XIV—Swimming Happily in Chinese Logic

Abstract

Dipping in Chinese waters

pulled and pushed by Mo

we see how Zhuangzi caught us

like the happy fish we know

we follow their flow

words matching as they sort us

I

The Chinese Logic Dialogue. Let us dive straight in with a story, which I shall call the Chinese logic dialogue.

Joanne and Hugh were discussing a passage from an ancient Chinese philosophical text.

Joanne said, ‘Look at the dialogue scripted here, this is cleverly argued.’

Hugh said, ‘You’re not a Chinese logician, how do you know it is cleverly argued?’

Joanne said, ‘You’re not me, how do you know that I don’t know it is cleverly argued?’

Hugh said, ‘I am not you, so certainly don’t know you; you are certainly not a Chinese logician, so the case is complete for your not knowing it is cleverly argued.’

Joanne said, ‘Let’s go back to the beginning. You said, ‘How do you know it is cleverly argued?’, in asking me which you already knew I knew it; I know it from using our shared reason.’

This transforms a famous story in the Zhuangzi, the happy fish dialogue, which we will examine shortly. The Zhuangzi is the richest and most intellectually challenging of all the texts of ancient Chinese philosophy, and it is full of stories which have been interpreted in a wide variety of ways over the subsequent two millennia. It is one of the founding texts of Daoism, but it is also central to the entire landscape of ancient Chinese thought. In this paper I want to explore some of the logical features of this landscape by focusing on the happy fish dialogue and addressing what we can call the problem of Chinese logic. Are there distinctive forms of argumentation and analysis in ancient Chinese thinking and/or distinctive logical conceptions and theories? Or can Chinese reasoning be analysed and entirely explained by modern forms of logic, such as propositional logic and quantificational theory? What implications does this have for how we interpret historical texts?

In what follows, I will first say more about the problem of Chinese logic (§ii) before introducing the Zhuangzi and the happy fish dialogue (§iii). I will then elucidate the key concepts involved in this dialogue (§iv) and discuss selected interpretations of it (§v). I conclude by returning to the Chinese logic dialogue, which frames my analysis of the happy fish dialogue, in drawing out the hermeneutic implications (§vi).

II

The Problem of Chinese Logic. What is today the standard Chinese term for ‘logic’, luójí, was only introduced (as 邏輯) in 1902, although there were all sorts of attempts to translate ‘logic’ into Chinese from the time of the Jesuit mission in the early seventeenth century (see Kurtz 2011). But the process of translating Western works of logic and philosophy only started in earnest at the turn of the twentieth century, as the Qing dynasty disintegrated. By the time it finally fell, in 1912, debate was already under way about the relationship between Western and Chinese forms of thinking.

Central in this debate was the role of logic in ancient Chinese philosophy. Some Chinese scholars sought to show that the various forms of reasoning recognized in Western logic could be found exemplified in the ancient Chinese texts. There was nothing to compare with the formal logic developed by Aristotle and the Stoics, but there was certainly concern with argumentation (biàn 辯) and explicit discussion of names (míng 名). Other scholars were determined to identify a distinct form of ‘Chinese logic’. All great civilizations, it was felt, developed some form of logic in their ‘axial age’, as Karl Jaspers was later to call it. European philosophy had Greek logic, and Indian philosophy had Buddhist logic, yet there was nothing comparable in Confucianism, and Daoism was seen as hostile to ‘logic’. There was one tradition that had been important in the ancient period, however, but which disappeared after China was unified in the second century bce. This was the Mohist tradition, and it was to Mohism that scholars looked to find ‘Chinese logic’.

2.1 Mohism

The Mohist tradition was inspired by the teachings of Mòzǐ (墨子; c.470–390 bce). The text named after him was compiled by his followers, and in the first three parts, in promoting their key idea of ‘inclusive care’ (jiān’ài 兼愛), there is a sustained critique of Confucianism. Precisely in countering Confucianism, the Mohists realized that they needed to provide arguments for their claims, which they set out with great clarity.

It is in the fourth part of the Mòzǐ, however, that we find the fullest account of Mohist logic. Known as the ‘Dialectics’, this is the work of the later Mohists, and consists of six books of which the most relevant for our purposes (Book 45) is the ‘Lesser Selection’ (xiǎo qǔ 小取). Mohist logic can be characterized as primarily concerned with one-step inferences involving the compounding of names, inferences of the form ‘A is B; so FA is FB’. Some of these inferences are good, such as ‘white horse is horse; so riding white horse is riding horse’; some are bad, such as ‘robbers are people; so caring about robbers is caring about people’ (Mòzǐ, 45.2b–c). The Mohists offered no systematic theory of why some such inferences are good and some bad, but they discussed different kinds of examples and identified analogies and disanalogies between them.

To give just one illustration, the Mohists held both that robbers are people and that killing people is wrong, but they also wanted to allow that killing robbers is right (in certain circumstances). This might strike us as inconsistent, but on their view, the inference from ‘robbers are people’ to ‘killing robbers is killing people’ is analogous to the bad inference from ‘robbers are people’ to ‘caring about robbers is caring about people’ (Mòzǐ, 45.2c). Of course, one way to claim that killing people is wrong but that killing robbers is not wrong is to distinguish two kinds of killing, say, ‘murder’ and ‘execution’: murder is wrong but not execution (cf. Graham 2003, pp. 488–9; Lai 2017, pp. 150–5). The relevant text makes clear what is meant by ‘caring about people’ (that is, ‘inclusive care’), on the basis of which we can reconstruct why they thought that ‘caring about people’ is a different kind of caring from ‘caring about robbers’. But the reason that is given for the supposed failure of the inference in the case of killing robbers is not the ambiguity of ‘killing’ but simply the analogy to the failure of the inference in the case of caring about robbers.

2.2 Analogical Reasoning

The fundamental role of analogical reasoning in Chinese argumentation comes out most clearly in the ‘Lesser Selection’, which elucidates the Mohist conception of ‘disputation’ (biàn 辯). The key passage is the following:

Analogy [pì 譬] is mentioning other things and using them to clarify it. ‘Paralleling’ [móu 侔] is placing expressions side by side and jointly proceeding. ‘Pulling’ [yuán 援] is saying, ‘You are so, how is it that I alone cannot be so?’ ‘Pushing’ [tuī 推] is, on this basis that what they don’t accept is the same as what they do accept, proposing it. …

Things have respects in which they are the same, yet it doesn’t follow that they are completely the same. Parallels between expressions are correct only up to a point. When things are so, there is that by which they are so. Their being so is the same, but that by which they are so isn’t necessarily the same. When people accept things, there is that by which they accept them. Their accepting them is the same, but that by which they accept them isn’t necessarily the same. Thus expression in analogies, paralleling, pulling, and pushing become different as they proceed, become dangerous as they change direction, fail when taken too far, and leave their roots as they flow, and so one cannot fail to be cautious and cannot invariably use them. (Mòzǐ, 45.1d–e)

The notions of ‘analogy’, ‘paralleling’, ‘pulling’ and ‘pushing’ can be illustrated by returning to the example of killing robbers. The analogy between killing and caring about people is what the Mohists appeal to in ‘clarifying’ why killing robbers is not wrong. The inferences involved in this analogy are ‘robbers are people; so killing robbers is killing people’ and ‘robbers are people; so caring about robbers is caring about people’. Here there is both a parallel between ‘robbers are people’ and ‘F-ing robbers is F-ing people’ and a parallel between the two inferences. Placing them ‘side by side’ is what constitutes or enables ‘proceeding’ in our disputation.

‘Pulling’ is the literal translation of ‘yuán’ (援), which has also been translated as ‘adducing’, and here we can imagine someone disagreeing with the Mohists about the analogy between the two inferences. The Mohists challenge them by asking, ‘If you hold that the second inference (concerning caring about robbers) is bad, then how is it that we cannot hold that the first inference (concerning killing robbers) is bad too?’ The question is intended to ‘pull’ their opponent towards their own point of view.

‘Pushing’ is the literal translation of ‘tuī’ (推), which has also been translated as ‘inferring’. Here we can imagine the Mohists arguing with their opponent as follows, making explicit what was conversationally implied in their attempt at ‘pulling’: ‘You hold that the second inference is bad; but the first inference is analogous; so you should accept that the first inference is bad.’ The aim here is to ‘push’ them into accepting what they do not currently accept, in other words, to infer something from what they do believe.

The second paragraph of the passage cited above shows that the Mohists recognized the dangers of analogical reasoning. Some analogies are convincing, some are not, and transitivity soon fails in chains of such reasoning. So we must always be on the lookout for failures. The Mohists, among the various philosophical traditions in ancient China, were the most reflective about this.

2.3 The School of Names

There is one other philosophical tradition that must also be mentioned in sketching, however briefly, Chinese logic. This is the so-called School of Names (Míngjiā 名家). Two of its most important representatives were Hui Shi (c.370–310 bce) and Gongsun Long (c.320–250 bce). Both are famous for the paradoxes they formulated: Hui Shi for his ‘ten theses’, which raise problems of relativity to perspective, and Gongsun Long for his notorious claim that ‘white horse is not horse’ (bái mă fēi mă 白馬非馬). Paradoxes have always been a stimulus to philosophical thinking, and these were no exception, motivating the work of the later Mohists, in particular. As we will see, Hui Shi was also Zhuangzi’s main sparring partner, taken in the text of the Zhuangzi as the representative of the ‘logician’.

2.4 The Answer in Outline

This sketch already suggests an answer to the problem of Chinese logic. Analogical reasoning, in which one-step inferences involving the compounding of names play an important role, is what is characteristic of Chinese argumentation, and reflection on this lay at the heart of the work of the later Mohists. So there is indeed something distinctive about Chinese logic. On the other hand, the use of analogies throughout ancient Chinese philosophy, and the ways in which they were accepted or rejected, involve forms of analysis and reasoning that were not systematically codified. So this is where an understanding of modern logic can help—to make explicit what was implicit. Taking the happy fish dialogue in the Zhuangzi as our example, and analysing it in detail, will fill out this answer.

III

Zhuangzi and the Happy Fish Dialogue. The Zhuangzi (莊子) is named after Zhuāng Zhōu (莊周), who flourished in the later half of the fourth century bce. Zhuangzi himself has traditionally been taken as the author of some of the chapters, although his precise contribution is uncertain. Whoever the author(s) may have been, however, there is a consistency of view and style across many of the chapters that makes it appropriate to treat them as a unit and to use the name ‘Zhuangzi’ as a shorthand way of referring to their author(s).

3.1 Perspectivism in the Zhuangzi

Perspectivist, relativist and sceptical themes pervade the Zhuangzi. Perspectivism can be characterized as the view that all knowledge is perspectival, and Zhuangzi was (broadly) a perspectivist in this sense: we can only know something from a particular point of view or within a particular conceptual scheme or ethical practice. The key question is whether he also held that all perspectives are equally valid, which might be taken to imply relativism and the repudiation of objective knowledge. My own view is that he refrained from drawing relativist conclusions, although relativist—and sceptical—arguments are deployed throughout his thinking to combat dogmatism. Arguably, his (epistemological) perspectivism is embedded in a broader (metaphysical) conception of a oneness that underlies and unites the various perspectives, on the basis of which Zhuangzi urges us to recognize how perspectives connect with one another, thereby undermining dogmatism. Given the centrality of this idea, I shall describe his core philosophical outlook as connective perspectivism.

The key passage, in this respect, occurs in chapter 2 of the Zhuangzi, entitled ‘Discourse on Equalizing Things’ (Qí wù lùn 齊物論), in which Zhuangzi elaborates on his conception of a dào (道). Often translated as ‘way’ or ‘path’, it can be understood here as any kind of linguistic or ethical practice. Such daos, he writes, are formed by walking them, and for any such dao, there is thus some perspective (suǒ 所) from which it can be deemed right or acceptable. But he nevertheless stresses that for any given perspectives, there is a dao that connects them into one (dào tōng wéi yī 道通為一; ICS Zhuangzi, 2/5/2). This is the heart of Zhuangzi’s Daoism. He is a pluralist about daos, but he renounces the radical relativism that this might seem to imply by emphasizing how any one perspective—or rather, the perspective one inevitably occupies at any one time—opens out into another. It is fine to occupy that perspective (for appropriate purposes), on Zhuangzi’s view, but the fear of other, perhaps alien, perspectives is removed—and dogmatism repudiated—by appreciating how they interconnect.

3.2 The Happy Fish Dialogue

The happy fish dialogue is one of the most famous stories in Chinese literature. It concludes chapter 17 of the Zhuangzi, which is entitled ‘Autumn Waters’ (Qiū shuǐ 秋水). Perspectivism is the central theme of this chapter as well, developed through a series of scenes that might indeed be interpreted as different perspectives opening up into one another, the realization of which is what the text itself performs (see Meyer 2015). The final scene, in which Zhuangzi and Huizi (Hui Shi) engage in an argument over whether the fish they see are happy, can then be taken as encapsulating and expressing Zhuangzi’s connective perspectivism.

The dialogue, though, has generated a range of different interpretations, as we will see in §v. Indeed, the dialogue can itself be read as raising the problems of interpretation and understanding that bedevil many areas of philosophy: the problem of other minds, of interspecies understanding, of multicultural dialogue, of interpreting historical texts, of comparative philosophy, and so on. To adapt a remark of Wittgenstein’s (2009, p. 233), what we have here is a whole cloud of philosophy condensed into a drop of dialogue. It also raises the problem of Chinese logic, and the interpretations I discuss in §v have been selected to focus on this problem.

3.3 Text and Translation

Here is the text itself, in both Chinese characters and pinyin, with an English translation.¹ The lines have been numbered, for ease of reference later.

1. 莊子與惠子遊於濠梁之上。Zhuāngzĭ yŭ Huìzĭ yóu yú háo liáng zhī shàng.Zhuangzi and Huizi were roaming on the bridge over the river Hao.

2. 莊子曰：「儵魚出遊從容, 是魚樂也。」Zhuāngzĭ yuē: ‘shū yú chū yóu cóng róng, shì yú lè yĕ.’Zhuangzi said, ‘Look at the darting fish coming out to roam around, this is fish happiness.’

3. 惠子曰：「子非魚, 安知魚之樂？」Huìzĭ yuē: ‘zĭ fēi yú, ān zhī yú zhī lè?’Huizi said, ‘You are not a fish, how do you know fishes’ happiness?’

4. 莊子曰：「子非我, 安知我不知魚之樂？」Zhuāngzĭ yuē: ‘zĭ fēi wǒ, ān zhī wǒ bù zhī yú zhī lè?’Zhuangzi said, ‘You’re not me, how do you know that I don’t know fishes’ happiness?’

5. 惠子曰：「我非子, 固不知子矣；子固非魚也, 子之不知魚之樂全矣。」

   Huìzĭ yuē: ‘wǒ fēi zĭ, gù bù zhī zĭ yĭ; zĭ gù fēi yú yĕ, zĭ zhī bù zhī yú zhī lè quán yĭ.’Huizi said, ‘I am not you, so certainly don’t know you; you are certainly not a fish, so the case for your not knowing fishes’ happiness is complete.’

6. 莊子曰：「請循其本。子曰『汝安知魚樂』云者, 既已知吾知之而問我, 我知之濠上也。」Zhuāngzĭ yuē: ‘qǐng xún qí běn. zĭ yuē ‘rŭ ān zhī yú lè’ yún zhĕ, jì yĭ zhī wú zhī zhī ér wèn wǒ , wǒ zhī zhī háo shàng yĕ.’Zhuangzi said, ‘Let’s go back to the beginning. You said “How do you know fishes’ happiness?”, in asking me which you already knew I knew it; I know it from here on the river Hao.’

IV

The Happy Fish Dialogue: Conceptual Clarification. There are six terms that are key to the understanding of the happy fish dialogue, which I will briefly elucidate to set the stage for discussion of the interpretations that follow.

4.1 yóu (遊)

‘yóu’ is one of the most important terms in the Zhuangzi, being used some ninety-five times. It has the sense of ‘roam’, ‘ramble’, ‘stroll’, ‘wander at ease’, and is used, most notably, in the title of the very first chapter, ‘xiāo yáo yóu (逍遙遊)’, which can be translated as ‘freely and widely roaming’. The basic idea of yóu is of a moving around that is unconstrained by any specific purpose. The term is used here for both the roaming of Zhuangzi and Huizi and the swimming of the fish, the (presumably intended) parallel indicating some kind of shared experience, supporting—or at least anticipating—Zhuangzi’s argument.²

4.2 yú (魚)

‘yú’ is the standard term for fish. Fish are mentioned around thirty times in the Zhuangzi, more than any other kind of non-human living being. Of particular significance is the fact that fish live in water: water is a common metaphor in Chinese literature for the dao (see esp. Allan 1997). As Franklin Perkins (2015) has argued, fish play three roles in the Zhuangzi. First, they draw attention to the limitations of any mode of being or perspective. Fish are clearly constrained by living in water. Secondly, they illustrate what it is to be at home in an environment or to inhabit a perspective. Fish have their own form of life, which is fine for them, though obviously unsuitable for us as humans. Thirdly, they highlight the problem of understanding across perspectives. It seems scarcely credible that we could ever ‘see things’ as a fish does, or ‘know what it’s like’ to be a fish. It is this third role that is especially relevant here: Huizi represents the view that there is incommensurability, while Zhuangzi suggests that there is some kind of connection between us.

4.3 lè (樂)

‘lè’, as a noun, means happiness, joy, or enjoyment, and as an adjective, happy or joyful. The key interpretive question concerns whether ‘happiness’ was understood in ancient China as an inner state or an embodied activity. Is the question raised by the dialogue ‘What is it like to be a happy fish, and how can we know it?’ or ‘What is happy fish activity, and how can we know it?’? If the former, then we might take Zhuangzi to be stressing the analogy to the happiness that he and Huizi are experiencing in their shared roaming. If the latter, then it makes more convincing Zhuangzi’s answer, ‘That is fish happiness’, as he points to the fish that they both see swimming around below the bridge from where they are looking. The two questions—and answers—may not be incompatible, however.

4.4 ān (安)

‘ān’, as an interrogative, means ‘how’ or ‘whence’. A natural translation of ‘ān zhī (安知)’ would be ‘How do you know?’ In English, this has two meanings, depending on the emphasis: ‘How do you know?’ (as in ‘How can you possibly know?’) versus ‘How do you know?’ (as in ‘How did you come to know?’). The first asks for justification that what you have is genuine knowledge, and this seems to be the primary meaning of ‘ān zhī’ in the ancient Chinese texts. The second presupposes that you have knowledge and asks for clarification of how you acquired that knowledge. In particular, in its sense of ‘Whence (that is, from where) do you know?’, it requests specification of location. It is natural that Huizi, as the logician, should ask for justification, but equally natural that Zhuangzi, as the Daoist, should be offering clarification—of the relevant perspective. The interpretative issue concerns the use to which the ambiguity is put in the text. Is it just a playful pun exemplifying Zhuangzi’s wit, or is there some more serious purpose? Does he realize that Huizi has got the better of him and so quickly changes tack, or has he somehow drawn Huizi into contradicting himself and now shows him the way out of the fish bottle?

4.5 zhī (知)

‘zhī’ is the central philosophical term in the dialogue. As a verb it can mean ‘know’, ‘understand’, ‘discern’, or ‘be acquainted with’, and as a noun ‘knowledge’, ‘wisdom’, or ‘consciousness’. The philologist Duan Yucai (1735–1815) explained the character ‘知’ as composed of ‘矢’ (shǐ), meaning arrow, and ‘口’ (ko˘u), meaning mouth, so that ‘知’ means ‘quick-witted’ (see Ames 2015, p. 273). We could gloss this as ‘verbally hitting the target’. As in the case of ‘happiness’, ‘knowing’ might be construed as less an inner (mental) state than an activity—in this case, the activity of using words correctly, fittingly attuned to the situation. To know what a ‘fish’ is, for example, is to be able to say ‘fish’ or ‘not-fish’ in the appropriate circumstances, a form of knowing-how rather than knowing-that—practical rather than theoretical knowledge. The opposition between Zhuangzi and Huizi might then be seen as one between representatives of different conceptions of knowledge, an opposition that has taken centre stage in recent debates in epistemology.

In my view, however, the fixation on knowing-that and knowing-how distorts our understanding of Chinese philosophy, since it obscures other conceptions—most notably, that of knowing-as, as I shall call it.³ The later Mohists define ‘knowing’ (zhī 知) as ‘connecting’ (jiē 接), explained as discerning the features of something (Mòzǐ, iv, a5), itself understood as recognizing its likeness to an appropriate model or standard (fǎ 法) for having those features (Mòzǐ, iv, a70–1). So what we might describe as knowing that something is X, or knowing how to name something ‘X’, is conceived by the Mohists as (known as!) knowing it as X—as like our standard for being X. Arguably, then, knowing-as is the more fundamental form of knowing, in Mohist epistemology, and this fits especially well with Zhuangzi’s perspectivism.

4.6 běn (本)

‘běn’ means root, origin, beginning, or starting point. It is translated here as ‘beginning’, since talk of ‘going back to the beginning’ sounds most natural in English. But it is important to recognize that its primary sense is ‘root’ or ‘origin’—what we might indeed call its root meaning. (The Chinese character itself suggests the idea of a root.) What Zhuangzi is arguably proposing to Huizi is that they return not just to the starting point of their argument but to the root experience (or perspective) from which the claimed knowing flows, an experience—of seeing fish swimming around—that the two of them shared in roaming together on the bridge.

V

The Happy Fish Dialogue: Logical Analysis. One interesting feature of accounts by ‘Western’ philosophers of Chinese ideas and texts is the way in which debates between ‘analytic’ and ‘“continental”’ philosophy are played out.⁴ Ancient Chinese thought is often taken to be characterized by greater fluidity in its concepts and less rigour in its argumentation than is typically deemed acceptable by analytic philosophers. Continental philosophers have tended to see this as a virtue, posing challenges to analytic philosophy. But more analytic approaches to Chinese philosophy also raise issues for analytic philosophy, as there are disputes within analytic philosophy, just as there are within any philosophical tradition. Indeed, it is important to recognize that there are disputes within Chinese philosophy, as the happy fish dialogue itself illustrates. Huizi represents a more analytic strand in Chinese philosophy, with Zhuangzi’s ideas partly formed in reaction to this. In what follows, I discuss two analytic interpretations that focus on the logic of the central argument between Zhuangzi and Huizi, and end by outlining two other readings. But before doing so I shall briefly survey some of the broadly continental ones.

5.1 ‘Continental’ Interpretations

Of all the ancient Chinese philosophical texts, the Zhuangzi is by far the most playful—in the stories that are told and in the dazzling inventiveness of the language used. It is unsurprising, then, that what I call ‘playful’ interpretations of the happy fish dialogue are quite popular. One example is that offered by Hans-Georg Moeller (2015), who makes a great deal of the play on the ambiguity of ‘ān zhī’ (‘How do you know?’) and ‘yóu’ (‘roam’), as used for both fish and humans. The dialogue, he suggests, should primarily be seen as just an enjoyable illustration—or parody—of the ‘intellectual ramblings’ of philosophers.

Daoism has a religious as well as philosophical dimension, and there are ‘mystical’ readings of the dialogue too. Eske Janus Møllgaard (2005), for example, has argued that Zhuangzi’s spontaneous exclamation, ‘That is the joy of fish!’, records a mystical experience of ‘pure coming-into-being’, reflecting his own realization of the joy of life that arises in ‘roaming’. Hans Peter Hoffmann (2015), by contrast, has interpreted this exclamation as expressing delight in the joy of language, offering what he himself calls a ‘literary’ reading, emphasizing Zhuangzi’s celebration of the richness and fluidity of language.

A more substantial interpretation is provided by Roger Ames (2015), which we can call ‘contextualist’ not just in providing context to the dialogue but also in articulating a contextualist conception of knowledge. Knowing, as Ames describes it, is holistic, perspectival, collaborative and situated: ‘For Zhuangzi, knowing, rather than being a true idea in the mind of some isolated experiencer, is always proximate as the quality of a particular, situated experience’ (2015, pp. 281, 285). The idea of knowing as a shared experience is also developed in the ‘phenomenological’ interpretation elaborated by two Japanese scholars, Toshio Kuwako (2015) and Takahiro Nakajima (2015). Experience, Nakajima argues, ‘must be open to what is not itself’ (2015, p. 173), and what Zhuangzi reminds Huizi of at the end of their dialogue is their openness to one another in their roaming together, which reflects Zhuangzi’s openness to the embodied happy experience of the fish.

All these interpretations enrich our understanding of the dialogue and its broader context. But they are all open to one major objection: they offer no analysis or evaluation of the argument between Zhuangzi and Huizi. So let us turn to the two analytic interpretations.

5.2 Hansen’s Analysis

The most detailed reconstruction of the argument in the happy fish dialogue, drawing on the resources of modern analytic philosophy, has been offered by Chad Hansen (2003). On his account, Zhuangzi and Huizi agree in what he calls their perspectival relativism, their disagreement lying in how to formulate its implications: ‘The issue is not simply ‘does Zhuangzi know?’ but ‘what is the appropriate standard of attributing knowledge?’’ (2003, p. 54). Their respective views and favoured standards are not stated explicitly in the dialogue, however, so have to be teased out inferentially, by attending to what is needed to make sense of the moves in the argument. Hansen’s analysis takes up eight pages (pp. 56–63), interweaving logical analysis with philosophical and linguistic clarifications. Since some of the latter has already been provided, let me reconstruct his reconstruction as succinctly as I can to bring out the logic of the argument, as he interprets it, under his own four headings, which correspond to lines 2 to 5 of the dialogue.⁵

1

See the fish swimming at leisure (Zhuangzi)

From seeing the fish swimming around, Zhuangzi asserts:

(f) This is fish happiness.

(f) could be interpreted in one of three ways:

(F1) This is what it is like to be a happy fish.

(F2) These fish are happy.

(F3) This is happy fish activity.

(f1) opens up Nagel-style scepticism (how do you know what it is like to be a happy fish?), (f2) yields traditional ‘other minds’ scepticism (how do you know that fish are happy?), and (f3) suggests a more ‘direct’ report and makes Zhuangzi’s later ‘return to the root’ more plausible.

How do you know? (Huizi)

In asking Zhuangzi how he knows fishes’ happiness, Huizi invokes the following conversational norm:

(c) In asserting something, one should be prepared to respond to a challenge of ‘How do you know?’

Huizi’s challenge to Zhuangzi’s assertion is made after asserting something himself, namely, that Zhuangzi is not a fish. The implication is that there is a principle from which it then follows that Zhuangzi does not know fishes’ happiness. Let us formulate this principle in a schematic form:

(p) One must be X to know X’s F.

This schema can be interpreted in a number of ways, depending on what the ‘X’ and ‘F’ stand for. As far as the dialogue is concerned, the relevant (partial) instantiations, formulated as two pairs, are (pn) and (pk), interpreting the ‘X’, and (pa) and (pc), interpreting the ‘F’:

(pn) One must be N to know N’s F, where ‘N’ names a particular individual.

(pk) One must be of kind K to know N’s F, where ‘N’ names any individual of kind K.

(pa) One must be X to know X’s affective state.

(pc) One must be X to know X’s cognitive state.

All that is needed to derive Huizi’s conclusion that Zhuangzi does not know fishes’ happiness is (pk) and (pa), combined as (pka):

(pka) One must be of kind K to know N’s affective state, where ‘N’ names any individual of kind K.

With the kind K here being fish, and the affective state being happiness, we then have the following specific instantiation:

(pfh) One must be a fish to know fishes’ happiness.

Since Zhuangzi is not a fish, it therefore follows by (pfh) that Zhuangzi cannot know fishes’ happiness.

3

How do you know I don’t know? (Zhuangzi)

In Zhuangzi’s first response to Huizi’s challenge, he does not answer directly by giving his grounds for claiming to know fishes’ happiness. He counterattacks—by accepting both (c), in challenging Huizi’s own assertion, as well as (p), interpreted as combining (pn) and (pc), which can be formulated as follows:

(pnc) One must be N to know N’s cognitive state, where ‘N’ names a particular individual.

Zhuangzi correctly asserts that Huizi is not Zhuangzi, in which case it follows by (pnc) that Huizi cannot know Zhuangzi’s cognitive state. So Huizi cannot know that Zhuangzi doesn’t know fishes’ happiness.

The crucial move that Zhuangzi makes here is to interpret (p) in a stronger form than was needed by Huizi to derive his implied claim that Zhuangzi doesn’t know fishes’ happiness. This is the trap that Zhuangzi sets, which Huizi falls into.

4

Right! Not being you, I don’t know you; you, not being fish, don’t know fish—that is the whole of it! (Huizi)

Huizi accepts both Zhuangzi’s correct assertion that Huizi is not Zhuangzi as well as (pnc), the stronger version of (p). Indeed, given that he has already accepted (pa), albeit as combined with (pk), we can see him as endorsing the even stronger principle:

(pnac) One must be N to know N’s state, whether affective or cognitive, where ‘N’ names a particular individual.

From (pnac) it then follows that Huizi doesn’t know any of Zhuangzi’s states, whether affective or cognitive—that is, in short, that he simply doesn’t know Zhuangzi.

However, it is at this point that Huizi oversteps, falling into Zhuangzi’s trap. He thought he had trapped Zhuangzi, by getting him to agree to the principle (p), and in an even stronger form than was needed. Happy with himself, he simply turns his original rhetorical question (‘You are not a fish, how do you know fishes’ happiness?’) into an (enthymatic) argument (‘You are certainly not a fish, so the case for your not knowing fishes’ happiness is complete’), with the principle that governs this argument now accepted by Zhuangzi. The problem is that the conclusion of this argument, that Zhuangzi does not know fishes’ happiness, as asserted by Huizi, conflicts with Huizi’s assertion that he doesn’t know Zhuangzi. If Huizi is held to (c), the conversational norm that they have both also endorsed, then he has to show that he knows both that he doesn’t know Zhuangzi and that Zhuangzi does not know fishes’ happiness; but the former rules out knowing the latter. (Both assertions might be true, but Huizi cannot claim to know both.) His happiness evaporating, Huizi is snared by his own trap.

5.3 Hansen’s Evaluation

There is more in Hansen’s analysis than I have just reconstructed. He also argues that there is a conflict between Huizi’s implicit acceptance of knowledge by inference, in the use he makes of principle (p), and his endorsement of (p) in its strongest form, that is, (pnac), which Hansen calls the inner-perspective principle (2003, p. 62), which can be stated as follows:

(ipp) One can only know something from one’s own first-person, subjective perspective.

There may or may not be such a conflict in Huizi’s thinking, but it is irrelevant to the trap that is sprung, and unnecessarily complicates the analysis.

Hansen’s diagnosis of what goes on in the final line of the dialogue can also be simplified. Having caught Huizi in a trap, Zhuangzi then offers him a way out by going back to the beginning. There is no verbal trickery on Zhuangzi’s part, only a helping hand. Huizi’s challenge of ‘How do you know?’ had been issued in the context of implicit acceptance of (p), understood in some form. But having led Huizi into contradicting himself, the question can be asked again without inferentially assuming (p). How does Zhuangzi know fish happiness? From here on the river Hao, where Huizi is too—and hence in a position to know just what Zhuangzi knows. It was Huizi’s adherence to (p) that had blocked him from recognizing this, and once that obstacle is removed, the way is open for him to accept that Zhuangzi does indeed know fish happiness.

Of course, Huizi may still refuse the helping hand. (The dialogue does not end: ‘Of course, Zhuangzi, I now see that you are absolutely right. You must know how happy I am!’) As Hansen points out, there are two other ways that Huizi could escape from the trap. One is to accept only the weaker form of (p), as represented by (pk) rather than (pn): he can know Zhuangzi’s states, since he is human himself; but no human can know fishes’ states. The other way is to refrain from asserting that he knows that Zhuangzi does not know fishes’ happiness. He should remain agnostic. Hansen suggests that Huizi errs in drawing absolutist conclusions from his relativism (pp. 63, 65, 70–2). Before we can assess this, however, we need to consider another logical analysis of the argument, according to which Huizi comes out better.

5.4 Teng’s Analysis

Rather than analysing the argument between Huizi and Zhuangzi inferentially, as Hansen does, Norman Teng (2006) has suggested that it is best analysed in Mohist terms, drawing on the ‘Lesser Selection’.

As we saw in §2.2, paralleling is defined as ‘placing expressions side by side and jointly proceeding’. Teng notes that there are two parallel patterns in the dialogue:

(p1) X is not Y; whence does X know the state Y is in?

(p2) X is not Y, so X does not know the state Y is in.

(p1) is used in Zhuangzi’s first response to Huizi’s challenge, in lines 3 and 4 of the dialogue:

Huizi: You are not a fish; whence do you know fish happiness?

Zhuangzi: You are not me; whence do you know that I don’t know

fish happiness?

Here we have paralleling together with adducing (‘pulling’). ‘If you can question how I know fish happiness, then I can question how you know that I don’t know fish happiness?’ is what Zhuangzi is adducing in this case, in throwing the challenge back at Huizi.

(p2) is used in Huizi’s response to this counter-challenge, in line 5 of the dialogue:

I am not you, so I don’t know you.

You are not a fish, so you don’t know fish happiness.

Here we have paralleling together with inferring (‘pushing’). From Zhuangzi’s counter-challenge, Huizi is assuming that Zhuangzi would accept his first argument (‘I am not you, so I don’t know you’) as well as the premiss of the second argument (‘You are not a fish’), and hence, given the parallel, would be led to infer what he had earlier refused to accept (‘You don’t know fish happiness’).

On Teng’s account, then, Zhuangzi’s first response to Huizi is ‘a combined exercise of parallelizing and adducing rather than a contrived, logical trap’, and Huizi’s counter-response is ‘a combined exercise of parallelizing and inferring rather than a mishandling of the dialectic’ (Teng 2006, p. 134). From the perspective of Mohist logic, he writes, ‘debating is a joint enterprise. It entails an exchange of the debaters’ views, and demands that those who are engaging in a debate consistently align what they themselves approve or disapprove of with what is to be demanded of the others’ (2006, p. 134).

With this in mind, Teng criticizes Hansen’s account of the attribution to Huizi of the inner-perspective principle—(pnac) or (ipp), as formulated above. As Teng notes, Huizi’s initial challenge only invokes the weaker, species-relativist principle—(pka). On the species-relativist view, we cannot know the affective or cognitive states of members of other species, but we may know those of members of our own species, namely, other humans. But what Zhuangzi tries to do, as we have seen, is get Huizi to endorse the stronger principle (pnac), which we can now describe Zhuangzi as doing by paralleling. Consider the two key arguments (substituting for the indexicals):

(ZF) Zhuangzi is not a fish, so doesn’t know fish happiness.

(HZ)  Huizi is not Zhuangzi, so doesn’t know Zhuangzi’s (affective or cognitive) states.

Huizi implicitly appeals to (zf) in his initial challenge, while Zhuangzi implicitly appeals to (hz) in his counter-challenge. But if Zhuangzi is right in paralleling and inferring from (zf) to (hz), then Huizi is no less right in paralleling and inferring back from (hz) to (zf). In other words, if it is right to ‘derive’ (pnac) from (pka), then it is no less right to ‘derive’ (pka) from (pnac). ‘Paralleling’ suggests that they are ‘equivalent’. Zhuangzi and Huizi are cooperating in establishing this purported equivalence. But if this is so, then Zhuangzi should agree that he doesn’t know fish happiness.

The problem, of course, is that Huizi cannot consistently assert the conclusions in (zf) and (hz), and if Zhuangzi and Huizi are indeed ‘aligning’ themselves in their debate, then they will both recognize this. What this therefore shows is that the paralleling is faulty. And this is the final insight into the dialogue that Teng provides from the perspective of Mohist logic. For as we saw in §2.2, the Mohists were clear about the dangers of analogical reasoning. Zhuangzi and Huizi were presumably aware of this too, and would not have been surprised that their paralleling leads to contradiction. This then motivates Zhuangzi’s going back to the ‘root’, which Teng accepts is a ‘trick’ but one which should now be seen as ‘a device for reorienting oneself to a different viewpoint and asking oneself whether one knew it all along’ (2006, p. 137).

Reading the dialogue from the perspective of Mohist logic does indeed help make sense of the form that Zhuangzi’s and Huizi’s reasoning takes. What is absent in both the dialogue and the ‘Lesser Selection’, however, is any diagnosis of where such paralleling goes wrong. Zhuangzi just returns to the ‘root’, and few resources are provided in Mohist logic (at least in what survives) for assessment: the Mohists simply identified (supposedly) good and bad cases of paralleling, without any systematic account of what made good cases good and bad cases bad. Teng suggests that once we recognize the paralleling in the dialogue, ‘Hui Shi’s response is both elegant and powerful from an ancient Chinese dialectical viewpoint’ (2006, p. 135). But it is nevertheless flawed, and Teng’s appeal to Mohist logic is insufficient for analysing the dialogue. We need to recognize that (ipp) is a stronger and more problematic principle than (pka), and that an option remains for Huizi to retreat to species-relativism.

5.5 The Species-Relativist Interpretation

What the logical analysis just given, drawing on both modern and Mohist logic, thus opens up is the space for a perspective that judges neither Zhuangzi nor Huizi as the ‘winner’. Confirmation of this is provided by the ‘species-relativist’ interpretation that has recently been advanced by Lea Cantor (2020). On her reading, Zhuangzi’s and Huizi’s shared experience in roaming together and their looking at the fish from the same location shows the possibility of their knowing one another’s states, but ‘we have no idea whether the happiness of fish would amount to anything outside our human purview’, as she puts it (2020, p. 226). Once the stronger principle—(pnac) or (ipp)—is shown to lead to contradiction, then the way is open for Huizi to recognize, with Zhuangzi, that what knowledge he has of fish happiness comes only from our human perspective. It is not being in the relevant state that is important, but having a certain perspective.

On both Hansen’s and Cantor’s interpretations, then, Zhuangzi is seen as exposing an assumption that blocks Huizi from appreciating what (human) knowledge of fish happiness is. But unlike Hansen, Cantor takes the dialogue as supporting species-relativism, not a stronger individual-relativism that just refrains from drawing absolutist claims. And this means that a better response to Zhuangzi’s counter-challenge was available to Huizi than the one he actually made. He could have retreated to the species-relativist principle inferentially invoked in his initial challenge, and hence found common ground with Zhuangzi. Perhaps that is what was intended in Zhuangzi’s talk of going back to the ‘root’: Zhuangzi and Huizi were indeed on common ground, looking at the fish, and Huizi did indeed know fish happiness in just the same way as Zhuangzi—whatever exactly that knowledge amounts to.

5.6 The Connective Perspectivist Interpretation

My own view is that Zhuangzi is not himself a species-relativist, at least about happy activity, even if Huizi is (charitably interpreted). The whole setting of the dialogue suggests a parallel between the fishes’ happy activity and the happy roaming of Zhuangzi and Huizi, and what Zhuangzi tries to do in the final line is to get Huizi to recognize the analogy: he can know happy fish activity as like their own happy human activity, taken as the standard. Their shared perspective and the ‘perspective’ of the fish open up to one another to the extent that there is this underlying similarity.

If this is right, then the deeper argument is implicit in the framing of the ‘logical’ dispute between Zhuangzi and Huizi, and here we might return to the contrast between knowing-that and knowing-as drawn in §4.5. Huizi interprets Zhuangzi as claiming to know that the fish are happy, but the attempts to justify this—or its negation—break down. So perhaps we should indeed switch to a conception of knowing-as, with corresponding modesty about claims to knowledge. Reflecting his perspectivism, Zhuangzi only claims to know fish happiness to the extent that it is like his and Huizi’s happiness, to which attention is drawn in the careful construction of the happy fish dialogue. Like all analogies, however, this may or may not be convincing, so perhaps in the end the dialogue simply expresses the basic problem of Mohist logic, concerning the evaluation of analogical and parallel reasoning. (“The softness of the analogical must.”) There may be no compulsion, but only an invitation to connect.

VI

The Dialogues as Analogues. Let’s go back to the beginning—the Chinese logic dialogue. Is the argument in the text that we have been considering—the happy fish dialogue—cleverly argued? This is the implication of both the logical analyses we considered, even if the argument needs considerable unpacking. But can it only really be judged so from the perspective of a Chinese logician, as Hugh claims? In other words, generalizing, can a text only be analysed—and the arguments evaluated—using the logical resources that we can reasonably assume that the author of the text had at their disposal? ‘Yes’ seems to be Teng’s answer in his dispute with Hansen. But, at least from my hermeneutic perspective, this seems wrong. Knowing what kinds of reasoning were used at the time, both implicitly and explicitly, is important for understanding any text, but the core of philosophical understanding is evaluation of the arguments, and in this respect neither Teng’s nor the ‘continental’ readings we considered do this. We need to draw on modern logic, which provides resources for dealing with a wide range of our inferential practices.

Of course, applying any logic only works if the relevant forms of reasoning and principles inferentially invoked are at least implicit in the text (or context). As I reconstructed Hansen’s analysis, this is arguably the case, although I would offer Huizi, more charitably, the retreat to the weaker, species-relativist principle that the analysis itself opened up. Most importantly, though, the kind of analysis that Hansen provides enables us to evaluate the argument. As we have seen, and this is where appreciation of Mohist logic does come in, analogical reasoning can easily lead us astray, which is precisely what happens in the argument between Zhuangzi and Huizi, as both of them implicitly recognize. But we need other forms of logic to identify and diagnose what goes wrong.

So is Joanne right? There are many ways of judging that something is cleverly argued. The continental readings we considered also make this judgement, but they do so by appreciating how the argument between Zhuangzi and Huizi is framed. The shared experience that Zhuangzi and Huizi are having comes out in the setting of the dialogue and the deliberate choice of words, which might be regarded as an implicit or ‘silent’ argument, which the reader is meant to make explicit or voice for themselves.⁶ So too, in the Chinese logic dialogue, it is relevant that Joanne and Hugh are discussing the passage together. It is through their discussion, assumed to involve conceptual and logical analysis—in other words, through the operation of their shared reason—that understanding of the relevant argument is achieved, just as the analysis of the happy fish dialogue offered in the present text is achieved through engagement with other interpreters and in response to actual comments on earlier versions and imagined criticisms by later readers. And this suggests a hermeneutic principle that is thoroughly Zhuangzian. Any interpreter must be open to as many different readings as possible. Each reading offers something, and we must endeavour to see how they connect with one another, as illustrated in the way that the interpretations were presented in §v.

My analysis of the happy fish dialogue has been framed by discussion of the problem of Chinese logic highlighted by the Chinese logic dialogue. The parallelism between the dialogues achieves two connected aims. The happy fish dialogue offers an instructive case study by means of which to think through and resolve the problem of Chinese logic. The Chinese logic dialogue provides a setting that brings both Mohist and modern logic into play in analysing the argument between Zhuangzi and Huizi, and appreciating the contribution they jointly make in understanding the happy fish dialogue. Mohist and modern logic offer different perspectives on the text: they both help open up the argumentation, but each one also opens out into the other.

Analogies encourage a form of perspectival thinking in which we roam between different perspectives in enriching our understanding of what it is that they are perspectives on. We construct our world through analogies, and the wider we roam in drawing them, the richer our thinking. At the core of Mohist logic is the idea that something is (deemed) what it is in being like something else that is our standard in this respect. The root form of knowing for the Mohists is knowing-as—knowing something as like something else. If analogies are turtles, then it’s turtles all the way down. And this is reflected in Zhuangzi’s view that knowing fish happiness is knowing it as like his and Huizi’s happiness. Analogy lies at the heart—the xīn (心)—of Chinese thinking, and we can enrich our own thinking by exploring analogies and parallels with Chinese philosophy and using logic to unpack and evaluate them. Analogies, both explicit and implicit, saturate the text you read before you, and I hope their felicity is manifest. How do you know this felicity? You know it from your happily swimming around in the text.⁷

Footnotes

References