Xunzi (荀子) and Virtue Epistemology Cheng-hung Tsai Department of Philosophy, Soochow University chtsai@scu.edu.tw Abstract: Regulative virtue epistemology is the view that the possession of intellectual virtues regulates, guides, and enhances one’s epistemic practices, and that such intellectual virtues are something that can be cultivated to a higher degree. The question is, what kind of intellectual virtues, faculty virtues (such as sight and hearing) or character virtues (such as intellectual courage and open-mindedness), can be a candidate? Most assume that it cannot be the former. However, this paper shows that there can be a regulative faculty-based virtue epistemology, which takes cognitive faculties as intellectual virtues. I do not intend to establish such a version of virtue epistemology from scratch. Instead, I suggest that this form of virtue epistemology can be constructed from the philosophical works of Xunzi 荀子, one of the founders of Confucianism. Keywords: Intellectual Virtues, Sosa, Xunzi, Xin (the mind-heart), Perceptual Knowledge 1. Introduction Virtue epistemology generally aims to explain our epistemic practices through the notion of intellectual virtues. However, there is no general agreement among virtue epistemologists about the nature of intellectual virtues. Some virtue epistemologists (e.g., Sosa 1991; Greco 2000) consider intellectual virtues as cognitive faculties, such as perception, memory, introspection, and reasoning, whereas others (e.g., Montmarquet 1993; Zagzebski 1996) consider intellectual virtues as an epistemic agent’s character traits, such as intellectual courage, honesty, humility, and open-mindedness. Regardless of the different understandings of intellectual virtues, both camps of virtue epistemology share a common trait in the primary aim of epistemology. According to Robert Roberts and Jay Wood (2007), there are two kinds of epistemology according to how the aim of inquiry is established, namely, “analytic” and “regulative”.1 Analytic epistemology “aims to produce theories of knowledge, rationality, warrant, justification and so forth, and proceeds by attempting to define these terms” (R&W 2007: 20; emphasis added). On the contrary, regulative epistemology, which “does not aim to produce a theory of knowledge”,2 is an 1 The distinction was originally made by Wolterstorff (1996). To say that regulative epistemology does not produce a theory of knowledge is not to say that it provides no analysis of the concept of knowledge. Regulative epistemology rejects a “theory of knowledge” that is understood as a theory that aims to give an “e-definition” of knowledge, that is, to 2 epistemology that “tries to generate guidance for epistemic practice, ‘how we ought to conduct our understandings, what we ought to do by way of forming beliefs’ ” and offers “a response to perceived deficiencies in people’s epistemic conduct” (R&W 2007: 21). According to Roberts and Wood, all contemporary epistemologies, including virtue epistemology, are analytic epistemologies.3 As such, virtue epistemology aims to define knowledge (that is, specifying the individually necessary and jointly sufficient conditions for knowledge) in terms of intellectual virtues. Roberts and Wood regard analytic epistemology as problematic;4 thus, analytic virtue epistemology is also problematic. Instead, they offer a kind of virtue epistemology that takes regulative epistemology as its meta-epistemology. For regulative virtue epistemology, the possession of intellectual virtues regulates and enhances one’s epistemic practices. The question is, what kind of intellectual virtues, faculty virtues or character virtues, can be a candidate? Roberts and Wood choose character virtues. They study intellectual virtues such as love of knowledge, firmness, courage, humility, autonomy, generosity, and practical wisdom. These virtues are acquired excellences and can be cultivated to a higher degree. Faculty virtues (such as sight and hearing) are intellectual excellences because they reliably lead one to the truth; however, these virtues are not acquired but natural or inborn. If a philosopher attempts to offer a regulative virtue epistemology that yields characterizations of various intellectual virtues, to persuade ordinary people to cultivate such intellectual virtues so as to regulate and enhance their intellectual practices and lives, and to evaluate (i.e., to praise or blame) one’s intellectual practices according to the intellectual virtues, it is natural for the philosopher to choose character virtues as his focus of inquiry because, after all, it is character virtues rather than faculty virtues that can be cultivated by agents and for which agents are specify individually necessary and jointly sufficient conditions for knowledge (cf. R&W 2007: 9). But regulative epistemology does not reject an analysis of the concept of knowledge that does not aim at an e-definition of knowledge. Moreover, an analysis or non-e-definition of the concept of knowledge is required for regulative epistemologists to generate guidance for epistemic practices. Here a non-e-definition of knowledge is used as an “expedient of regulation” (R&W 2007: 27): “If we think of a definition not as a single formula that captures without remainder the essential characteristics of every instance of some kind, but rather as an expedient for making a concept more ‘definite’ for some person or group of persons, then we too are offering ‘definitions’ of various concepts” (R&W 2007: 20). 3 Some might question whether all contemporary epistemologies are analytic epistemologies. See, e.g., Turri (2012). 4 According to Roberts and Wood, “the reason why simple definitions fail is the complexity and diversity within the concept of knowledge. The concept may be held together by a set of overlapping resemblances between kinds of cases, as Wittgenstein argued that the concept of game is, rather than by a single set of properties that are both individually necessary and jointly sufficient for any case to belong to the class” (R&W 2007: 19). In this paper, I shall not examine Roberts and Wood’s criticism of analytic epistemology and focus instead on their classification of epistemologies. 2 responsible. Now we have three distinctive forms of virtue epistemology based on their different primary aims of epistemology and intellectual virtues: analytic faculty-based virtue epistemology, analytic character-based virtue epistemology, and regulative character-based virtue epistemology. However, a possible form remains, which is regulative faculty-based virtue epistemology. Let me sketch the four possible forms of virtue epistemology as follows: (“VE” is used as an abbreviation for “Virtue Epistemology) Concerning Intellectual Virtue Concerning the Aim of Epistemology Faculty-Based Character-Based Analytic Analytic Faculty-Based VE (Sosa) Analytic Character-Based VE (Zagzebski) Regulative Regulative Faculty-Based VE (?) Regulative Character-Based VE (Roberts &Wood) Table 1: Taxonomy of Virtue Epistemologies The pioneers or representatives of analytic faculty-based virtue epistemology, analytic character-based virtue epistemology, and regulative character-based virtue epistemology are Ernest Sosa (2007, 2009b, 2011), Linda Zagzebski (1996, 2009), and Roberts & Wood (2007), respectively. Who is the representative of regulative faculty-based virtue epistemology? It seems difficult for Western philosophers to conceive of this kind of virtue epistemology. The question with which I am concerned in this paper is how to conceived of a faculty-based virtue epistemology as a regulative one; that is, if a regulative virtue epistemology is possible, or if there can be a virtue epistemology that takes cognitive faculties as intellectual virtues and treats such faculty virtues as something that can be cultivated to a higher degree, what is the nature of such faculty virtues? I address this question in this paper. I do not intend to establish such a form of virtue epistemology from scratch. Instead, I suggest that this form of virtue epistemology can be constructed from the philosophical works of Xunzi 荀子 (c. 310-219 BCE), one of the founders of Confucianism. I show that the key to constructing such epistemology lies in Xunzi’s commitment to a both natural and cultivated faculty of perception due to his understanding of the exercise of the faculty of perception as the co-exercise of xin (the mind-heart) and the sense organs. The aim of this paper is to establish a version of regulative faculty-based virtue epistemology, and I will achieve this aim by exploring and interpreting Xunzi’s epistemology. I hope that this work presents a cross-cultural as well as a theoretical interest. The structure of this paper is as follows: In Section 3, I will explain why, or in what sense, Xunzi’s 3 epistemology is a virtue (§3.1), faculty-based (§3.2), and regulative epistemology (§3.3). Prior to Section 3, in Section 2, I will introduce Sosa’s version of virtue epistemology, the paradigm of analytic faculty-based virtue epistemology. With this paradigm, we can understand in what sense Xunzi’s epistemology can be classified as a faculty-based virtue epistemology, what distinctive features it possesses, and what contribution it may make to contemporary virtue epistemology. 2. Sosa’s Faculty-Based Virtue Epistemology 2.1 The Core According to Sosa, the core ideas of his virtue epistemology are 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 bi-level virtue epistemology distinguishes “animal knowledge” and “reflective knowledge”. Animal knowledge is apt belief, and reflective knowledge is apt belief aptly noted. It is evident that understanding the notion of aptness is crucial for understanding Sosa’s notions of animal knowledge and reflective knowledge (and then how he uses them to deal with several central issues in contemporary epistemology, such as the debates between foundationalism and coherentism, between internalism and externalism, and the value problem). The notion of aptness can be best understood by Sosa’s general account of performance normativity in which the aptness of a belief is treated as a special case of the aptness of a performance. Sosa’s account of performance normativity has five key concepts: full aptness, meta-aptness, aptness, adroitness, and accuracy. The first concept is defined by the second and third concepts together, and the second and third concepts are defined by the fourth and fifth concepts together at different levels. We proceed from the last to the first concepts. In assessing a performance, the last three concepts (accuracy, adroitness, and aptness) constitute what Sosa calls the “AAA structure” of a performance, which can be illustrated by the example of archery: When an archer takes aim and shoots, that shot is assessable in three respects. [...] First, we can assess whether it succeeds in its aim, in hitting the target. Although we can also assess how accurate a shot it is, how close to the bull’s-eye, we here put degrees aside, in favor of the on/off question: whether it hits the target or not. [...] Second, we can assess whether it is adroit, whether it manifests skill on the part of the archer. Skill too comes in degrees, but here again we focus on the on/off question: 4 whether it manifests relevant skill or not, whether it is or is not adroit. [...] A shot can be both accurate and adroit, however, without being a success creditable to its author. 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. [...] An archer’s shot is thus a performance that can have the AAA structure: accuracy, adroitness, aptness. (Sosa 2007: 22) A performance with an aim can be assessed by whether it succeeds in its aim (i.e., whether it is accurate or successful), whether it manifests relevant competence (i.e., whether it is adroit or competent), and whether it is accurate because of its adroitness (i.e., whether it is apt). The concept of “aptness” is epistemic in the sense that it is used to deal with Gettier-type cases in particular (i.e., to exclude as cases of knowledge those instances in which a belief is true because of epistemic luck) and in general (i.e., to exclude as cases of [better] performance those times when a skilled operation is successful because of luck). A performance can be meta-apt. A skilled operation might not be performed at a particular time t because the agent in question decides not to perform it then. The agent might make this decision because he knows that the operation will not succeed at t for some reason. Neither aptness nor inaptness applies in this case because (ground-level) success is not an objective. However, this conclusion does not mean that no performance occurs. According to Sosa, the agent’s forbearing is a kind of performance with an aim, i.e., avoiding ground-level failure (Sosa 2009a: 11). This performance can also be assessed through the AAA structure. The performance of forbearing is meta-accurate if and only if it succeeds in avoiding ground-level failure; the performance of forbearing is meta-adroit if and only if it manifests the agent’s meta-competence in risk-assessment; and finally, the performance of forbearing is meta-apt if and only if it is meta-accurate because meta-adroit. A performance of forbearing might be meta-apt. However, this definition does not mean that all performances at the meta-level must entail forbearing, which is negative in the sense that it does not suggest the execution of ground-level skilled operations. For Sosa, “[t]he forbearing might be meta-apt ... in being a proper response to the perceived level of risk.... Sometimes an agent responds properly by performing on the ground level, in which case that positive performance is meta-apt” (Sosa 2009a: 12). Suppose an agent’s performance at the meta-level aims to avoid the failure of his ground-level performance. Two possible cases follow. First, if the agent perceives that the rate of failure is too high, then he might respond by forbearing. Second, if the agent perceives that the rate of failure is low enough that the success of the ground-level performance can be secured, then he might respond by executing his ability. In the second case, we see that the execution of the ground-level ability can be connected to the meta-level perspective on that execution. Here, I introduce the fifth and 5 final concept in Sosa’s account of performance normativity, full aptness: A performance attains thus a special status when it is apt at the ground level and also its aptness is explained through competent risk assessment. Suppose this risk-assessment issues in the performer’s knowing that his situation (constitutional and circumstantial) is favorable (where the risk of failure is low enough) for issuing such a performance. If these conditions all obtain, then the performance’s aptness might stem from its meta-aptness; that is to say, its aptness might be relevantly explicable through the performer’s meta-knowledge that his first-order performance is likely enough to succeed and be apt. [...] This applies to performances such as a shot that hits its prey. That shot is superior, more admirable and creditable, if it is not only apt, but also meta-apt, and, further, fully apt: that is, apt because meta-apt. (Sosa 2009a: 13) A performance at the ground level is fully apt if and only if it is apt and its aptness stems from its meta-aptness. In Sosa’s model of full aptness, an agent’s meta-knowledge about his situation can contribute to the manifestation of his ground-level competence; that is, such meta-knowledge can make the manifestation or performance of the competence fully apt. For Sosa, fully apt performances are better or more valuable than either inapt or merely apt performances (Sosa 2009a: 13-4). As Sosa suggests, once we regard a belief as a performance, the belief has the AAA structure, or it can be assessed from three aspects. So, what is the particular AAA structure of a belief? What is the aim of a belief? What is the adroitness of a belief? For Sosa, “[w]e can distinguish between a belief’s accuracy, i.e., its truth; its adroitness, i.e., its manifesting epistemic virtue or competence; and its aptness, i.e., its being true because competent” (Sosa 2007: 23). As a belief, no matter in which order it is located, it has the AAA structure. Thus, an agent might have a first-order apt belief about a certain issue and a corresponding (second-order) apt belief about whether she should form or withhold the first-order belief. When the aptness of the former stems from the aptness of the latter, the first-order belief is not only apt but fully apt—it is not only animal knowledge but reflective knowledge.5 2.2 Intellectual Faculty Virtues Epistemic competences or faculties play a crucial role in Sosa’s epistemology. The core of Sosa’s epistemology is the distinction between animal knowledge and reflective knowledge; both kinds of knowledge are defined in terms of aptness, and aptness requires a manifestation of a competence (cf. Sosa 2007: 29). Sosa has provided a very detailed explanation of an intellectual virtue (Sosa 1991, 5 For a critical discussion of Sosa’s account, see Pritchard (2009), where he argues that apt belief is neither necessary nor sufficient for knowledge. 6 Ch.16).6 However, I ask readers to note that the instances of intellectual virtue in Sosa’s epistemology include perception (sight, hearing, smell, taste, and touch), introspection, memory, and reasoning (deductive, inductive, and intuitive)7 and to note further that the following discussion of intellectual virtues is limited to perceptual faculties, such as faculties of sight, hearing, smell, taste, and touch. Faculties of perception, memory, and reasoning are all natural faculties; that is, they are inborn rather than acquired. However, these faculties seem to have something different. It seems possible for an agent to cultivate his natural faculties of memory and reasoning (deductive and inductive) by training or learning some tricks.8 However, it seems relatively difficult for one to cultivate his natural faculties of sight, hearing, smell, taste, or touch. Focusing on the intellectual faculty of perception sharpens the contrast between the faculty-based and the character-based virtue epistemologies. But what is it that makes intellectual faculty of perception to be classified as something that cannot be cultivated? Inquiring into the acquisition of faculties is not helpful because some natural faculties can be cultivated. We can find a possible answer by understanding why Sosa calls an epistemic “faculty” an epistemic “virtue”: 6 Here is Sosa’s definition: “One has an intellectual virtue or faculty relative to an environment E if and only if one has an inner nature I in virtue of which one would mostly attain the truth and avoid error in a certain field of propositions F, when in certain conditions C” (1991: 284). 7 These cognitive faculties, according to Sosa, can be classifies into two broad sorts: “transmission” faculties and “generation” faculties. The former “lead[s] to beliefs from beliefs already formed”, while the later “lead[s] to beliefs but not from beliefs” (Sosa 1985: 225). For Sosa, intuition, perception, and introspection are generation faculties; memory and reasoning (deductive, inductive, and explanatory) are transmission faculties. In Sosa’s bi-level virtue epistemology, all these faculties are first-order faculties, distinguished from the second-order faculty, i.e., the faculty of reflection. 8 Jason Baehr distinguishes cognitive faculties (faculty virtues) from intellectual virtues (character virtues) in three ways, two of which concern us here. First, cognitive faculties are natural endowment, while character virtues are cultivated traits. Second, the “operation of cognitive faculties does not typically require an exercise of agency” (Baehr 2011: 23), while an “exercise of intellectual character virtues ... does characteristically involve agency” (Baehr 2011: 24). However, the two differences are not totally parallel. The two more general points about human qualities or dispositions that I want to emphasize are: (a) To say that a quality or disposition is cultivated implies that the quality or disposition characteristically involves agency because the cultivation of the quality or disposition requires an agent’s exercise of active efforts and active control, directly or indirectly, over the exercise of the quality or disposition, i.e., over the formulation of a set of heterogeneous trigger-manifestation pairs. (b) To say that a quality or disposition is natural, however, does not imply that the quality or disposition must have nothing characteristically to do with agency. Baehr would acknowledge my point (b) since he adds a footnote to the second difference mentioned above: “The [natural] faculty of reason [such as deductive and inductive reasoning] may seem to be an exception here, since its operation is commonly tied to an exercise of the will” (Baehr 2011: 23). See also Zagzebski’s Virtues of the Mind, where she argues that “the distinction between natural and acquired is somewhat vague since even natural qualities can often improve with training and practice” (1996: 103). 7 [P]erceptual ... beliefs are often acquired willy-nilly. And yet even where deliberate choice is thus absent, some mechanism may yet generate one’s belief. 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 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) I do not wish to quarrel about whether the term “virtue” can be applied to faculty. The point here is that unlike character virtues, the faculty virtues of perception do not involve agency such as an agent’s deliberate choice or voluntary control over belief-forming mechanisms. When his faculty virtues or belief-generating mechanisms are triggered to exercise, an agent forms perceptual beliefs willy-nilly.9 The inquiry into the exercise of the faculty or mechanism of perception suggests that if the faculty of perception has components, it has no component whose exercise characteristically involves agency (or, the will). Let me summarize three points established in this section. First, virtue epistemology is distinctive because in the study of knowledge, it greatly emphasizes epistemic sources. Second, the intellectual virtues can be understood as faculty virtues or belief-generating mechanisms. Third, the faculty of perception has no component whose exercise characteristically involves agency. The third point makes Sosa’s virtue epistemology difficult to be a regulative faculty-based virtue epistemology. In the following sections, I will show that Xunzi would agree with the first two points (which makes him a virtue epistemologist) but not with the third. 3. Xunzi’s Faculty-Based Virtue Epistemology 3.1 Early Chinese Epistemology and Virtue Epistemology In what sense can we treat Xunzi’s epistemology as virtue epistemology? In his entry on “Xunzi” in the Stanford Encyclopedia of Philosophy, Dan Robins (2007) explains a general feature of early Chinese epistemology that Xunzi’s epistemology shares: Early Chinese philosophers usually thought of knowledge in practical terms. They took it to consist in the mastery not of facts but of ways of acting (dao). Especially important was the knowledge of how to draw distinctions. Drawing distinctions was the 9 Here and hereafter, “faculty virtues” refer to “belief-generating faculties” rather than “transmission faculties”, unless otherwise noted. 8 closest analog to conceptualization recognized by early Chinese philosophers, and they took it to be the fundamental cognitive operation. ... This takes knowledge to be a kind of ability rather than a sort of representation of facts, and it should come as no surprise that Xunzi did not explain cognitive errors by appealing to mistakes of representation. For Xunzi, we make mistakes not because we picture [or represent] the facts incorrectly but because we lack some ability; knowledge contrasts not with false belief but with confusion. Xunzi twice (in Books 6 and 21 of the Xunzi) provides lists of his philosophical opponents and diagnoses their errors, and in neither case does he accuse them of misrepresenting the facts, or of confusing appearance with reality. Instead, he charges that they placed too much emphasis on some part of the Way, and thus failed to understand the whole. (Robins 2007; emphasis mine) Robins points out, though implicitly, a distinctive feature of early Chinese epistemology that aid the interpretation of Xunzi’s thought as virtue epistemology. I formulate the feature as a conjunction of the following two claims: (1) Early Chinese epistemology merely focuses on knowing-how, whereas contemporary Western epistemology merely focuses on knowing-that. (2) The kind of knowing-how with which early Chinese epistemology is concerned is knowing how to draw distinctions or knowing how to differentiate/discriminate,10 whereas Western epistemology is concerned with knowing that such and such is the case or knowing that a proposition p represents the facts.11 I agree with Robins’ general observation but offer something new regarding the first claim. When philosophers discuss the distinction between knowing-how and knowing-that, they tend to understand it as a distinction between knowledge of how to do something and knowledge that such and such is the case, and they wonder whether the former is a species of the latter (cf. Stanley & Williamson 2001; see Author 2011a and b for discussion). However, there is a possible and plausible understanding of the distinction in the field of epistemology; that is, the distinction concerns the distinction between epistemic competence (or sources) 10 There might be other kinds of intellectual know-how, such as knowing how to ask good questions. But such kind of know-how is not what early Chinese epistemologists are interested in. 11 Chris Fraser also claims that “early Chinese thinkers understand mind and knowledge mainly in terms of competence or ability, not representation. For them, the major function of the xin 心, or ‘heart’, is to guide action by discriminating different kinds (lei 類) of things, thus triggering skilled responses to them” (Fraser 2011: 128). An interesting side-issue here is that, based on the above claim, Fraser argues that skepticism about the external world (supported by the argument from illusion) cannot get off the ground in Chinese philosophy because Chinese epistemologists focus on competence or ability rather than on representation which is required for constructing the argument from illusion. With regard to the issue whether skepticism cannot get off the ground in Chinese philosophy, see Author 2006 for discussion. 9 and epistemic performance (or products), between an agent’s ability to know and what is known by an agent. So construed, it would be misguided to ask whether epistemic know-how is a species of know-that for they are two aspects of knowledge. It would also be wrong to derive, from the distinction, the view that early Chinese epistemology does not have the concept of propositional knowledge just because it focuses on the concept of knowledge-how. This understanding sheds light on the true difference between the two epistemologies; that is, contemporary Western epistemology (virtue epistemology excluded) is concerned with belief or knowledge per se, whereas early Chinese epistemology is concerned with cognitive abilities that generate belief or knowledge. Some scholars have indicated that the salient feature of virtue epistemology in general is the change in the direction of analysis (cf. Axtell 2000, xiii): virtue epistemology primarily focuses on the properties of epistemic agents or sources (properties such as character traits or reliability), whereas non-virtue epistemology merely focuses on the properties of belief or knowledge, which are the states or products of an agent or epistemic sources. With regard to the primary object of inquiry, early Chinese epistemology, including Xunzi’s epistemology, is close to virtue epistemology in its broadest sense.12 3.2 The Faulty of Perception as a Combination of the Sense Organs and Xin More interesting than merely treating Xunzi as a virtue epistemologist are the questions of what kind of virtue epistemology Xunzi implicitly proposes, and more important, why and how Xunzi’s version is distinctive among contemporary virtue epistemology. We pursue these questions from the idea of intellectual faculty in Xunzi’s work. The instances of intellectual faculty in Xunzi’s thought include the five sense organs and xin (the mind-heart). Xin itself is an organ,13 but it has a supervisory role over the five sense organs: The eye, ear, nose, mouth, and body each have the capacity to provide sense contact, 12 In their entry on “Virtue Epistemology” in the Oxford Bibliographies Online, Turri and Sosa (2010) treat historical figures such as Descartes, Hume, Reid, Peirce, and Russell as precursors of contemporary virtue epistemology. Hume is included because he thinks that “‘natural instincts’ or innate mental ‘mechanical tendencies’ enable us to gain knowledge beyond the ‘narrow sphere of our memory and senses.’ ”(Turri and Sosa 2010). Additionally, Reid is included because he thinks that “our knowledge derives from the exercise of our reliable intellectual powers and other dispositions that form part of our natural constitution” (Turri and Sosa 2010). Besides the historical features, contemporary philosopher John McDowell is included because he “explains central epistemological concepts—explicitly empirical knowledge and justification—in terms of the ‘exercise’ of ‘capacities’ ” (Turri and Sosa 2010). I believe that Xunzi can be regarded as a precursor of virtue epistemology in part for the same reason, although the point requires further elaboration. 13 In Xunzi, xin is treated as an organ or at least hard to be distinguished from the sense organs. See Geaney (2002: 97), Lee (2004: 33-4), and Hagen (2007: 160). 10 but their capacities are not interchangeable—these are termed “the faculties given us by nature”. Xin that dwells within the central cavity is used to control the five faculties—it is called “the lord provided by nature”. (Xunzi, “Tianlun”; tr. Knoblock, 17.3a) 耳、目、鼻、口、形，能各有接而不相能也，夫是之謂天官。心居中虛，以治五官，夫是之謂 天君。（ 〈天論〉） This passage contains two important claims: First, the five senses or perceptual faculties are not interchangeable. Second, the five senses are governed by xin. Let us start with the first claim and ask: What are the five senses (or the specific functions of the five sense organs)? According to Xunzi, The eye differentiates white from black, the beautiful from the ugly. The ear differentiates sounds and tones as to their shrillness or sonority. The mouth differentiates the sour and salty, the sweet and bitter. The nose differentiates perfumes and fragrances, rancid and fetid odors. The bones, flesh, and skin-lines differentiate hot and cold, pain and itching. These ... are part of the nature that man is born possessing, that he does not have to acquire. (Xunzi, “Rongru”; tr. Knoblock, 4.9; modified) 目辨白黑美惡，耳辨音聲清濁，口辨酸鹹甘苦，鼻辨芬芳腥臊，骨體膚理辨寒暑疾養，是又人 之所常生而有也，是無待而然者也。（〈榮辱〉） The five senses or perceptual faculties are not interchangeable because they are individuated by means of the functions of their respective organs (the eye, the ear, the nose, the mouth [or the tongue], and the body [or the skin]).14 Xunzi characterizes the functions of the five sense organs as the power of differentiating or discriminating. I propose that Xunzi’s use of the term “bian 辨” (differentiating or recognizing) must be taken seriously because the term indicates that Xunzi understands the faculty of perception (or the five senses as a whole) as the faculty of recognitional perception. There are two main ways that one might understand the expression “S perceives O” (where “S” is a subject and “O” is an object or a thing): as “S perceives O non-epistemically” and as “S perceives O epistemically”. The former says that S perceives O without possessing any concepts about O. The latter says that S perceives O with possessing and applying the concepts about O; that is, S epistemically perceives O in such a way that he perceives O as O.15 Corresponding to the two understandings of perception, as non-epistemic perceiving and as epistemic perceiving, there are two understandings of the 14 In contemporary philosophy of perception, it is still a problem that how different senses are individuated. Xunzi’s view might be classified as the “sense organ view”, which “individuate[s] sense modalities by appeal to their respective organs” (Fish 2010: 150). With regard to other approaches to individuating the senses, see, e.g., Fish (2010: Ch. 9). 15 See especially Fred Dretske (1969: Ch. 2; 2000: Essay 6) where he distinguishes “epistemic seeing” and “non-epistemic seeing”. 11 faculty of perception, as the faculty of experiencing perception and as the faculty of recognitional perception. It seems that Xunzi does not understand the functions of the five sense organs merely as the power of simply perceiving a thing but rather as the power of perceiving a thing as belonging to a certain kind or as possessing certain defining characteristics, or as the power of recognizing the things as what they are and what they are not. Further, as stated above, the power of the five senses or faculties is “given by nature” or inborn. Although, as explained in the previous section (§2.2), to state that the nature of a faculty is inborn does not imply that the exercise of the faculty must have nothing characteristically to do with human agency, this statement is not applicable to the faculty of perception. The faculties of memory and reasoning are inborn while their exercise can be controlled and improved by agency (that is, an agent can have active control over the transmission of beliefs formed or held); the faculties of sight, hearing, smell, taste, and touch are inborn and their exercise is out of the agent’s control (that is, the agent has no voluntary control over the generating of beliefs from non-doxastic sensory inputs), let alone that they can be cultivated in practice. However, the idea that the exercise of the faculty of perception is out of agent’s control and cannot be cultivated would not be accepted by Xunzi. Now let us turn to the second claim made above. The second claim is that the five senses are governed by xin. I shall explain this by focusing on the components of the faculty of perception. In the previous passages I explain what the faculty of perception (or the five senses as a whole) is in Xunzi without explaining what the components of the faculty of perception are and how they are related to each other. In Xunzi, the faculty of perception is composed of xin (the mind-heart) and the five sense organs because the exercise of the former is explained in terms of the co-exercise of the latter two. First, Xunzi asserts that the exercise of xin requires the exercise of the five sense organs: [The basis upon which we judge that things are the same or different is] the awareness that xin has of the defining characteristics that distinguish things. Only when it rests on the data provided by the ear is it possible for this awareness of the defining characteristics to know sound, and only when it rests on the data provided by the eye is it possible to know shape. This being so, xin’s awareness of defining characteristics necessarily requires that the sense organ be impressed by the type of thing to which that sense organ [is sensitive]. (Xunzi, “Zhengming”; tr. Knoblock, 22.2e) 心有徵知。徵知，則緣耳而知聲可也，緣目而知形可也。然而徵知必將待天官之當簿其類然後 可也。（ 〈正名〉） Xunzi’s notion of xin (in the context of discussing the epistemology of perception) is understood as a mental faculty that has the power of being aware of the defining characteristics of a thing, or as I shall call it, the power of conceptual recognition. Its exercise 12 “requires”, among other things, the exercise of the sense organs, and here the term “require” refers to the enabling condition rather than the constitutive condition for the exercise of xin. The power of xin in itself is not defined in terms of the function of sense organs. Second, Xunzi asserts that the exercise of the faculty of perception requires the exercise of xin. There are two main passages that support my attributing the assertion to Xunzi. Here is the first passage (P1): When xin is not employed (xinbushiyen), then although black and white are in front of a person’s own eyes, he will not see them, or although the thunder drums are sounding on either side of him, his ears will not hear them. (Xunzi, “Jiebi”; tr. Knoblock, 21.1) 心不使焉，則白黑在前而目不見，雷鼓在側而耳不聞。（ 〈解蔽〉 ） Does Xunzi mean in P1 that when xin is not employed, then although black and white are in front of a subject’s own eyes the subject does not “see” anything at all as if his eyes are closed? This seems to be, phenomenologically speaking, absurd. The term “see” in P1 must be understood as “epistemic seeing”. That is, the subject in question still non-epistemically sees the object in front of his eye but has no epistemic seeing of the object; the subject does not recognize the object in front of his eyes as the black object or as the white object. Thus, what P1 suggests is that without the exercise of xin there is no (recognitional) perception. This reading of P1, especially Xunzi’s use of the term such as “see”, can be supported by another passage (P2) in Xunzi: If the five senses come into contact with a thing and you do not become aware of it, or if xin notes its defining characteristics and you can offer no explanation, then everyone will agree that there is “no knowing.” (Xunzi, “Zhengming”; tr. Knoblock, 22.2e) 五官簿之而不知，心徵之而無說，則人莫不然謂之不知。 （〈正名〉 ） P2 contains two conditionals. The first conditional says that if the five sense organs come into contact with an object while the subject in question does not become aware of it (that is, his xin is not employed), then everyone will agree that there is “no knowing”. This conditional can be used to support my reading of P1 since Xunzi does not use the expression such as “no (non-epistemic) seeing” or “no experiencing” to describe the result of exercising the five sense organs but not exercising xin. Thus, again, without the exercise of xin there is “no knowing” (which, in this case, had better be understood as “no recognitional perception”). The claim that the five senses are governed by xin amounts to the claim that it is xin that determines the (conceptual) content of recognitional perception. The above two assertions—that the exercise of the faculty of perception requires the exercise of xin as its content determinant, and that the exercise of xin requires the exercise of the sense organs as its enabling condition—suggest that in Xunzi’s thought the faculty of perception is a combination of xin and the five sense organs. Thus, expressed in a more 13 contemporary vein, a perception can be conceived as an output of a dual process: The five sense organs issue sensations as an output, and the output, in turn, becomes the input of xin which issues awareness of defining characteristics that distinguish things as the output (i.e., the recognitional perception). Here we should note that to say that a perception requires awareness which is constitutive of the perception does not imply that one must be aware of what he is aware of. One can have a second-order reflective awareness of what he is aware of, but this second-order awareness is not a requisite constituent of perception, which requires only first-order constitutive awareness. The idea that the faculty of perception is a combination of xin and the five sense organs makes Xunzi’s notion of faculty virtue have a potential to be distinguished from the purely mechanical notion of faculty virtue because xin, when acting as the power of conceptual recognition, is something that involves agency (or the will) and can be cultivated in practice. Here I use the expression “have a potential to be distinguished from” rather than the expression “is distinguished from” because if I used the latter someone would object as follows: that the faculty of perception has xin as its component does not mean that the faculty is thus a cultivated faculty because xin might be something whose exercise is mechanical like that of sense organs. I admit the thrust of this objection. Thus I shall show later in the next section why Xunzi thinks that xin is a cultivated faculty. So far I have shown that in Xunzi’s epistemology the faculty of perception as a whole is constituted by xin and the sense organs from Xunzi’s thought that the formation of (recognitional) perception or perceptual beliefs requires the co-exercise of xin and the sense organs. The idea that xin is a component of the faculty of perception creates a potential for a faculty-based virtue epistemology to be regulative, because the faculty virtue of perception, thus properly understood, is not as mechanical as originally thought, but can be cultivated to a greater or lesser degree as the character virtues can. In what follows I turn to the issue of why and how to cultivate xin. 3.3 The Cultivation of Xin Why does xin, when acting as the power of conceptual recognition, need to be cultivated? This is because, to put in Xunzi’s term, xin might be “blinded” in exercise, or to put it in the Sosaian terminology, xin might be maladroit. Why does Xunzi choose the term “blindness” to characterize the maladroitness of the exercise of xin? Xin’s power is not merely to be aware of something, but to recognize something as belonging to a certain kind or as possessing certain defining characteristics. Assume that the nature of an object or a thing O can be represented by its possessing the defining characteristics C1, C2, C3, C4 and C5. A subject S’s perceiving of O requires the exercise of S’s xin to recognize C1 to C5 of O, ideally speaking. But it is possible that xin’s act of recognizing is performed worse, that is, S recognizes C1 as O’s mere defining characteristic. In such a case, S’s xin is blinded in the sense that it does not recognize 14 C2 to C5 as O’s defining characteristics. Xunzi, however, does not think that xin has no power to recognize C2 to C5 as O’s defining characteristics; for him, xin’s being blinded can be dispelled by self-cultivation. In exercising the faculty of perception (or in the co-exercising of xin and the sense organs), there are some cases in which xin is blinded or maladroit: As a general rule, when examining things about which there are doubts, if xin is not inwardly settled, then external things will not be clear. If my deliberations are not clear, then I will never be able to settle what is so of a thing and what is not so of it. [1] Someone walking along a road in the dark may see a fallen stone and think it a tiger crouching in ambush, or he may see an upright tree and think it a standing man. The darkness has beclouded the clarity of his vision. [2] 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. The drink has disordered his spirit. [3] Pressing against the eye while looking at an object will make it appear double; covering the ears when listening will make silence seem like a clamor. The force applied to the sense organs has disordered them. (Xunzi, “Jiebi”; tr. Knoblock, 21.8) 凡觀物有疑，中心不定，則外物不清；吾慮不清，則未可定然否也。[1] 冥冥而行者，見寢石 以為伏虎也，見植林以為後人也：冥冥蔽其明也。[2] 醉者越百步之溝，以為蹞步之澮也，俯 而出城門，以為小之閨也：酒亂其神也。[3] 厭目而視者，視一以為兩；掩耳而聽者，聽漠漠 而以為哅哅：埶亂其官也。（〈解蔽〉 ） In case [1], xin is blinded because of its environment; in case [2], xin is blinded because of its neurological condition; in case [3], xin is blinded because of its co-operators (the sense organs). Beyond the field of perceptual knowledge, xin might be blinded in the fields of political and philosophical knowledge. Xunzi mentions several philosophers whose xins are blinded: Mo Di was blinded by utility and was insensible to the value of good form. Song Xing was blinded by desire and was insensible to satisfaction. Shen Dao was blinded by law and was insensible to worth. Shen Buhai was blinded by technique and was insensible to knowledge. Hui Shi was blinded by propositions and was insensible to realities. Zhuang Zhou was blinded by Nature and was insensible to men. (Xunzi, “Jiebi”; tr. Knoblock, 21.4) 墨子蔽於用而不知文。宋子蔽於欲而不知得。慎子蔽於法而不知賢。申子蔽於埶而不知知。惠 子蔽於辭而不知實。莊子蔽於天而不知人。（〈解蔽〉） Without regard to a particular field, the blindness or the maladroitness of xin’s operation can be characterized as follows: What makes for blindness? One can be blinded by desire or aversion, by the beginnings 15 of things or their end, by what is remote or what is near, by broadness or shallowness, by antiquity or modernity. Since each of the myriad things evokes a different reaction, there is none that could not obsess xin. This is the universal flaw of the operation of xin. (Xunzi, “Jiebi”; tr. Knoblock, 21.2) 故為蔽：欲為蔽，惡為蔽，始為蔽，終為蔽，遠為蔽，近為蔽，博為蔽，淺為蔽，古為蔽，今 為蔽。凡萬物異則莫不相為蔽，此心術之公患也。（〈解蔽〉 ） The above statements exemplify the blindness of xin’s operation in various fields (such as perception, politics, and philosophy). Xunzi attempts to teach us how to dispel such blindness. However, Xunzi does not offer different prescriptions for different fields. He deals with blindness as a whole. Nonetheless, I assume that Xunzi’s formula for dispelling blindness can apply to all fields, including perception. For Xunzi, the achievement of dispelling blindness can be found in a sage: The sage knows the flaws of xin’s operation and perceives the misfortunes of blindness and being closed to the truth. This is why he is without desires and aversions, without beginnings and ends of things, without the remote or near, without broadness or shallowness, without antiquity or modernity. He lays out all the myriad things and causes himself to exactly match how each settles on the suspended balance. This is why for the sage, the multitude of different reactions to things cannot produce obsession by one thing’s beclouding another and so disturbing their proper position. (Xunzi, “Jiebi”; tr. Knoblock, 21.5a) 聖人知心術之患，見蔽塞之禍，故無欲，無惡，無始，無終，無近，無遠，無博，無淺，無古， 無今，兼陳萬物而中懸衡焉。是故眾異不得相蔽以亂其倫也。（〈解蔽〉 ） The sage always has a comprehensive view, rather than a partial view, of a thing. The question is how can the sage achieve such a state, that is, “to be without merely focusing on X or on its contrary”? According to Xunzi, “the critical factor necessary to put things in order consists in understanding dao” (Xunzi, “Jiebi”; tr. Knoblock, 21.5c). This is because “Dao itself is constant in its form yet completely changeable; one corner is an insufficient basis for drawing conclusions about it [夫道者，體常而盡變，一隅不足以舉之]” (Xunzi, “Jiebi”; tr. Knoblock, 21.4). So the question now is: How can an ordinary agent, in order to become a sage whose xin is not blinded, understand or know dao? What do men use to know dao? I say that it is xin. How does xin know? I say by its emptiness, unity, and stillness. Xin never stops storing; nonetheless it possesses what is called emptiness (xu 虛). Xin never lacks duality; nonetheless it possesses what is called unity (yi 壹). Xin never stops moving; nonetheless it possesses what is called stillness (jing 靜). (Xunzi, “Jiebi”; tr. Knoblock, 21.5d) 人何以知道？曰：心。心何以知？曰：虛壹而靜。心未嘗不臧也，然而有所謂虛；心未嘗不滿 16 [兩]也，然而有所謂壹；心未嘗不動也，然而有所謂靜。（〈解蔽〉 ） To know dao (or to have a comprehensive view of all things), xin must be empty, united, and still. These three qualities or powers can be possessed by xin through self-cultivation. The three qualities or powers are explained in what follows in turn: [1] Men from birth have awareness. Having awareness, there is memory. Memories are what is stored, yet xin has the property called emptiness. Not allowing what has previously been stored to interfere with what is being received in xin is called emptiness. (Xunzi, “Jiebi”; tr. Knoblock, 21.5d) 人生而有知，知而有志；志也者，臧也；然而有所謂虛；不以所已臧害所將受，謂之虛。 （〈解 蔽〉 ） [2] Xin from birth has awareness. Having awareness, there is perception of difference. Perception of difference consists in awareness of two aspects of things at the same time. Awareness of two aspects of things all at the same time entails duality; nonetheless xin has the quality called unity. Not allowing the one thing to interfere with the other is called unity. (Xunzi, “Jiebi”; tr. Knoblock, 21.5d) 心生而有知，知而有異；異也者，同時兼知之；同時兼知之，兩也；然而有所謂一；不以夫一 害此一謂之壹。（ 〈解蔽〉 ） [3] When xin is asleep, it dreams. When it relaxes, it moves of its own accord. When it is employed in a task, it plans. Thus xin never stops moving; nonetheless it possesses the quality called stillness. Not allowing dreams and fantasies to bring disorder to awareness is called stillness. (Xunzi, “Jiebi”; tr. Knoblock, 21.5d) 心臥則夢，偷則自行，使之則謀；故心未嘗不動也；然而有所謂靜；不以夢劇亂知，謂之靜。 （ 〈解蔽〉） I will explain these three notions in the case of recognitional perception. Let us assume that the defining characteristics of an object O are C1, C2, C3, C4 and C5. When a subject S perceptually encounters O (which he has never encountered before) and exercises his xin (together with his faculty of memory) to recognize O, it is probable that S, whose memory contains no concepts of C1 to C5, is disposed to recognize O as possessing certain defining characteristics that he already possessed. To avoid such an epistemically obstructive scenario, xin must cultivate itself to have the power of being “empty”, i.e., the power of “not allowing what has previously been stored to interfere with what is being received in xin”. Let us consider the second scenario. Assume that S’s memory contains the concepts of C1 to C5. When S perceptually encounters O and exercises his xin to recognize O as possessing C1 and as possessing C2, it is probable that S is disposed to recognize O as either possessing C1 or as possessing C2 but not both. To avoid such an epistemically obstructive 17 scenario, xin must cultivate itself to have the power of being “united”, i.e., the power of “not allowing the one thing to interfere with the other” and of synthesizing the two into a unity. Finally, when S perceptually encounters O and exercises his xin (together with his faculty of memory) to recognize O, it is probable that S, whose memory does contain the concepts of C1 to C5, is disposed to recognize O as possessing C6 to C10, none of which presents the true nature of O. To avoid such an epistemically obstructive scenario, xin must cultivate itself to have the power of being “still”, i.e., the power of “not allowing dreams and fantasies to bring disorder to awareness”, or the power of making itself effective. I will not elaborate these qualities or powers further. My aim is to show that xin can be cultivated. In my interpretation of Xunzi’s epistemology of perception, a perceptual belief is something derived from the exercise of the faculty virtue of perception, or to put it more clearly, from the co-exercise of xin and the relevant sense organs. The faculty virtue of perception can be cultivated because its essential component, xin, can be cultivated. Why does the cultivation of xin imply the cultivation of the faculty virtue of perception as a whole? It is because xin is superior to the five sense organs in the sense that it actively recognizes sensory information issued from the sense organs and determines the content of perception. Recognize more adroitly, perceive more accurately. The faculty virtue of perception, in which xin plays the pivotal role, is the faculty of perceptual recognition, which is inborn and cultivated in character.. 4. Conclusion Three tasks are achieved in this paper. First, I have shown that a regulative faculty-based virtue epistemology is possible through a (re)construction of Xunzi’s epistemology. Second, I have shown that we can achieve a slightly but significantly different understanding of the faculty virtue of perception—that is, a combination of xin and the sense organs. Third, I have shown why and in what manner Xunzi’s epistemology can be understood as a regulative faculty-based virtue epistemology. References Axtell, Guy (2000). “Introduction”, in Guy Axtell (ed.), Knowledge, Belief, and Character: Readings in Virtue Epistemology (Lanham, MD: Rowman & Littlefield). Baehr, Jason (2011). Inquiring Mind: On Intellectual Virtues and Virtue Epistemology (Oxford: Oxford University Press). Dretske, Fred (1969). Seeing and Knowing (London: Routledge and Kegan Paul). Dretske, Fred (2000). Perception, Knowledge and Belief (Cambridge: Cambridge University Press). Fish, William (2010). Philosophy of Perception (London: Routledge). Fraser, Chris (2011). “Knowledge and Error in Early Chinese Thought”, Dao 10: 127-48. Geaney, Jane (2002). On the Epistemology of the Senses in Early Chinese Thought (Honolulu: University of 18 Hawaii Press). Greco, John (2000). Putting Skeptics in their Place (Cambridge: Cambridge University Press). Hagen, Kurtis (2007). The Philosophy of Xunzi: A Reconstruction (Chicago: Open Court). Lee, Janghee (2004). Xunzi and Early Chinese Naturalism (Albany: State University of New York Press). Montmarquet, James (1993). Epistemic Virtue and Doxastic Responsibility (Lanham, MD: Rowman and Littlefield). Pritchard, Duncan (2009). “Apt Performance and Epistemic Value”, Philosophical Studies 143: 407-16. Roberts, Robert and Wood, W. Jay (2007). Intellectual Virtues: An Essay in Regulative Epistemology (Oxford: Oxford University Press). Robins, Dan (2007). “Xunzi”, in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2008 Edition), URL = <http://plato.stanford.edu/archives/fall2008/entries/xunzi/>. Sosa, Ernest (1991). Knowledge in Perspective (Cambridge: Cambridge University Press). Sosa, Ernest (2007). A Virtue Epistemology: Apt Belief and Reflective Knowledge, Volume 1 (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 2 (Oxford: Oxford University Press). Sosa, Ernest (2011). Knowing Full Well (New Jersey: Princeton University Press). Stanley, Jason & Williamson, Timothy (2001). “Knowing How”, Journal of Philosophy 98(8): 411-44. Tsai, Cheng-hung (2006). “Can Knowing-How Skepticism Exist?” accepted by The Electronic Journal of Analytic Philosophy. Tsai, Cheng-hung (2011a). “Linguistic Know-How: The Limits of Intellectualism”, Theoria: A Swedish Journal of Philosophy 77(1): 71-86. Tsai, Cheng-hung (2011b). “The Metaepistemology of Knowing-How”, Phenomenology and the Cognitive Sciences 10(4): 541-56. Turri, John (2012). “Critical Notices of Robert C. Roberts and W. Jay Wood, Intellectual Virtues: An Essay in Regulative Epistemology”, Philosophy and Phenomenological Research 82(3): 793–97. Turri, John & Sosa, Ernest (2010). “Virtue Epistemology”, in Duncan Pritchard (ed.), Oxford Bibliographies Online: Philosophy. Wolterstorff, Nicholas (1996). John Locke and the Ethics of Belief (Cambridge: Cambridge University Press). Xunzi: A Translation and Study of the Complete Works, trans. John Knoblock, Stanford: Stanford University Press, vols 1-3; vol. 1 (1988), books 1-6; vol. 2 (1990), books 7-16; vol. 3 (1994), books 17-32. Zagzebski, Linda (1996). Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge (Cambridge: Cambridge University Press). Zagzebski, Linda (2009). On Epistemology (Belmont, Calif.: Wadsworth). 19