will not love it; and if you don’t love it, you will not believe that it is beautiful. Similarly, when we smell a bad odor, we form a belief that it is a bad smell and a desire to hate it. If you don’t

214

Ernest Sosa Encountering Chinese Philosophy

believe that it is a bad odor, you will not hate it; and if you don’t hate it, you will not believe that it is a bad smell. In Wang’s view, “if we can love the good as we love the beautiful flowers and hate the evil as we hate the bad order, we are already sages” (Wang 1992: 97). As Wang’s moral knowing as knowing-to is one single mental state which is simultaneously belief and desire, it can be fittingly labeled as “besire,” a term coined by J. E. J Altham to describe “a unitary mental state which has properties both of belief and desire . . . . It would not be two mental states, one a belief and the other a desire” (Altham 1986: 284). Michael Smith agrees that “besire” is a good label to designate such a mental state, “because, though this state is belief-like, it is also desire-like”; instead of “the contingent co-existence of beliefs and desires,” it is a single, unitary mental state in addition to the mental state of belief and that of desire (Smith 1994: 118–119). Now, it is this conception of moral knowing as constituted of both belief and desire that can explain why such a mental state is motivating. Let me explain. The dominating theory explaining human actions in contemporary philosophy remains the Humean model, which appeals to belief and desire as two separate mental states: belief tells me what to do, while desire motivates me to do it. While there are both anti-Humeans who are rationalists, claiming human actions can be explained by belief alone, and anti-Humeans who are emotivists, claiming that human actions can be explained purely by desire, I think the most plausible alternative to the Humean model is the one developed by such philosophers as John McDowell and Mark Platts, who claim that human actions can at least also be explained by this unitary mental state of besire (although none of them use the term, the term is coined initially to describe their view) to explain human action. They are still Humeans, in the sense that they also think you need both belief and desire to explain human actions, but are anti-Humeans, in the sense that they consider belief and desire to be not two separate mental states but two mutually inseparable aspects of one single mental state. However, I think none of them has developed a theory of besire as clearly, explicitly, fully, and, I emphasize, convincingly as Wang Yangming, as I have briefly shown above.6 Now I would like to discuss the third and the last feature of Wang Yangming’s moral knowing. If the first feature is about its nature and the second its constitution, the third is about its origin: how we come to acquire such moral knowing.7 Gilbert Ryle, in his distinction between knowing-that and knowing-how, states that the way to acquire the former is purely intellectual, while the way to acquire the latter is practical. For example, one cannot learn how to ride a bicycle simply by reading relevant books, listening to relevant lectures, or making relevant observations, the ways that may be appropriate to acquire knowing-that; instead one has to practice riding a bicycle. This shows that the ways to acquire these two different types of knowledge are also different. Wang’s moral knowledge as knowing-to, as we have seen above, is different from both knowing-that and knowing-how and thus is a third type of knowledge. What makes it unique, in comparison with the other two type of knowledge, is that this is a type of knowledge with the motivation to act accordingly, and such a motivation to act is clearly not something that one can acquire either by reading books, listening to lectures, making observations, or by practicing (one need motivation to practice in the first place).

A Third Platonic Problem for Sosa?

215

Then how to acquire such moral knowledge as knowing-to or besire? Responding to a student, Wang states that “one has to make effort with one’s heart-mind (xin ᗳ). Whenever one is not understanding and lacking the motivation to act, one only needs to return to one’s own heart-mind and try to ti 億 it” (Wang 1992: 14). In this passage, there are two crucial terms, one is xin and one is ti. In Chinese traditions in general, and in Confucian tradition in particular, xin is the single human faculty for both thinking (mind) and feeling (heart), and so it is now often translated as heart-mind, a perfect faculty to acquire knowing-to, which, as we have seen is both belief-like and desire-like. Ti is often translated as body. However, there are two things noteworthy as it is used by Wang in this context. First, since Mencius, ti refers to both body in Western philosophy and xin, the heart-mind. The former is called small ti, while the latter is called great ti. Wang is clearly using it to refer to the great ti here. Second, ti is used here as a verb and not a noun, referring to the heart-mind’s reflective inner experiences, which are both cognitive and affective.8 This idea of the self-reflective inner experience (ti) of the heart-mind is central to Wang’s teaching. For example, when one of his students asks where to seek the Confucian way (dao), which ancient sages talk about in different ways, Wang states that “if you would like to really grasp it, you need to experience it in your own heartmind; you should not try to seek it from outside” (Wang 1999: 25–26). This shows that one cannot acquire moral knowledge in the way one seeks scientific knowledge. Elsewhere, Wang states that the Confucian way “must be first experienced internally before one can understand it; it is not the case that one first understands it and then tries to make an effort to internally experience it” (Wang 1992: 74). This shows that Wang’s moral knowing, as a knowing-to, is not initially a knowing-that, whose possessor only later acquires a motivation to act according to this knowing-that; or Wang’s moral knowing, as besire, is not initially a belief, whose possessor only later acquires a desire to act upon this belief. Rather, Wang’s moral knowledge as knowing-to or besire is acquired in the single process of ti, the inner experience, where both the process, the inner experience, and its achievement, knowing-to or besire, as we have seen, are both cognitive and affective. Due to this nature, there is another unique aspect of the acquisition of moral knowledge. This is about the extent of help one can expect from others in this process. One can rely upon others to a great extent to acquire knowledge-that. Indeed, most of our knowing-that is learned from others. Scientific knowledge is a clear case, where not only do we common people rely upon scientists, but even scientists cannot but rely upon other scientists to a great extent. Knowing-how is somewhat different, as this is something that we have to acquire by repeatedly practicing it ourselves. Yet, teachers or coaches are still very helpful, if not absolutely indispensable, in our acquiring it. However, in the case of moral knowledge as knowing-to or besire, we have to almost entirely rely upon ourselves, as no one can do the inner experiences for us or even help us do the inner experiences. It is in this sense that Wang Yangming emphasizes the idea of getting it oneself (zide 㠚ᗇ), an idea originally from Mencius. So Wang states that “what is valuable in learning is to get it oneself ” (Wang 1992: 186). In another place, he says that one can only get a rough idea if one tries to do exegesis of Confucian classics; in order to get the subtlety and the miraculousness, “something that cannot be

216

Ernest Sosa Encountering Chinese Philosophy

expressed by language, one has to do the self-reflection and get it by oneself ” (Wang 1992: 195). In Wang’s view, Confucian classics are about what sages get through their own internal experiences (ti), but since what they get through such internal experiences is not something that can be fully and adequately expressed in language, in order to experience what sages experience in their heart-mind, we cannot rely upon such classics, but we have to experience them ourselves (zide).

5 Conclusion In this chapter, I have argued that, in order to know full well, it is not enough to just provide solutions, however adequate they are, to the two Platonic problems involving the nature and value of knowledge respectively, as Sosa does. A person who knows full well in Sosa’s sense that, for example, he ought to help the needy when he can but is not motivated to act accordingly cannot be regarded as knowing full well in a genuine sense. This involves what I regard as the third Platonic problem involving knowledge. To address this problem I have drawn on the insights of the neo-Confucian philosopher Wang Yangming, who argues that genuine moral knowledge, i.e. moral knowledge, must be one that inclines one to act accordingly. I introduced this third Platonic problem as involving the impact of knowledge, in contrast to the two Platonic problems Sosa pays his exclusive attention to that involve the nature and value of knowledge respectively. Through our discussion of Wang Yangming’s solution to the third Platonic problem, we have now realized that this third Platonic problem is indeed not irrelevant to the first two Platonic problems: only with a proper solution to this third problem is it possible to have proper solutions to the first two problems. With Wang’s solution to the third Platonic problem, we can see now that, in terms of its nature, knowledge is knowing-to constituted of both belief and desire, and part of the value knowledge has, in contrast to merely justified true belief, is that it enables one to act accordingly. Admittedly, the third Platonic problem is not a problem to epistemology in general but is only specific to moral or, more broadly, normative epistemology. To moral anti-realists who think morality does not involve beliefs, this Platonic problem may be regarded as a pseudo problem and can be harmlessly ignored in epistemology. I think, and argue elsewhere, that moral anti-realists are wrong, and Sosa himself is not a moral anti-realist, regarding moral belief as a legitimate concept, and so Wang’s solution to this third Platonic problem deserves our attention.

Notes 1

2

Aristotle says that “virtue of character is a mean . . . in feelings and actions” (Aristotle NE II.9, 1109a); “virtue is to do with feelings and actions” (Aristotle NE III.1, 1109b); and “virtue is concerned with feelings and actions” (Aristotle NE II.6, 1106b). Noah Lemos also affirms that Sosa is a moral realist: “in a variety of essays, Sosa defends the broadly Platonic tradition that there is an objective reality knowable to us, a reality that exists independent of its being known, conceived, or minded by us. The

A Third Platonic Problem for Sosa?

3

4

5 6

7

8

217

objectively real includes, for example, the facts that the earth is round and 7+5=12. Such things are knowable to us and true independently of being conceived or believed by us. But, it also includes, according to Sosa, ethical and evaluative truths. He defends the view that there are truths about what is good or right that are knowable to us and whose truth is independent of our believing them, independent of the belief of an individual or collective or community” (Lemos 2013: 21). Distinguished from theory of knowledge, intellectual ethics “concerns evaluation and norms pertinent to intellectual matters generally, with sensitivity to the full span of intellectual values. It is therefore a much broader discipline than a theory of knowledge focused on the nature, conditions, and extent of human knowledge” (Sosa 2007: 89). In another place, Sosa states that “a pure epistemology is restricted to purely epistemic virtues or competences. That’s how epistemology is properly restricted both in theory of knowledge and in theory of inquiry. And so, virtues or competences that include both epistemic and practical objectives would be better classified under the heading of ‘ethical’ virtues. Such mixed virtues or competences are not directly relevant for a theory of what knowledge is, however, nor even for a theory of intellectual inquiry, of how knowledge is best pursued when extraneous values are excluded” (Sosa 2019: 22). Sosa also states that: “A competence is hence necessarily a competence to ø successfully, for some ø, one tied to a conditional of the following form: that if one tried to ø, one would (likely enough) succeed (given one’s relevant SSS profile)” (Sosa 2017: 192). If this is true, then it seems that there is some problem for Sosa to understand virtue in terms of competence, as virtue is not merely knowing-how. For many philosophers, including Altham, who coined the term, and Smith who also used the term in the passage quoted above, besire is an impossible mental state, and that is why, immediately after coining the term, Altham states parenthetically that “If that [“besire”] reminds one ‘bizarre’, I have no objection” (Altham 1986: 284). The issue here is the so-called conflict of directions of fit. The direction of fit between belief and the world is belief to the world: when they don’t fit, we change the belief to fit the world; the direction of fit between desire and the world is the world to desire: when they don’t fit, we change the world to fit our desire. However, if there is a single mental state which is both belief-like and desire-like, there will be a conflict of directions of fit between this mental state and the world. In a separate paper, I specifically address this issue, showing that this apparently serious difficulty with this mental state is not insurmountable and besire can emerge to be a very promising idea (see Huang 2020). Although Wang claims that such moral knowing is something that we are born with, he also argues that, except in-born sages, whose existence is merely allowed, everyone has selfish desires that obstruct it. So virtually everyone needs to make an effort to acquire or recover it. It is understandable, then, that Tu Wei-ming, perhaps the most influential living Confucian scholar writing today, coined the term tizhi 億⸕, knowledge through ti, to characterize Confucian knowledge. While most of his writings involving this concept are in Chinese, in a couple of articles in English, Tu himself uses “inner experiences” and “embodied knowledge” to translate it, with the former referring to the process of acquiring the knowledge and the latter the end result of the process. See Tu (1979a and 1979b). For my discussion of this idea of Tu, see Huang (2022).

218

Ernest Sosa Encountering Chinese Philosophy

References Altham, J. E. J. 1986. “The Legacy of Emotivism.” In Graham Macdonald and Crispin Wright (eds), Fact, Science, and Morality: Essays on A.J. Ayer’s Language, Truth, and Logic. Malden, MA and Oxford: Basil Blackwell. Aristotle. 2004. Nicomachean Ethics, translated and edited by Roger Crisp. Cambridge: Cambridge University Press. Beck, Cline M. 1968. “Knowing That, Knowing How To, Knowing To, and Knowing How.” Philosophy of Education 24: 171–178. Copi, Irving, and Cohen, Carl. 1994. Introduction to Logic, 9th edn. Upper Saddle River, NJ: Prentice Hall. Hansen, Chad. 1992. A Daoist Theory of Chinese Thought: A Philosophical Interpretation. Oxford: Oxford University Press. Huang, Yong. 2008. “How Is Weakness of the Will Not Possible?: Cheng Yi on Moral Knowledge.” In Roger Ames and Peter D. Hershock (eds), Educations and Their Purposes: A Philosophical Dialogue among Cultures. Honolulu: University of Hawaii Press. Huang, Yong. 2014. Why Be Moral: Learning from the Cheng Brothers. Albany, NY: State University of New York Press. Huang, Yong. 2017. “Knowing-that, Knowing-how, or Knowing-to: Wang Yangming’s Conception of Moral Knowledge.” Journal of Philosophical Research 42: 65–94. Huang, Yong. 2020. “Belief, Desire, and Besire: Slote and Wang Yangming on Moral Motivation.” In Yong Huang (ed.), Michael Slote Encountering Chinese Philosophy: A Cross Cultural Approaches to Ethics and Moral Philosophy. London: Bloomsbury. Huang, Yong. 2022. “Tu Weiming’s Tizhi and Confucian Contribution to Contemporary Epistemology.” Philosophy East and West 72.1. Lemos, Noah 2013. “Objective Value and Requirements.” In John Turri (ed.), Virtuous Thoughts: The Philosophy of Ernest Sosa. Dordrecht, Heidelberg, New York, London: Springer. Plato. 1963. Protagoras, Meno, and Theaetetus. In Edith Hamilton and Huntington Cairns (eds), The Collected Dialogues of Plato. Princeton, NJ: Princeton University Press. Ryle, Gilbert. 2009. The Concept of Mind (The 60th Anniversary Edition). Oxford and New York: Oxford University Press. Smith, Michael. 1994. The Moral Problem. London: Basil Blackwell. Sosa, Ernest. 1987. “Serious Philosophy and Freedom of Spirit.” The Journal of Philosophy 84(12): 707–726. Sosa, Ernest. 1994. “Moral Relativism, Cognitivism, and Defeasible Rules.” Social Philosophy and Policy 11(1): 116–138. Sosa, Ernest 2007. A Virtue Epistemology: Apt Belief and Reflective Knowledge, Volume 1. Oxford: Clarendon Press. Sosa, Ernest 2011. Knowing Full Well. Princeton, NJ: Princeton University Press. Sosa, Ernest 2015. Judgment and Agency. Oxford: Oxford University Press. Sosa, Ernest 2016a. “Knowledge in Action.” In Amrei Bahr and Markus Seidel (eds), Ernest Sosa: Targeting His Philosophy. New York: Springer, pp. 1–13. Sosa, Ernest 2016b. “Knowledge as Action.” In Chienkuo Mi, Michael Slote, and Ernest Sosa (eds), Moral and Intellectual Virtues in Western and Chinese Philosophy: The Turn toward Virtue. New York and Oxford: Routledge, pp. 5–15.

A Third Platonic Problem for Sosa?

219

Sosa, Ernest 2017. Epistemology. Princeton and Oxford: Princeton University Press. Sosa, Ernest 2019. “A Telic Virtue Epistemology.” In Heather Battaly (ed.), The Routledge Handbook of Virtue Epistemology. New York and London: Routledge. Tu, Weiming. 1979a. “ ‘Inner Experience’: The Basis of Creativity in Neo-Confucian Thinking.” In Tu, Humanity and Self-Cultivation: Essays in Confucian Thought. Berkeley: Asia Humanities Press. Tu Weiming. 1979b. “The Unity of Knowing and Acting—From a Neo-Confucian Perspective.” In Collected Works of Tu Weiming ᶌ㏝᰾᮷䳶, vol. 5. Wuhan: Wuhan Chubanshe ↖╒ࠪ⡸⽮. Wang, Yangming ⦻䲭᰾. 1992. The Complete Works of Wang Yangming ⦻䲭᰾‫ޘ‬䳶. Shanghai к⎧: Shanghai Guji Chubanshe к⎧ਔ㉽ࠪ⡸⽮.

220

13

Responses Ernest Sosa

1 Introduction First of all my sincere gratitude to everyone: to Yong Huang, to our hosts at Fudan, and to the colleagues who wrote papers for the conference. Thanks to all for participating, and for all I’ve learned about Chinese philosophy and the ways in which it anticipates contemporary virtue epistemology. I appreciate this opportunity to respond. I don’t want to overtax readers’ patience, so I will restrict myself to critical questions raised about my work. I will be responding to these, but mostly I will respond more globally on critical themes that overarch most of the chapters. My constraints preclude responding extensively to every chapter, but I have read the chapters closely and have found much insight and food for thought in every one of them. I will be focusing mainly on common themes of the chapters, which are said to bring my virtue epistemology close to important themes in Chinese philosophy.1 This will take up Part I of my commentary. Especially towards the end of that discussion I am led to some new extensions of the virtue theoretic account, and I am especially grateful for the probing and comparisons that have led to those extensions. But I do want to respond, even if perforce briefly, to all of the chapters. So, in Part II I will take up specific issues that do not fit into those overarching themes of Part I.

2 Part I A. A question of methodology: What sort of analysis might we pursue in epistemology? Is it semantic, conceptual, or metaphysical? This question arose in our conference. Since it has a sort of priority, I will essay at least a brief answer preliminary to our treatment of substantive issues to follow. We can discern three quite distinct sets of methodological issues, however closely they may be interrelated. First are issues of the semantic or pragmatic linguistic analysis of epistemic expressions such as “S knows that p.” Second are issues of conceptual analysis of one or another sort. Here one asks about how a certain concept is constituted, 221

222

Ernest Sosa Encountering Chinese Philosophy

what it necessarily involves. About a concept of knowledge, for example, does it necessarily involve both a concept of belief and a concept of truth, so that necessarily the former concept will apply to something only if the latter concepts do so as well? Third are issues of metaphysics. Here we focus on an objective phenomenon that need be neither expression nor concept. Our focus is rather on a state that people host, or an act that they perform. This is the phenomenon whose ontology we now wish to understand. What is human knowledge? What is its nature? What in general constitutes a case of knowledge? The questions in this third set, which are those I have mainly addressed in virtue epistemology, arise in several of the chapters, with important bearing on a supposed division between two varieties of virtue epistemology, the “responsibilist” and the “reliabilist.” In my view, this alleged division is a false dichotomy. Although I remain convinced of that, I now see a better way to articulate my dissent.

B. The place of testimony 1. Winnie Sung’s perceptive chapter (Chapter 7) correctly attributes to me a position at neither the reductivist nor the non-reductivist extreme on the epistemology of that important source of knowledge. So, I accept with thanks Sung’s account of my view, and would like now to develop that view a bit further than in earlier publications. 2. According to my view, there is a deep analogy concerning types of knowledge, between the instrumental and the testimonial. Both instruments and testifiers deliver information to us, instruments through their readings, testifiers through their assertions. In each case there is a deliverance of

that makes it seem that p to the receiver. There is thus a sort of uptake of a proposition on the part of the receiver. This “uptake” has propositional content, and can be characterized with at least two distinguishable modalities: mere uptake as if p, and “gathering” that p. Such deliverances can be “taken” through “uptake” in either of these modalities. On my now expanded view, uptake can be either instrumental or testimonial. In either case, uptake as if p is non-factive: you as receiver take it that p, and it thus seems to you that p. This is non-factive, since it can seem to you that p while it is false that p. 3. However, by resolving an ambiguity in “gathering,” or even by stretching English just a bit, we can recognize a factive sort of “gathering” such that what you thus gather must be a fact, and the gathering must be more than just an accidental acquisition of that fact, which happens to fall into your hands. Rather, by definition, what you thus “gather” must not just accidentally fall into your hands. In order for you to gather that fact, you must gain possession of it through your own doing, not just by accident. This doing can be aided by other entities that are “instrumental” in your gathering. Thus an instrument can aid your gathering of a fact through proper issuing of a certain reading. Alternatively, a testifier can aid your gathering that fact by properly reporting it to you. I hasten to acknowledge, however, that English does not unambiguously assign such meaning to the term “gather.” I am inclined to think that English does recognize my

Responses

223

preferred meaning. But I need not insist on that, I hope, in these days of conceptual engineering. 4. In any case, we can assess uptake as if p, which we here understand as just a seeming as if p. If the instrument or the testifier is not proceeding properly, if they are malfunctioning or lying, then the seeming as if p does not graduate into a proper gathering of the fact that p. On my virtue-theoretic understanding of these phenomena, we can understand this distinction between the non-factive seeming as if p and the more accomplished gathering that p, as follows. Upon accepting a deliverance as if p from an instrument or a testifier, it may thereby seem to you as if p. That may be all that the mere deliverance delivers to you on that occasion. In order for a deliverance to do more, however, in order for it to enable you to gather a fact thereby, three conditions must be met: first the proposition delivered, whereby it seems to you as if p, must of course be a fact. Second, for you to gather that fact, for it not just to fall into your hands by luck, you must somehow bring about your acquisition of it, and you must do so in a way that is causally creditable to you, through some exercise of pertinent competence. In these cases of instrument-aided or testifier-aided gathering, your gathering is aided by external help. The instrument may deliver to you the proposition that p without delivering the fact that p. Even when the proposition delivered is true, that fact would not be delivered through a malfunction. A malfunctioning instrument might still give you the right, true deliverance by lucky accident. In this case you do not gather the corresponding fact. In order for you to gather the fact that p through an instrument’s deliverance of that fact, the instrument must then function properly, so that aided by its proper functioning you can yourself function properly in its seeming to you that p. So, the instrument-aided seeming that p must here be through your exercise of instrument-aided competence, with the instrument functioning properly. And, what is more, your thereby getting it right that p must itself manifest pertinent competence on your part. 5. And we can now see interesting relations among testimonial, instrumental, and perceptual sources of knowledge and epistemic competence or justification. All three are naturally conceived as sources that deliver propositional deliverances. In each case we have the deliverance of mere seemings as if p. All three such seemings can then graduate to gatherings that p. You gather a fact that p when it seems to you (all things considered) that p and this seeming is veridically apt: i.e. when its correctness is through competence and not just pertinent luck. Suppose it seems to you that p (whether testimonially, instrumentally, or perceptually). In that case, you will thereby gather that p if and only if it seems to you that p, all things considered, and you thereby get it right not just by luck but through competence, through the exercise of a competence. This competence may involve your perception, or your ability to properly read one or more instruments, or your ability to properly “read” your testifying interlocutors. 6. Plausibly, however, our testimonial source of thereby knowledgeable deliverances is fundamentally different from any instrumental source. How so? They differ importantly

224

Ernest Sosa Encountering Chinese Philosophy

in that we learn to trust our various gauges or other instruments by relying on other sources, such as perception and testimony. By contrast, our intellectual agency is social deeply enough that our testimonial source of propositional deliverances plausibly has the sort of default standing found also in our perception.

C. Responsibilism vs. reliabilism in virtue theory 1. Let us consider the nature of intellectual virtues and their place in the constitution of knowledge. On this question, responsibilists and reliabilists are often thought to differ in that reliabilist virtue epistemology restricts itself to basic faculties, such as vision, memory, and inference, as opposed to deliberative judgment with its distinctive character traits such as open mindedness and intellectual courage. This surfaced repeatedly in our conference. Moreover, our conference was not an outlier in that regard. In fact, that division of two “sorts” of virtue epistemology has congealed into a standard account. But it is seriously misconceived, deriving as it does from a simple misreading. 2. Nothing in virtue reliabilism denies to the free and deliberative exercise of judgment its proper place in human knowledge. In order to be knowledgeable, judgment too needs to be apt, which means that it must get it right through the exercise of a reliable enough resultant competence.2 3. That being so, it is hard to see how anything in Chinese virtue theory could favor responsibilist over reliabilist virtue epistemology. A detective needs more cognitive sophistication for arriving at apt judgments than does a lookout atop a ship’s mast. And that is of course typical of the immense sophistication required for success in artistic, athletic, and professional judgment, and even in ordinary human interaction. But why should reliabilist virtue epistemology be any less interested in such sophisticated knowledge than in the simplest perceptual knowledge of the lookout? One detective can be more imaginative than another, to his great advantage, by considering a much broader array of possible answers for determining guilt. And a better detective might also be helped by reasoning more reliably, besides imagining more imaginatively. Finally, such greater reliability is in no tension with the freedom that each detective exercises in judging based on the deliverances of their senses. 4. What then is the place of reliable free judgment in human cognition? This issue was raised repeatedly in our conference. It concerns the relation between analytic virtue epistemology and its closest correlates in Chinese philosophy. And it concerns also how epistemology is related to ethics. Accordingly, these will be main questions addressed in the remainder of Part I of my contribution. (And my later remarks on Pyrrhonian skepticism and Buddhist detachment, in Part II, also fit under natural extensions of this overall theme.) How then is epistemology related to ethics? In earlier writings I have claimed that epistemology is not a department of ethics, despite both being normative disciplines. I will now try to develop the essentials of that

Responses

225

view. The view I lay out still conceives the normativity of epistemology as a form of “telic normativity.” In now stating the view, I will be responding to critical points about it in one or another of our chapters.3 I begin with a brief summary of my presupposed telic normativity.

D. Telic virtue epistemology 1. The five main ideas of my account are those of attempt, success, competence, aptness, and achievement. Archery not only provides an example of a telic triple-A normativity constituted by those five ideas; it also shows how achievement comes in degrees within two dimensions. One dimension is that of the apt shot, accurate because adroit. The other is that of the fully apt shot, where the agent aims not just at accuracy but at aptness and succeeds through competence in this more complex endeavor. If a shot is too risky, it is ill-advised. A shot can attain quality in the specific regard of being well selected. A well selected shot can thus rate higher in that regard than one that falls short through pertinent negligence or recklessness. A dimension of second-order evaluation of huntress Diana’s shot thus involves more than its aptness, its success through (first-order) competence. Also relevant is whether the attempt is well selected so as to avoid recklessness, and even negligence. That is so even if the dimension of interest to us is that of creditability, independent of moral concerns such as whether the act is a murder, and independent of the admirability that involves degree of difficulty. An instance of absolutely certain knowledge, of the highest degree of relevant epistemic quality, need involve no difficulty at all. Think of the cogito! When successfully enough guided that way, an attempt rises to the level of the fully apt. Nothing short of this will suffice for achievement full well. If an attempt succeeds aptly without being fully apt, there is an element of relevant luck in its success. Its aptness is not secured through the guidance of the agent’s second-order competence. 2. Applying such virtue theory to the epistemic domain, we reach a main thesis of virtue epistemology: that the normativity of knowledge is a special case of such telic normativity.

Knowledge is thus a central sort of epistemic achievement. Here we find the traditional issues of skepticism, and other issues of the nature, scope, and value of knowledge.4 3. Agency in my view takes two broad forms. Aside from the agency of consciously intentional judgment, epistemic agency can alternatively take the form of functional agency, where one pursues an end not through conscious intention and deliberation but through functional teleology, as when the heart beats aiming to move the blood (so as to aid the survival and continuing health of the organism). That is an agency exercised by humans through the telic deliverances of implicit cognitive functioning,

226

Ernest Sosa Encountering Chinese Philosophy

as when our perceptual systems aim at correctly representing our surroundings. Such human cognitive functioning begins thus with deliverances of the senses and leads functionally to full-fledged belief. This then is teleological, aimed representational agency, even if it does not involve the conscious intention constitutive of what I distinguish as “judgment.” In his excellent account of my virtue epistemology, Chris Fraser (Chapter 3) worries that much epistemic action falls below the more exalted level of conscious judgment based on intentional deliberation. Although I agree with that claim, we have also just seen how occurrent belief can alternatively involve the agency of functional representation, even if this is the agency of a subconscious functional teleology. This distinction between functional and intentional teleology is explicit in my Judgment and Agency (Sosa 2015). That book makes explicit the pertinence of virtue epistemology to both sides of epistemic agency, both the functional and the consciously intentional. That book of mine focuses mainly on judgmental belief, however, and on judgment, where the agent does aim with conscious intention to get it right aptly on a given question. What is involved in this? An alethic affirmation might be just a guess, as when a contestant tries to affirm the correct answer to a quiz show question. But a physician would aim not just to guess but to affirm competently, indeed aptly. Only an alethic affirmation can amount to a judgment, which it can do only if it aims not just at truth but also at aptness. This yields the following hierarchy. Saying of “p.” Affirmation: saying that p. Alethic affirmation: endeavor to get it right by affirming that p. Judgment: endeavor to get it right aptly by alethically affirming that p.

These can all be public, in outer speech, or private, in silent soliloquy. They are commonly and generally free acts.

E. Responsibilist vs. reliabilist virtue epistemology? 1. Such judgment-involving virtue epistemology is not just an epistemology of automatically operative faculty mechanisms, such as our perceptual mechanisms functionally aimed at correct representation. How then does virtue epistemology make room for the virtues involved in sophisticated human judgment? First, as I will argue, not in the way of responsibilist virtue epistemology. True, the higher faculties of deliberation and judgment have always been welcome in virtue epistemology, but always with a clear recognition that the character traits of responsibilists have limited significance for the history of epistemology, with its central focus on issues of philosophical skepticism. It might be argued that philosophical skepticism concerns not so much knowledgeable judgment as justified judgment. But the sort of epistemic justification relevant to philosophical skepticism must be understood through the contribution

Responses

227

that such a belief would make toward the subject’s knowledge. (If you threaten to shoot someone unless they believe in the external world, you give them a pragmatic reason to so believe, but that is not the sort of “epistemic” reason required for a solution to the problem of the external world.) And it is hard to see how such epistemic justification could be explained without reference to knowledge, which would take us after all to issues of theory of knowledge, or, for short, to issues of “gnoseology.” Let me now try to explain my sharp contrast between such gnoseology and intellectual ethics. These I believe to be quite distinct departments of epistemology. Several chapters touch on this issue. Xiang Huang (Chapter 6) focuses his main critical attention on it, for example, and concludes his insightful discussion with the critical remark that “the inner circle of traditional epistemology need not be considered an important constraint on settling the inner circle of virtue epistemology.” This is quite in line with the responsibilist critique of my views, and so in what follows I will be responding to Huang’s critique. But before I focus on that critique, I would like to set the stage more broadly by explaining the scope and motivation of my approach. 2. Consider how we might draw the contrast between, first, gnoseology or theory of knowledge, and, second, intellectual ethics; and how we might show these to be quite distinct departments of epistemology.5 The virtues highlighted by responsibilists are those that satisfy criteria such as the following. They are exercised in intentional agency, are developed through the repeated practice of such agency, bear on the personal worth of the agent, and aid agential success; and, in epistemology they concern intentionally conducted inquiry. Such responsibilist virtues are also thought to involve an important motivational component. They supposedly bear on the personal worth of the agent largely because of their distinctive motivation and respect for truth. Knowledge itself is said to attain its special value when the success of its constitutive belief is attained through the distinctive motivation constituted by respect for truth. This value derives from the dignity of a human motivation that freely aims at truth in epistemic domains (as it might aim at the good in the ethical, and the beautiful in the aesthetic). There is no similar dignity in how a mere device mechanically attains its success, however consequentially good that success may be, and however reliably the device might deliver its truths. 3. There is insight in that account, but it misfires as a critique of virtue reliabilism. The reliability required by virtue reliabilism for the epistemic standing of a belief is a necessary condition. It is never supposed that all knowledge is to be found at the level of animal knowledge. Nowhere in the writings that develop virtue reliabilism is there any restriction to such basic faculties or to the representations that we humans share with lower brutes so as to give us our animal knowledge. Virtue reliabilism does not restrict the representations of human animals to those of functional teleology, as with the deliverances of our perceptual systems and their immediately corresponding functional beliefs. A higher human judgment is also explicitly recognized, one involving higher faculties also subject to the AAA/SSS virtue-epistemic structure.6 And such higher epistemic performance is given the very same analysis that is left at a more determinable level in the earlier discussions of

228

Ernest Sosa Encountering Chinese Philosophy

animal knowledge. Earlier the focus is on examples of animal knowledge at elementary levels such as the Cartesian or the Moorean . The main target of analysis is then simple, first-order animal knowledge, which admits functional teleology, though it need not remain at that implicit functional level. Thus, if someone asks René Is the fire on in the fireplace? he would presumably have a look and might conclude through conscious inquiry that Yes, here in the fireplace there is a fire. But it is absurd to suppose that the focus of Descartes and Moore on such elementary cases revealed a lack of interest in more sophisticated knowledge, or an interest restricted to only the elementary cases. Their focus is obviously meant to reveal the depth of problems that arise more clearly even in such simple, elementary cases. It is certainly not due to a myopia that puts the more sophisticated cases beyond them. And that is certainly how it is with reliabilist virtue epistemology. The objective is initially an account of knowledge defensible against the skeptical attacks found in the history of epistemology from the Greeks to Descartes and beyond. Moreover, a higher reflective knowledge is always explicitly recognized. And the basic concept of adroit competence has never been restricted to the basic animal faculties of perception, memory, etc. What has misled many, I surmise, is the terminology of a knowledge that is “animal.” This may have unfortunately suggested that the sort of first-order knowledge of interest to reliabilist virtue epistemology is only the sort of knowledge that we share with animals generally. But my “animal” knowledge, like Descartes’ “cognitio,” is the knowledge that a knower would have even unaided by any reflective higher-order perspective. It is left open that no human knowledge is ever purely “animal,” unlike perhaps the sort of knowledge that may be found in animals lower on the evolutionary scale. So, purely animal knowledge may never appear among adult humans, who always have an at least implicit involvement of higher order endorsement in any of their knowledgeable beliefs. Moreover, the judgments and beliefs that qualify as cases of animal knowledge among adult humans are mostly derived from competences that go far beyond basic perception, memory, etc. Even without rising much above the level of first-order perception, that is because of the essential involvement in such adult human perceptual knowledge of perceptual discrimination and perceptual judgment that far surpasses the basic levels of animal belief. We need not even invoke art critics, wine connoisseurs, or our sensitivity to other minds. Just consider the enormous scope for human perceptual uptake in a technological setting such as ours, where we perceptually identify innumerable devices and settings as we navigate our physical and virtual environments. That suggests how much gathering of complex fact is accomplished constantly by adult humans, even without invoking the aid of higher-order assessment. So, we can distinguish the sufficiently competent exercise of such competences, and understand the first-order aptness that a thinker might attain thereby, even if that is never unaccompanied by higher-order endorsement. Such aptness of alethic judgment and belief can then count as “first-order” aptness, and hence as “animal” knowledge, even if no human beyond the age or reason ever attains such aptness with perfect first-order purity untouched by reflective competence.

Responses

229

Although my view has stayed the same, its focus has shifted explicitly to the judgmental knowledge that requires the conscious pondering of a detective, an art critic, or a philosopher. That is not to change the view, in the sense of recanting. It is rather to develop a view that was already determinably there. The new emphasis (especially in Judgment and Agency) is on the more sophisticated pondering or thought that leads agentially to the making of a judgment, or the forming of a belief. 4. Human knowledge can attain the highest levels of such distinctively human dignity, but must still satisfy the requirements of competent reliability and pertinent aptness. These requirements must be met not only by the functional knowledge that attains its proper aims of accurate representation found in the elementary perceptual knowledge of a fire or a hand. They must also be met by the judgmental knowledge that attains its epistemic aims through the consciously deliberative engagement of our higher, more distinctively human intellectual faculties. In fact, the point can be made by showing how Descartes himself is a virtue reliabilist, a radical one. The Internalist-in-Chief of the tradition, whose main epistemic faculty of Judgment resides in the will, is an extremist reliabilist. After all, what sort of reliability did he require for epistemic certainty? As is clear and distinct in the early paragraphs of the Third Meditation, what is required for Cartesian certainty is nothing less than infallible reliability, and the account of such certainty does not stop with truth and psychological assurance. Rather, it involves a virtue reliabilism of aptness, of full aptness indeed. Or so I have argued in detail elsewhere.7

F. What sorts of virtues are constitutive of knowledge? 1. Finally, we return to the relation between ethics and epistemology. By focusing on the titular question of the present section, can we not discern a further critique according to which virtue reliabilism overlooks character traits of special interest to virtue responsibilism, among which open mindedness and intellectual courage have held center stage along with intellectual humility. Responsibilist virtues are said to involve giving proper weight to moral and other value considerations. Here’s an example of how that is supposed to go. Take intellectual courage. That is supposed to concern a judgment that weighs personal safety and moral considerations along with whatever needs to be considered in determining truth reliably enough. Prudential and moral concerns are thus in play along with accuracy in the exercise of such responsibilist virtues. The exercise of such intellectual courage thus often depends on the agent’s weighing of personal risk. 2. I agree that there are such virtues, even obviously so. They are part of the wisdom of intellectual ethics, which weighs such values when they come into conflict in human life. Far from denying that, I myself would emphatically affirm it. What virtue reliabilism denies is that these responsibilist virtues are constitutive of human knowledge. Someone negligent or even reckless with regard to the morality or prudence of their inquiry may still attain the very highest levels of proper epistemic certainty as an outcome of their researches. And that remains true regardless of how objectionable

230

Ernest Sosa Encountering Chinese Philosophy

their inquiry may be in prudential or moral respects. The gnoseological assessment of inquiry and its resulting judgments is unaffected by such prudential or moral assessment. Suppose we understand conscientiousness as a virtue whose exercise enables agents to avoid pertinent negligence and recklessness. In my view that is a broad, determinable category, with determinate sorts falling under it. In particular, conscientiousness comes in sharply distinguishable prudential, moral, and gnoseological sorts. 3. My reference to “gnoseology” is meant to distinguish a subdiscipline of epistemology concerned to understand knowledge and skepticism and pure inquiry in particular. A broader epistemology also includes questions of intellectual applied ethics and prudence. These are akin to the questions of medical ethics and business ethics. Applied ethics of the X sort is generally ethics restricted to performances of the X sort. They might be of the medical sort in one case, of the business sort in a second, and of the intellectual sort (broadly conceived) in a third. Intellectual performances include those of inquiry and those of judgment, among others. And these would seem assessable in whatever ways human performances are assessable, which include prudential and moral assessment, and not only assessment of judgments and beliefs with a view to whether they amount to knowledge. Still it is this latter restriction that is distinctive of the theory of knowledge, of how knowledge is constituted and attained. This is the sort of theory that we now label “gnoseology” for convenience of thought and communication.

G. What is required for fully apt X success? 1. I have advanced several distinctions and theses for the case of knowledge (in theory of knowledge, in gnoseology). Confusion and miscommunication can come of ignoring these distinctions and theses and smudging together all ways of assessing performances in our intellectual domains. We can better appreciate how and why this is so if we extend the controversy to all domains of human performance: not only the intellectual, but also the domains of athletic, artistic, and professional performance. Take simple examples like those of competition tennis and chess. The AAA/SSS theoretical structure applies straightforwardly to these. Here again we can distinguish the apt and fully apt performances. A commentator at Wimbledon or at a world chess championship might effusively praise a particularly impressive move as brilliantly apt, where it would be obvious that the assessment is restricted to considerations relevant to the ends and competences proper to the domain. Commentators might possibly take note of something cruel in the performance of a participant, but this would hardly affect the tennis assessment of a brilliant ace (despite the cruelty you might find in that ace as a concluding demolition of a much weaker opponent who had won not a single point in that match). Cruelty is just irrelevant to the correct assessment of that blazing ace as a great serve. 2. In making that point we of course ascend also to whether the commentator’s assessment is itself correct and indeed apt. What anyhow are the proper aims of a TV

Responses

231

commentator in their running commentary? Even if we recognize that there are aims proper to the art of such commentary, which go beyond simply stating the plain truth, it remains plainly true that the winning serve was a wonderful, fully apt performance as a tennis shot, as is evident and known to every witness of it in the stands. That moreover still seems correct even for the mind and speech of anyone among the losing player’s family and close friends. It might be improper and even cruel for someone personally close to rub in the losing player’s disgrace by highlighting the brilliance of that concluding serve. But that still does not affect the aptness of the commentator’s belief, and even the aptness of their public statement, nor would it affect the tennis aptness of the private judgment by the close friend or relative, nor even the tennis aptness of their public statement, if they were cruel enough to make such a statement. Suppose that in making that affirmation you were trying to get it right on the tennis quality of the serve. Of course, you might have been trying to do any number of other things in making that affirmation, including embarrassing and insulting the loser. But if we focus simply on your alethic attempt to get it right on whether the blazing ace was a fine shot, then no doubt this attempt of yours was correct, aptly so, and even fully aptly so. To think otherwise is to fall prey to confusion. This is especially so if we just focus on the private thoughts of an onlooking assessor, but they seem true even of the pertinent public affirmations, so long as we focus on merely alethic attempts. Of course, the full aptness of an alethic attempt is compatible with the awful unwisdom of the public utterance involved in that attempt. The utterance might be unwise, cruel, inappropriate, wholly deplorable, despite the excellent and full aptness of the alethic attempt that the speaker makes through that utterance. Finally, even less would the cruelty affect the aptness and the full aptness of the serve itself, not if this aptness and full aptness are restricted to the tennis aptness and full aptness of the tennis player’s performance as they attempt to ace the opponent. This attempt clearly does attain its objective, and does so aptly and fully aptly. 3. In conclusion, I suggest that the domain of intellectual performances is one more domain of assessable human performance. Cruelty is neither here nor there as concerns our distinctively gnoseological assessment, our assessment of judgment and inquiry with respect to the proper aims distinctive of such performance. On the other hand, conscientiousness is very much involved in epistemic assessment, but it must be conscientiousness of the epistemic sort, which leaves aside any moral or prudential concerns. It must be the conscientiousness that avoids any epistemic negligence or recklessness in arriving at one’s judgment, such as the negligence or recklessness involved in rash judgment that omits sufficient inquiry. It is such gnoseological conscientiousness that is constitutive of human knowledge, and it alone is a form of conscientiousness in the charmed inner circle of the telic normativity of theory of knowledge, where the traditional issues of knowledge and skepticism have their home. This is just to state a historical fact. It is not to disparage an intellectual ethics with its own important place alongside other forms of applied

232

Ernest Sosa Encountering Chinese Philosophy

ethics, such as medical ethics and business ethics. Such disparagement could not possibly be further from my intent.

H. More on reliabilist vs. responsibilist virtue theory and the charmed inner circle 1. Xiang Huang’s contribution (Chapter 6) is a thorough and insightful discussion of the two supposed branches of virtue epistemology, along with an intriguing comparison with Wang Yangming’s meta-ethics. 2. An important part of my response to Huang is found in Section E, above. The problem with the view that separates the two supposed branches of virtue theory is that it misconceives the content of virtue reliabilism. My virtue reliabilism has never restricted itself to the so-called faculties that we share with the other animals. What it requires for knowledge, even at the level of apt belief, is that a competence be exercised, an adroitness, such that its exercise guides an attempt to get it right on a question whether p, such as a judgment or a judgmental belief. And this exercise of competence or adroitness must then manifest in the success of that attempt. Never has my view been restricted to basic faculties such as perception, memory, etc. The AAA/SSS structure requires a middle A of adroitness, but this adroitness would definitely include the higher rational and reflective competences that are the pride of humanity. And such rational, reflective competences would require the avoidance of any negligence or recklessness that would spoil performance, particularly at the reflective level, where risk of failure must be assessed with competent adroitness. Despite the disagreement I have noted, I admire Huang’s chapter and its deep understanding of the issues and of my own take on those issues.

I. On Buddhist detachment and its virtue-theoretic fit 1. Tao Jiang (Chapter 5) takes note of a distinction between two sides of virtue epistemology, the reliabilist and the responsibilist. As I have argued, this is a distinction too often invoked as a basis for a false dichotomy. On one side we find a reliabilist epistemology of competences such as those of perception, memory, and basic inference, basic faculties that we share with the lower animals. On the other side we are invited to consider the more sophisticated reflective competences distinctive of the animal that is rational. Here are found distinctive competences of professions such as medicine and law, which enable sophisticated diagnoses and adjudication, along with the intellectual skills and abilities distinctive of detectives, scientists, and aesthetic critics. Such sophisticated knowledge lies beyond the competence of even the most intelligent furry mammals. Any suggestion that reliabilist virtue epistemology is restricted to such animal knowledge has I hope been put to rest already. Is it only such basic faculties that deliver knowledge when manifest in the truth and the aptness of a judgment or belief? Clearly not. Humans excel intellectually with competence to discern truth reliably enough through the sorts of reflective access distinctive of the rational animal.

Responses

233

2. Granted, I do distinguish between the broadly epistemic character traits of special interest to responsibilists, such as open mindedness and intellectual courage, but I say that these have a place in epistemology mainly by putting us in a position to know the answers to questions into which we may inquire. Are they competences whose exercise is constitutive of knowledge? Should we welcome them under the umbrella of virtue reliabilism? Responsibilists emphasize that the traits of special interest to them involve not just the concern for apt accuracy dear to the virtue reliabilist. Mixed into responsibilist intellectual courage and open-mindedness are moral concerns. Admirably courageous or open-minded agents are supposed to embody within their pertinent traits a discernment that enables them to give proper weight to distinctive moral concerns that help guide inquiry. Whether you are properly courageous or properly open-minded is thus determined by the proper weight that you allot to whatever moral issues may be in play as you decide how to conduct your inquiry. Here again I am not at all tempted to deny that there are such virtues, such character traits essentially constituted by ability to (a) discern what is morally desirable in one’s inquiry, and (b) to give that its proper weight in how one inquires. What I do want to deny is that the exercise of such morally constituted character traits is constitutive of knowledge or of the sort of epistemic justification or competence whose exercise can constitute such knowledge or justification. So, I conclude that traditional epistemology, which is focally concerned with knowledge and skepticism, is best recognized as a distinctive category under the broad tent of epistemology (broadly conceived). How so? 3. Gnoseology is concerned with knowledge and how it is constituted. It is concerned with the sort of epistemic justification that can help constitute such knowledge. And this sort of justification is importantly disjoint from many obviously weighty values. It is unaffected by values not concerned with accuracy and its truth-relatedly competent attainment. Thus it disregards moral, aesthetic, and prudential values as irrelevant to its focal concern to understand knowledge and how this is constituted. Clarity is at a premium here, as the issues are subtly evasive. I certainly do not mean that those who aspire to know must disregard the moral or other practical or aesthetic values that may be relevant to inquiry or to any other intellectual action. Rather, I mean only that in assessing whether a belief amounts to knowledge, we must focus on the gnoseological competences aimed at reliable enough attainment of truth and aptness. 4. There is much of great interest in Jiang’s discussion of Buddhist detachment, but I must focus on the matters of direct relevance to my virtue epistemology, and how it is related to that Buddhist virtue. Here Jiang proposes that when Buddhists are able to “detach” they gain insight that had been blocked (by the corresponding attachment, presumably). This is remarkably similar to St. Paul’s metaphor of scales that fall from his eyes. Both cases presuppose that humans have an underlying competence to get it right, a competence that is obstructed, whether by “attachment” or by “scales.” Any ability that enables us to unblock such intellectual competence is then a trait that goes

234

Ernest Sosa Encountering Chinese Philosophy

beyond mere animal faculties such as our faculties of vision or hearing. And the exercise of such ability can then be constitutive of true knowledge; it’s after all through its exercise that we attain reliable access to the truth, an access that is otherwise denied to our benighted human fellows. We have just now stumbled on another familiar metaphor: that of the obscurity or darkness that blocks us from attaining even basic perceptual knowledge, where we need relief from the darkness that prevents them from giving us access to some distinctive body of truths. 5. There is a common element to both familiar metaphors, that of scales falling from one’s eyes, and that of darkness removed by enlightenment. In each case, a human intellectual faculty is blocked from proper functioning by something removable through some procedure: the removal of the scales, or the provision of light. Eventually professionals might even be tasked with the competent performance of such procedures. The competences exercised by these professionals might then be aided by distinctive instruments, such as scalpels or lanterns. Suppose you had been unable to see the large tree before you (because of scales or because of darkness). Now the truth is revealed through the operation of your visual faculty, now obstructed by neither scales nor darkness. You now form your belief that the tree is there before you, something you had only suspected. Now you see clearly and distinctly the outlines of that tree before you in good light. Here there is an exercise of a vision-aided faculty, of a disposition to get it right reliably when you try to get it right aided by the deliverances of your eyesight in proper conditions for its exercise. In the new condition (no scales, good light) your disposition to get it right on such a question is highly likely to succeed; it is a highly reliable disposition, competence, intellectual virtue. And when the accuracy of your judgment or belief thus manifests a reliable enough competence, your intellectual performance is an apt success, here an apt belief, and a sort of knowledge. 6. Consider again the removal of the scales, the provision of the light, or the Buddhist detachment. Are these constitutive of the knowledge that they make possible? Arguably they are only auxiliary. They put one in a position to exercise one’s corresponding competences unblocked by such disabling factors. So, the true visual competence requires good light and eyes unblocked by scales. It is the good view with unblocked eyes in good light that is constitutively involved in the competence manifest in one’s getting it right about the tree. And, arguably, that is how it would be with Buddhist detachment. Attachment (to a conventional ontological framework installed culturally) functions like scales that preclude proper access to deep Buddhist truth. Proper access requires a mind’s eye unblocked by bad conventional preconceptions. Detachment enables access to truth made plain only when we overcome the conventional blinkers or scales. 7. Granted, we can debate how much to include in the constitution of the knowledge attained through proper competence. This is because we can disagree on how to conceive of the competences whose exercise has a true belief for upshot. Do we include what was done to remove the scales, or to provide the light? Or do we declare these to

Responses

235

be just auxiliary to the acquisition of the truly constitutive competence, which is simply unscaled eyes operating in good light, etc. Similarly, it would be debatable whether Buddhist detachment is constitutive of the attainment of the deep truths of Buddhist metaphysics. Or should we rather say that it is only an auxiliary competence that enables us to attain the access to that truth now unimpeded by such conventional scales on our mind’s eye? 8. Nevertheless, all of that should be put to one side. It is on the other side that we find the responsibilist character traits such as open-mindedness and intellectual courage. This is because of the responsibilist insistence that these are morally constituted at least in essential part. They involve the ability to discern the important moral values that may be involved in the pertinent conduct of inquiry, for example, and the ability to give them their proper weight in determining how to guide one’s inquiry. Such responsibilist traits are thus morally tinged, or otherwise practically tinged. By contrast, we should recognize gnoseological varieties of such traits as open-mindedness and intellectual courage. These gnoseological varieties would govern proper truthdirected inquiry. And it is this gnoseological conscientiousness that could help constitute apt alethic affirmation, or apt judgment, and thereby knowledge.

J. Knowing to: knowing even more fully well? 1. Yong Huang (Chapter 12) understands and expounds the essentials of my view full well, and has my gratitude for the close and insightful reading. But he concludes that the view falls short and requires a further development in line with the moral philosophy of Wang Yangming. So, that will be our topic in this section. 2. Let’s first focus on the case of moral judgment, decision, and action. You come to believe that you ought to ø (on a certain occasion), and this belief of yours may be true, apt, and fully apt. Its correctness manifests your pertinent competence, and so does even its aptness. Its aptness is not just a matter of luck, as you bring it about that your judgment is apt, and you do so through your competence to assess your pertinent first-order moral competence and your skill to judge aptly given your shape and situation. 3. Suppose, however, that despite your knowledge that you ought then to ø, you still fail to ø. Should we then leave it open that you might still know full well that you ought to ø? Huang answers in the negative, drawing inspiration from Wang. And Chad Hansen also writes in a similar vein about ancient Chinese philosophy, as when he says that “know-to is prescriptive knowledge . . .. The knowledge is expressed in actions, since know-to is not merely intellectual but includes attitudinal ingredients: desires and aversions” (Hansen 1992: 253). Huang finds a similar view in John McDowell and Mark Platts, who are said to appeal to a “unitary attitude” of “besire.”8 They “consider belief and desire to be not two separate mental states but two mutually inseparable aspects of one single mental state.”

236

Ernest Sosa Encountering Chinese Philosophy

This broad set of issues has an extensive literature that goes back to ancient Greek philosophy, and it is fascinating, and in a way reassuring, to see its important presence in similar forms (even if somewhat differently) also in ancient Chinese philosophy, as well as in very recent contemporary analytic philosophy. 4. One important issue in this vicinity is that of weakness of will (akrasia). What is its nature and is it really possible? Can an agent really believe with full conviction that they ought to ø even when the time for action comes and they fail to ø, not because they are unable but only because they are unwilling? Although these issues are in the vicinity, we need not settle them here, or even enter deeply into them, as we will focus on an independent question. Suppose the agent believes that they ought to ø but lacks the “desire” to ø, even just in being unwilling to ø. Would it not speak better of them if their will aligned better with their knowledgeable conviction that they ought to ø? Yes, that seems hard to deny. What is not so obvious is that this is because of some failure on the side of their belief. But wait: might they not fail simply because their confidence falls short of the required full conviction? This might be so, but it is far less plausible that it must be so. If we reject that this must be so, however, then why should we think that the knowledge involved must fall short of the epistemic heights? Why should we think that it must fall short of knowledge full well, at least knowledge that is as good as it gets in epistemic respects? It is not so plausible that it must fall short in that way. 5. As for the idea of an attitude of besire, here’s how I would propose to accommodate that. Suppose an agent S believes that they ought to ø and also desires (or, better, fully intends) to ø, even when the moment of action comes. It might be thought that there is then automatically an attitude of so desiring (intending) while so believing. This is just a conjunctive state, like the conjunctive state of walking while chewing gum. But that is not what is meant, comes the rejoinder: “What we mean is a unitary state of besiring, which in its unity is very unlike any mere conjunction of walking and chewing gum.” And, yes, that is a fair point. An agent who merely conjoined the knowledge that they ought to ø with their willing (“desiring”) to ø would be objectionably disjointed. We want a tighter unity. How can we make room in our virtue theory for the possibility of such a desirable tighter unity? 6. What is the (proper) aim of willing (intending, “desiring”)? Some will say that it’s the (deontological) right, some that it’s the (consequentialist) good (or the best of the competing options all things considered), and some will say that it often is (properly) the prudential good of the agent. Let us say, more ecumenically and determinably, that it is the “to be done,” so that we can avoid taking sides unnecessarily, so that we can cover all forms of (proper) agency, including even what is to be done at a certain juncture in a given tennis or chess match and, even more, what is “to be done, all things considered.” Consider then willing to ø (or intending to ø, or aiming to ø, any of which we might label “desiring” for present purposes, where our theoretical objective is just to

Responses

237

fashion a template). In so “desiring,” the agent then aims to do the thing that is then “to be done.” Again, for our purposes of theorizing with schematic generality, we focus on the structural place reserved by that template, one that can be filled in the way of prudence, or of morality, or of whatever practical objectives might be to the fore. Suppose øing is indeed what you are to do on that occasion. And suppose you do then ø. How do we assess your action (or deed)? Well, if you do ø and øing was the thing for you to do, then you ø successfully, whether this takes a form of satisfying a moral contract, or taking the consequentially best of the available options, or acting for the best prudentially, etc. Your then øing might or might not manifest pertinent adroitness (or competence). And, plausibly, a large part if not all of that adroit competence will reside in your deliberation on what is to be done. This deliberation may proceed more or less competently, and is intended to yield a judgment on your part as to what you are then to do, which more specifically would be a judgment on whether you are to ø. If your success is not by luck but through competence, then you will rely in your attempt to ø on your corresponding judgment that you are to ø. And the adroitness of your øing will derive from the adroitness of your judgment, combined with the practical-rational adroitness involved in your willing and attempting (trying) in accord with your best judgment, and because of that judgment. If your attempt succeeds, so that you do then ø, and thereby do what was to be done, then the success of your attempt, your doing what was to be done, will be in accord with the adroitness of your attempt, so that its success will come about not just by luck but as a manifestation of the adroitness of your attempt on that occasion. Your attempt will thus be apt.9 7. I have here aimed to apply a virtue theoretic approach beyond episteme to praxis, while aiming for a view that can cover forms of concern that have been labeled “moral” by philosophers of practice, whether consequentialist, or deontological, etc., but also other forms, such as the prudential, the aesthetic, and I would also argue the ludic, this last in order to cover domains of human sport and play. In all of that rich diversity of human performance we find a virtue theoretic structure in a theory of attempts, with its distinctive AAA/SSS constitution. I am delighted to have learned how deeply ingrained in Chinese philosophy is the interest in theorizing about such phenomena. Even if the greats of the Chinese traditions do not agree on every point with the contemporary virtue-theoretic approaches, just as they do not agree on every point with each other, and just as there is similar diversity among contemporary theorists, it is still gratifying indeed to learn of the deep similarities of theoretical aims along with parallel ways and means. That is my response to the chapters that focus on the main issues of my virtue epistemology, ones that relate it closely to the history of Chinese philosophy. I am also grateful for comments that do not focus specifically on those central issues of my work, and so I’d like now to conclude by commenting also on those chapters.

238

Ernest Sosa Encountering Chinese Philosophy

Part II A. Skepticism Skepticism is at the center of Yiu-Ming Fung’s chapter (Chapter 10), whose scope is vast and variedly rich, to an extent that I could not possibly cover in these brief comments. So, I will restrict myself to one pointed and important question about “brains in a vat” (BIV) skepticism posed to me. When I present the problem of external world skepticism, I do so in part through the example of a recently envatted brain, one excised from a normal human knower. I say that this BIV thinker might then conceivably be “detached” from the world beyond (in a sense of normal, not Buddhist, detachment), but Fung dissents by suggesting that my BIV would just be “attached in some way different from the normal.” And I take the point, even if arguably it is conceivable (logically possible, maybe even metaphysically possible) that there be a possible world where such a BIV arises spontaneously with a sequence of uncaused events at its periphery indistinguishable from a sequence of events at the periphery of some brain normally housed in a normal skull of a normal live body. Even abstaining from such wild science fiction, I could opt for a less radical notion of detachment that amounts simply to “not normally attached, in accordance with human normalcy.” And I believe my reasoning could then proceed in terms of this more specific concept of detachment. I regret that I cannot go further into Professor Fung’s rich and informative chapter.

B. The rectification of names and joining virtue epistemology with Confucianism Yingjin Xu’s chapter (Chapter 9) is intricate in ways that I find interesting but hard to grasp. Let me give just one example that will perhaps suggest the sort of difficulty that I am having, although the detailed and full discussion contributed by Professor Xu offers hope that such difficulties could be bridged, given enough time and discussion in depth. Xu writes: [A] Confucian scholar would simply cast doubt on the JTB-statuses of the target beliefs themselves by asking a further question: under what kind of reason can a Gettier case-builder say that both target beliefs are justified?

The two “target beliefs” are those featured in two examples used by Edmund Gettier in his famous refutation of the “justified true belief ” (JTB) account of propositional knowledge. Case II in Xu’s presentation is a case that according to Gettier falls under the following pattern. It is a case where a thinker justifiedly believes that p, deduces from this that either p or q, and believes this conclusion on the basis of their competent deduction, so that (as is supposed by Gettier and by just about everyone who has

Responses

239

responded to Gettier’s brief article) the thinker is thereby also justified in believing that p or q. However, it turns out false that p, and it is true that p or q only because it is true that q. And so, the intuitive conclusion to which readers have been powerfully drawn is that the thinker in the example has a justified true belief (JTB) that is not knowledge. The thinker is supposed to have a justified true belief that p or q without knowing that p or q. Here now is Xu’s treatment of that Gettier case: Now let’s turn to the Confucian analysis of Case II. . . . [What] is involved here is a disjunctive proposition. Suppose that a Confucian scholar now barely accepts the deductive rule that a disjunctive proposition is true if and only if at least one of its disjuncts is true. Hence, she may deduce from this that a disjunctive proposition is justified if and only if at least one of its disjuncts is justified. But she will immediately figure out that neither of the disjuncts . . . is justified.

Here I am unable to follow Xu’s lead. I cannot believe that a disjunction is justified if and only if at least one of the disjuncts is justified. From this it would follow that I cannot be justified in believing an immense number of disjunctions that I take myself to be obviously justified in believing: namely, every disjunction of the form

that I in fact do believe, and am justified in believing despite having no clue as to which of the disjuncts is true. Nevertheless, Xu’s chapter strikes me as a very interesting discussion that draws extensively on his familiarity with works in the philosophical tradition of Confucius. And so I welcome it for its potential to join together virtue epistemology and elements of Confucianism in a way that seems well worth exploring.

C. Reflective knowledge In his lucid and engaging chapter, Leo Cheung (Chapter 8) proposes that a form of reflective knowledge accords better than does mine with certain views in Chinese philosophy. The proposal is that reflective knowledge can be present through apt belief that one’s first-order belief is true. I require something stronger for reflective knowledge: it must be apt belief that one’s first-order belief is apt, rather than just true. But I am not convinced that the new, weaker proposal can really help. Here is why.10 The proposal is that reflective knowledge derives simply from apt meta-belief that one’s belief that p is true, not from apt meta-belief that one’s belief that p is apt. But the new reflective knowledge then seems equivalent to animal knowledge, to first-order apt belief. Once one believes (on the first order) that p, then if one so much as raises (on the second order) the question of whether one believes correctly, one must surely believe that one does. If either the first order or the second order belief is apt, moreover, then so is the other: one will believe aptly that p if and only if one believes aptly that one correctly believes that p (modulo one’s simply posing the question whether one does believe correctly). Such reflective knowledge hence collapses into animal knowledge, but, regrettably, without adding much of epistemic significance. Despite this disagreement, I found much to agree with in Cheung’s chapter.

240

Ernest Sosa Encountering Chinese Philosophy

D. Buddhism and Pyrrhonism 1. Chinese philosophy is often said to be practical, whereas Western philosophy is said to be theoretical and speculative, but this overlooks the importance of philosophy as a way of life in Greek philosophy, not only for Stoics and Epicureans, but also for Pyrrhonian Skeptics. Wisdom was of course important to all lovers of wisdom. 2. Despite much concurrence with Gregor Paul (Chapter 4), we disagree on one point. I cannot believe that the discussion of skepticism is just a conceited and unimportant display of cleverness. For the Pyrrhonian skeptics, it is a vital matter of how life ought to be lived. Their form of skepticism leads to the contentment that they require for a happy life. Let us consider how they are led to that outcome. Let’s look at the stages of their trajectory, if only briefly and abstractly.

a. First come the modes, the argument forms that lead to equipollence, to a balance of reasons on both sides of any issue.

b. Such equipollence then leads to suspension of judgment on any question they may take up.

c. And suspension finally results in a freedom from disturbance, a blissful calm. 3. This is in sharp contrast with what Tao Jiang sees as the Buddhist route to detachment, or non-attachment. I gather that Buddhist detachment is based on a metaphysical commitment to a certain truth about the world, that of emptiness, according to Jiang’s exposition of the nature and basis of Buddhist detachment. Earlier (in Section I above) I discussed how his highlighted “detachment” might bear on virtue epistemology, in its two supposed branches, the reliabilist and the responsibilist. Now I’d like to comment briefly on another aspect of Jiang’s chapter (Chapter 5), one that may hold considerable historical and substantive interest. 4. There is a striking divergence between Buddhist detachment and Pyrrhonian equipollence and suspension. For, it is precisely the absence of any metaphysical or other cognitive commitment that supposedly frees one from disturbance and grants peace. The Pyrrhonists would not have found the Buddhist route easy to accept. By use of their famous argumentative modes, they would have tried to undermine any metaphysical confidence allegedly supportive or Buddhist detachment, such as any confidence about metaphysical emptiness. For the Pyrrhonists it is not confidence but suspension that secures calm detachment from disturbing cares. And on reflection this raises a question at the heart of Pyrrhonian skepticism and its way of life allegedly freed from disturbance. The question arises because frustration is just one variety of disturbance. Frustration does of course follow the disappointment of disbelieving what you favor, or believing what you oppose. On any given question what Pyrrhonian equipollence entails is the absence of belief on that question. And once you take away belief you take away disappointment, since disappointment is constituted essentially by belief (together with favoring or opposing).

Responses

241

However, the frustration of disappointment is not our only source of disturbance. We are also subject to anxiety, which is disturbing in its own way. Here I mean the anxiety that results if you intensely favor or oppose some outcome while you are in serious doubt on whether that outcome will come about. This makes it unclear just how the epistemic practice of Pyrrhonism is supposed to further freedom from disturbance in its way of life. If suspension is an essential component in disturbing anxiety, it is unclear how Pyrrhonian practice can really protect us from disturbance and grant us blissful calm.11 So, it is not clear that the Pyrrhonian route to blissful calm is superior to the Buddhist route.

Notes 1

I am especially impressed by the specific connections with Zhuangzi, with Xunzi, and with Wang, as these are detailed in several of the chapters. 2 Unfortunately, because the contrary misattribution derives from an outright misreading, it cannot be corrected by considering any rationale provided for it. Only a close look at the relevant texts would correct it. But here we cannot go into even the brief textual review to be found in chapter 2 of my Judgment and Agency (Sosa 2015). 3 Sections E to J will now concern topics found in most of the chapters, topics central to my views. My discussion will take up interrelated themes that overarch those chapters. These are themes concerning the contrast between responsibilist and reliabilist virtue epistemology, and concerning the relations between ethics and epistemology. My focus on these themes allows me to offer a response with a unity that I hope will make it more interesting than a catalogue of retail responses and will thereby excuse the apparent neglect of the points and ideas of my commentators that would otherwise require and receive more attention in my response. 4 And Gettier cases may now be seen as ones in which the epistemic agent falls short either of aptness or of full aptness. 5 I have discussed in detail the responsibilist critique of virtue reliabilism, with references, in my Judgment and Agency (Sosa 2015: chapter 2). Here I will be brief. 6 Here the reference is to the Accuracy-Adroitness-Aptness/Skill-Shape-Situation structure of virtue reliabilism. 7 Repeatedly, most recently, in Epistemology (Sosa 2017). 8 A term introduced by J. E. J. Altham in “The Legacy of Emotivism,” in Fact, Science, and Morality: Essays on A.J. Ayer’s Language, Truth, and Logic, ed. by Graham Macdonald and Crispin Wright (Malden, MA, and Oxford: Basil Blackwell, 1986). 9 In a fuller development of this line of thought, we have also considered the requirements for the agent’s øing to be fully apt. 10 Laid out more fully, it goes as follows. As soon as one is aware that one believes that p, one will surely believe that one’s belief is true. This may be seen as follows. Suppose one does believe that p. From

one can deduce first , which combines with to deductively entail , which is tantamount to . Given the infallible reliability of such deduction, as soon as one believes aptly that p, one can deduce an equally apt belief that one’s belief that p is a true belief. So, consider one’s second-order belief that one’s first order belief is true. The aptness of that second order

242

Ernest Sosa Encountering Chinese Philosophy

belief is trivially guaranteed by the aptness of the first order belief, given our infallible deduction. So, the aptness of a first-order belief guarantees the aptness of the second-order belief, if the second-order belief is merely the belief that the first-order belief is true. At least, that is so modulo one’s posing the question of whether one does believe correctly that p, and conditional on the also obvious fact that one does believe that p. So, the collapse is on the assumption of these two trivialities. Contrast one’s question whether one’s belief that p is apt, and not just true. There’s no way of deducing an answer to that second question trivially from some facts that are trivially accessible once one’s first-order belief that p is assumed to be true, nor even once that first-order belief that p is assumed to be apt. So, there’s a sort of collapse of reflective knowledge into animal knowledge on the Cheung version of reflective knowledge but not on my version. 11 I cannot tarry over this conundrum in greater detail; anyone interested might consult chapter 10 of my Judgment and Agency (Sosa 2015).

References Altham, J. E. J. 1986. “The Legacy of Emotivism.” In Graham Macdonald and Crispin Wright (eds), Fact, Science, and Morality: Essays on A.J. Ayer’s Language, Truth, and Logic. Malden, MA and Oxford: Basil Blackwell. Hansen, Chad 1992. A Daoist Theory of Chinese Thought: A Philosophical Interpretation. Oxford: Oxford University Press. Sosa, E. 2015. Judgment and Agency. Oxford: Oxford University Press. Sosa, Ernest 2017. Epistemology. Princeton and Oxford: Princeton University Press.

Index AAA conception of knowledge 3, 137, 139, 176, 177 AAA/SSS theoretical structure 4, 71, 227, 230, 232, 237 animal knowledge, as within the scope of 7, 142 archery, using to illustrate AAA structure 84 n.2, 135 Gettierized beliefs, identifying with AAA model 138 UKA doctrine, as a theory of 87, 95 virtue epistemology, as the root of 11, 101–2 Wang, in meta-ethics of 10, 88–92, 96–8, 100, 103 Alfano, Mark 71 Altham, James E. J. 214, 217 n.8 Ames, Roger 28, 51 animal knowledge 19 n.6, 33, 39, 129, 132, 192 in the AAA model 3–4, 7, 142–3 apt belief, as based on 40, 41, 88, 130 craftsmen, animal skills of 36 n.5 Gettier cases, applying to 146–7 human knowledge as continuous with 179, 188 necessary condition for 156–7 reflective belief, distinguishing from 8, 19 n. 5, 42, 131, 239 reflective knowledge, distinguishing from 5, 11, 14–15, 30, 119–23 in the virtue-epistemic structure 227–8, 232 aptness, accuracy, and adroitness. See AAA conception of knowledge archer analogy 26, 35 n.1, 97, 222 Analects, as applied in 21 expertise of the sage archer 30–1 Sosa, employing 2, 3, 12, 21, 40, 49 n.3, 84 n.2, 135, 225 Aristotle 2, 97, 176, 198 n.3, 207

character, on the virtue of 208, 216 n.1 intellectual and moral virtues, distinguishing between 72 in virtue epistemology model 1, 175 virtue ethics in Aristotelian tradition 29, 175, 207–8 Aspects of Confucianism (Paul) 51, 64 n.18 Baehr, Jason 10, 132 n.1 character-based virtues, describing 98, 100, 101 courageous logical reasoning for knowledge 62 n.6 Sosa, on the virtue epistemology of 64 n.16, 66 n.25 Battaly, Heather 64 nn.16–17, 102 Beck, Cline M. 212 Behuniak, James, Jr. 22 benevolence 7, 10, 35 n.1, 54, 58, 150 Berkeley, George 157 besire 215, 236 describing and defining 214, 217 n.6 knowing full well and 17–18 unitary attitude of 235 BIV concept 154 BIV skepticism, rethinking 164–7, 238 self-refuting skepticism of 159–64 Sosa epistemology in response to 53, 137 Blackburn, Simon 181, 197, 199 n.14, 200 n.20 blinkering concept 24, 28, 44–6, 234 brains in a vat. See BIV concept Brueckner, Antony 160, 162 Buddhism 63 n.14, 69, 138, 157, 195, 200 n.23 Chan Buddhism 2, 10, 70, 73, 74, 78, 79, 80, 81–2 detachment, deeming as virtuous 10, 12, 232–5

243

244 Mahāyāna Buddhism 70, 74, 76, 77, 78, 79, 80 Pyrrhonism and 224, 240–1 reliabilist virtue in 73–9, 83–4 Chan Wing-Tsit 89–90, 93–4, 99 fate, on the Confucian doctrine of 97 ko-wu, not accepting Wang version of 92 on “the pure intelligence and clear consciousness of the mind” 104 n.3 Cheng Yi 6 Cheung, Leo K. C. 14–15, 239, 242 n.10 Chong Kim-chong 8–9, 66 n.26 Confucianism 9, 22, 32, 181 Analects 5–8, 21, 34–5, 51, 138, 140–1, 191 Confucian sages 66 n.27, 189–90 knowledge, potential danger in 53, 58 moral virtues, on luck in development of 97, 104 n.5 neo-Confucianism 17, 200 n.21, 216 rectification of names 15–16, 136, 140, 143–6 six arts of Confucian philosophy 135, 139 virtue epistemology, joining with 136, 238–9 consequentialism 1, 2, 208, 236, 237 Copi, Irving M. 164, 213 Craig, Edward 178, 191 dào 61, 131, 176, 179 as the Confucian way 146, 181, 190–1, 215 Daoist skepticism 177, 187, 200 n. 20 epistemic agency on the path of 48–9 following the dào 43, 44, 51, 53 as the Great Ordering Principle 22, 28, 30 social dàos 178, 186, 188, 189, 190, 191, 194–6, 197 Xúnzˇı , reflections on 14–15, 47, 59, 123–7, 127–30, 132 of Zhuangzi 168–9, 171 n.4, 172 n.8, 198 n. 6 Dao: A Journal of Comparative Philosophy 5 Davidson, Donald 149, 156, 172 n.2, 192

Index conceptual scheme, working within 199 nn.12–13 ontological predicament of 150–5, 167 Unorganized Content of 157 Dell’Utri, Massimo 160–2 deontology 1, 51, 188, 193, 208, 236–7 Descartes, René 1, 7, 16, 185, 228 Cartesian demon 77, 137, 175 Davidson theory and 150–2 on judgment and free will 207–8 as a virtue reliabilist 103, 229 virtuous circularity in philosophy of 149, 162, 163 Doris, John 29 Epistemology (Sosa) 241 n.7 Explaining Graphs and Analyzing Characters (Xu) 146, 147 n.2 Fairweather, Abrol 71 Four-Sentence Teaching (sijujiao) 93 AAA structure as underlying 91–2, 96 meta-ethics of Wang, as basic structure of 88–9, 100 on moral competence 94 Fraser, Chris 9, 50 n.14, 226 Frege, Gottlob 136, 140, 145, 147 n.6 Fricker, Elizabeth 112, 115, 116 Fung Yiu-Ming 16, 172 n.7, 238 Garfield, Jay 75–6 Gettier, Edmund 55, 238 Gettier Problem 33, 70, 102, 138 animal knowledge, applying to 139, 146–7 Buddhists, no interest in 78 Confucian approach to 15, 136, 141–2 counterexamples to justified true belief model 32 deductive values, appealing to 140 virtue epistemology, as tackling 10, 53, 69 Xu, treatment of 15–16, 145, 238–9 Gibbard, Alan 194 gnoseology 34 gnoseological competences 12, 233 gnoseological conscientiousness 103–4, 230, 231, 235 intellectual ethics, comparing to 227

Index Gongsun Long 55 Great Pure Understanding (da qing ming) 129, 131, 132 Greco, John 71, 73 Grice, Herbert P. 110 Hansen, Chad 17, 212, 235 Hobbes, Thomas 178, 189 Huang, Xiang, 10–11, 227, 232 Huang, Yong, 17–18, 104, 221, 235–6 Hui Shi 26, 27–8, 28–9, 169 humaneness 21, 35 n.1, 53, 54, 58, 59 Humean model in contemporary philosophy 214 Husserl, Edmund 154, 157, 159 Hutton, Eric 5, 61 Inquiry on the Great Learning (Wang) 89 Instructions for Practical Living (Wang) 93 intellectual virtues 1, 2, 21, 29, 33, 56, 60, 120, 152 in the AAA structure 11 apt belief and 13, 121, 122, 123, 130 character virtues, comparing to 54–5 constitution of knowledge, place in 224 formal belief, link with 131–2 justification, relating to 119 knowledge as a function of 57 reliabilism, within the system of 29, 61 scales falling from the eyes metaphor 234 Sosa’s emphasis on 52–3 true belief as linked with 156 virtuous performance, role in 208 Jiang Tao 10, 12, 83, 233, 240 jing (ensnaring circumstances) 80 Johnston, Mark 149, 171 n.1 Judgment and Agency (Sosa) epistemic agency, describing 42 forming of a belief, focus on 229 functional and intentional teleology, distinguishing between 226 knowledge and action, featuring 209 responsibilist approach to virtue reliabilism in 64 n.16, 241 n.5 justification 56, 64 n. 14, 102, 150, 155 bypassing justificating, difficulty of 121 competence-centered explanation for 39

245

epistemic justification 140, 152, 156, 223, 226–7, 233 JTB concept 32, 70, 78, 138, 140, 238–9 justification of knowledge 59–60, 158 as the rectification of names 136, 140, 143, 146 secondary justification 70–1, 119 Sosa, foundational justification of 149 testimonial justification 13, 107–8, 109, 112, 116, 117 Xúnzˇı on justification requirement for knowing 43–4 justified true belief. See JTB concept under justification Kant, Immanuel 154, 155, 157–8, 159, 165, 171, 199 n.10 K-internalism (knowledge internalism) 129–30 Knoblock, John 23, 30–1, 108, 124–9 Knowing Full Well (Sosa) 203 knowing full well concept 129, 131, 210, 216 besire, Huang linking with 17–18 fully apt belief and 41, 123 knowing-to as better than knowing full well 213 in Platonic context 203, 209 “Knowledge and Intellectual Virtue” (Sosa) 119 Knowledge and The State of Nature (Craig) 178 “Knowledge in Action” (Sosa) 209 Knowledge in Perspective (Sosa) 56–7 Kvanvig, Jonathan 101 Lau, D. C. 23, 35 n.1, 36 n.10 Legge, James 6, 65 n.20, 167–70 Lewis, Mark 22 li (ritual) 126–7, 127–8, 178, 185, 190 Linji school Chan Buddhism, in tradition of 10 Linji lu on the virtue of detachment 70, 79–83 Linji on the true person 74 List, Christian 193 Liu Shu-shien 87, 94, 104 n.1

246 McDowell, John 104 n.6, 116, 118 n.6, 214, 235 Mencius (Meng-zi) 51, 53, 104 n.2, 215 on Confucian doctrine of fate 97 humaneness-like-archery analogy 35 n.1 on innate goodness 22, 31, 66 n.26, 89, 90 meta-aptness 4, 123, 204, 209 mind maps 180, 181, 199 n.14 Minsky, Marvin 159 Moore, G. E. 7, 137–8, 162–3, 228 Mozi and Mohist canon 178, 186–9, 189–91, 192, 194, 196 Nivison, David 27 nomic-transcendentalism 127–30, 132 Owen, Stephen 22 Paul, Gregor 9–10, 65 n.18, 240 perception 30, 33, 39, 41, 60, 159, 163 as an animal faculty 119–20, 228 as a basic faculty 2, 9, 11, 102, 228, 232 default justification, providing 108 Descartes, in system of 150 epistemic competence in 10 faculty of, developing 12, 13, 54 forgetting, driving out perception to achieve 32 Greek vs. Ancient Chinese perception 191–2 instrument and perception 116–17 non-distorted perception 9, 31 in Platonic thought 203–4 reality as a “thing-in-itself ” causing perception 167 in reliabilist epistemology of competences 232 Sellers, on foundational knowledge through 156 testimony as akin to 14, 107, 109 veridical perception 8, 22, 23–4, 25, 29, 31, 34, 152, 223 virtue, perception as 52–3, 56, 58, 61, 71, 100 Platonism 176, 185, 194, 195, 197 epistemic virtue in 177–8 Meno 176, 205

Index Platonic noumena 157, 158 Protagoras 17, 205–6 supernatural theology and 179 Theaetetus 203–4 See also Third Platonic Problem Platts, Mark 214, 235 Pritchard, Duncan 97, 98, 137 Putnam, Hilary 16, 159–60, 161–2 Pyrrhonism 224, 240–1 Quine, W. Van Orman 166, 175, 181, 199 nn.12–13 Rawls, John 178, 199 n.10 reflective knowledge 7, 8, 53, 98, 139, 157, 228 in AAA structure 88 aptness of belief and 3, 4, 92, 130, 239 as epistemically valuable 33 in A Journal of Comparative Philosophy 5 in the Lunyu 59 n.2, 65 n.20 meta-competence, as formed through exercising 204 new generalized definition of 130–2 See also animal knowledge Reflective Knowledge (Sosa) 16, 121 reliabilism. See virtue reliabilism ren (benevolence/humaneness) 10, 21, 35 n.1, 53, 55, 58, 61 Rescher, Nicholas 158 Rorty, Richard 156, 157 Ryle, Gilbert 17, 177, 198 n.4, 212, 214 Santayana, George 155, 157, 158, 171 seat, shape, situation. See SSS conception of competence Sellars, Wilfrid S. 102, 156, 192 Six Principles of Composing Chinese Characters (liùshū) 143–4 Smith, Michael 214, 217 n.6 Socrates 203–4, 205, 206 Sosa, Ernest archer analogy, employing 2, 3, 12, 21, 40, 49 n.3, 84 n.2, 135, 225 Chinese philosophy, encountering 5–8, 34, 65 n.18 on epistemic competence and agency 39, 40–3

Index foundationalism and coherentism, on the middle way between 155–9 on intellectual ethics 11–12, 210, 217 n.3, 227, 229, 231–2 intellectual virtues, emphasis on 52–3 “Knowledge as Action” 209 “Moral Relativism” 209 “The Raft and the Pyramid” 56, 69, 119 “Reflective Knowledge in the Best Circles” 120, 122 reformulation of theory on a high level 136–9 “Reliabilism and Intellectual Virtue” 121 “Telic Virtue Epistemology” 64 n. 16 testimony, account of 107–8 Third Platonic Problem for 203–7, 216 “Two Forms of Virtue Epistemology” 57, 58 as a virtue reliabilist 9, 21, 57, 71, 150–5 Zhuangzi and Xúnzˇı , comparisons with 28–31, 31–5 See also Judgment and Agency; virtue epistemology SSS conception of competence 4–5, 71, 227, 230, 232, 237 Sung, Winnie 13–14, 222 Tao Jiang 10, 232, 240 Tarski, Alfred 197 testimonial knowledge 13–14, 102, 107–8, 109, 116 Third Platonic Problem 209 first and second Platonic problems 203–5 Huang as dissecting 17, 104 n.4 knowledge in general, not applying to 207 Wang’s solution for 211, 213, 216 tian (nature/heaven) 32 traditional epistemology 57, 93, 175, 227 in the AAA structure 102–3 charmed inner circle of 21, 34, 73, 98, 103, 104, 227 as a distinctive category of epistemology 233 individual belief as basis of knowledge in 70–1

247

intellectualist knowing-that, dealing with 212 Turri, John 29 UKA (unity of knowledge and action) doctrine 93, 104 dualist model, contrasting with 94 putting into practice 96, 99 virtue epistemology as a theory of 87, 95, 98, 102, 103 virtue epistemology 69, 71, 93, 95, 129, 209, 238 in AAA framework 91, 92, 100, 101–2, 177 anti-luck virtue epistemology 98 apt performance, role in 88 Chan epistemology as a version of 70 Chinese philosophy, pairing with 175, 221, 239 competence as part of 73, 208 contemporary Anglophone version 135–6 detachment, as related to 74, 83, 233, 240 Gettierized cases, engaging 136–7, 147 knowing-that concept, tackling 178 normativity, as a strategy to naturalize 176 of Sosa 2–5, 13–18, 39, 87, 175–6, 207 telic virtue epistemology 2, 64 n. 16, 103, 104, 225–6 virtue ethics, overlap with 51 virtue perspectivism and 119, 207 See also virtue reliabilism A Virtue Epistemology (Sosa) 121 virtue ethics 51, 178, 198 n.10 in Aristotelian tradition 29, 175, 207–8 naturalistic focus of 177, 188 situationist criticism of 5, 210 virtue epistemology, as parallel to 1–2 virtue perspectivism 119, 121, 158, 163 virtue reliabilism 233 in the AAA structure 100, 241 n.6 animal knowledge, not restricted to 157, 232 responsibilism vs. 8–13, 39–40, 98, 222, 224, 226–9, 232, 241 n.5

248 Sosa as a virtue reliabilist 9, 21, 57, 71, 150–5 Zagzebski critique of 72–3 Walsh, Sean Drysdale 97, 104 n.5 Wang Yangming 87, 104 n.4, 217, 232, 235 AAA structure in meta-ethics of 88–92, 97–9, 100, 103 better than full well, how to know 211–16 innate knowledge of the good, using analogy of 10–11 as a neo-Confucionist 2, 17, 203 unity of knowledge and action doctrine 93–6, 99, 103, 104 Watson, Burton 25, 27, 28, 36 n.4, 36 n.9 wei-yi (being one) concept 168–9, 172 nn. 9–10 Wittgenstein, Ludwig 34, 35 n.2, 136, 192 Wright, Dale 75 Xie Liangzuo 6 Xing Bing 6 Xu Ai 93, 94, 99 Xúnzˇı 51, 52, 56, 66, 181, 192, 195 character virtues, emphasis on 53–4 dào, reflecting on the knowledge of 14–15, 30 epistemic agency in writings of 39–40, 43–8, 48–9, 57 heart-mind connection, commenting on 24–5 K-internalism with respect to the dào 129–30 on knowledge 6–7, 31–2, 55, 58, 60, 65 n.20 know-to and know-how, naturalistic concept of 191–2 nomic-transcendentalist understanding 128, 130, 132

Index reflective knowledge of the dào 131–2 on testimonial knowledge 13–14, 59 n.3, 108–9, 116, 117 unification of nature and heaven via the dào 123–7 verdical perspective 22, 29, 34 virtue reliabilism in 9, 10, 61–2 Zhuangzi and 26, 28, 36 n.3 Xu Shen 143, 144, 145 Xu Yingjin 15–16, 147 n.2, 147 n.6, 238–9 Yan Yuan 34–5 Yong Huang 17, 104 n.4, 221, 235 Zagzebski, Linda 49 n.2, 72–3, 178 Zhuangzi 16, 24, 33, 159, 200 n.20 absolute truth, on the non-existence of 170 cognitive capacity, denying sole authority of 25–6 conceptual skill of Zhuangzi students 181 Confucian principles, criticizing 9, 32 dispositional character, on its role in cognition 8, 28–9 dream and waking state, distinguishing between 55 happy fish assertion 27–8 on the hundun concept of the natural 167–9 knowledge, on the notion of 9, 31–2, 131 nothingness, recognition of 149, 166–7 open-mindedness, call for 22 on the pre-established heart-mind 34 on the social dào 194 true person/true knowledge manifesto 73 Zhu Xi 19 n.3, 94, 95