In this chapter, we discuss the remaining necessary preconditions for interpretation, which includes further elaboration and integration of the necessary preconditions previously introduced.
•As discussed in chapter 5, the interpreter must assume that the other is a human being (the attitude-toward-a-soul principle). This is the individualistic formulation of the precondition that there must be mutually recognizable human practices (family resemblance between human forms of life). Without this fact or transcendental assumption, interpretation would not be possible. This precondition invites a further elaboration of the notion of quasi-universals with a limited range.
•Reflection on language shows that a necessary precondition for interpretation is the FR-principle (as elaborated in chapter 4). All concepts are FR-concepts. The FR-principle includes the NNSSL-principle and the possibility of FR-extension. The latter practice forms the basis of finding (deciding on) quasi-universals. In this chapter we present a substantial discussion of different variants of FR-extension.
•Perhaps the precondition for interpretation most often discussed in the literature is the principle of charity, which is necessary to curtail the underdetermination of interpretations caused by having to interpret numerous things at the same time. The interpreter must assume that the other is usually sincere and right, at least in the radical stages of interpretation.
•Interpretation is constrained by the background of the interpreter (hermeneutic relativity). Apart from the quasi-necessary constraints as discussed in chapter 8, the major precondition that ensues from hermeneutic relativity is that interpreters have no choice but to assume particular epistemic virtues, such as coherence with background knowledge and empirical adequacy, in deciding, for example, what quasi-universals to use or which FR-extensions to exploit.
•Interpretation is primarily something a group is involved in (instead of isolated individuals). Communicative interaction within, between, and among groups is subject to the principle of mutual attunement. The latter principle takes on different forms in different stages of interpretation (principle of reasonableness, principle of charity, principle of humanity, to be discussed below).
In a final section we discuss true/false, right/wrong, and shifei 是非, which illustrates the possibly needed revision of the formulation of the principle of charity and other fundamental principles of interpretation in intercultural philosophy.
Perhaps Leibniz was the first Western philosopher who explicitly addressed the question of necessary preconditions for interpretation.1 Leibniz proposed four principles of interpretation. First, one has to assume that rationality or reason is shared. Reason (i.e., logic plus some universal innate ideas, for example, the concept of substance) is “relatively universal and experience is relatively limited or perspectival.” Second, one should adhere to a principle of generosity; that is, one should interpret a text as something “reasonable and consistent,” giving it the most favorable interpretation. This principle is similar to the principle of charity. Third, one should give precedence to original texts, instead of commentaries; this is an epistemic virtue. Fourth, one should interpret alien concepts as similar to familiar concepts (principle of accommodation). The latter principle has some similarity to our notion of choosing quasi-universals. In the twentieth century, a range of similar principles has been proposed, among which the principle of charity and the hermeneutic principle of dialogical consensus are the best known.
Perhaps Zhu Xi was the first Chinese philosopher who proposed a number of principles guiding the methodology of interpreting classical texts. According to Pan Derong and Peng Qifu (2006), Zhu Xi distinguished between three levels of interpretation: “to master the meaning of the text,” “to understand the author’s original intention,” and “to develop a refined comprehension that is different from and yet based on the original meaning of the text” (135). The first step includes parsing the text (which contains no punctuation marks and only limited end markers). The three steps, understanding, interpreting, and applying cannot be separated (142). Zhu Xi also remarks that one must study the whole of the author’s thoughts, grasping both what he has said and how he has acted (139). Preconceptions (xianru zhi jian 先入之見), which will hinder understanding the sages’ meaning, should be suppressed.
The most basic preconditions for interpretation and communicative interaction are the soul principle and the necessity of their being mutually recognizable human practices. A human being has a learnable language, which uses classifications, is embedded in a community (human forms of life), and has a notion or feeling of right and wrong. This precondition is taken for granted in the practice of every form of intercultural philosophy. Although this principle is self-evident for almost every situation, it raises the question of “how much” universality common humanity entails. That is to say, which universals (if any) do all human beings share?
The notion of quasi-universal was introduced in chapter 2. We repeat the list of features of quasi-universals given there, adding illustrative examples, and extending it.
•Quasi-universals are family resemblance concepts. Like other FR-concepts, a quasi-universal has no core and is open-ended in its use (see chapter 4). However, not all FR-concepts are quasi-universals. For example, later in this chapter we argue that Being (Sein) is not a quasi-universal. There exists a family resemblance between the Being of, say, Aristotle and Heidegger, but there is no family resemblance with a classical Chinese concept (and in modern Chinese only with the help of a stipulated translation).
•Quasi-universals connect notions from a limited number of traditions by FR-extension. For example, the extension of FR(games) with respect to youxi 游戏 and the extension of FR(youxi) with respect to games display a family resemblance (see chapter 4). In this situation, we may say either that the quasi-universal belongs to a passing metalanguage used in the interpretative practice or that the extended FR(youxi) is the quasi-universal in Chinese and, reversely, that the extended FR(games) is the quasi-universal in English. Hence, quasi-universals are extended FR-concepts presupposed in a particular interpretative practice, while the interpreter is ready at any moment to drop this particular quasi-universal or to extend (or contract) it in unexpected FR-directions.
•Interpretation would not be possible if the interpreter could not be prepared to assume quasi-universals and to engage in FR-extension. Therefore, quasi-universals fulfill a heuristic and necessary role in the interpretative practice. Without assuming a large number of quasi-universals, interpretation across traditions is impossible. A particular interpreter may not be aware of this when, in fact using quasi-universals, is seeing connections.
•Quasi-universals are working hypotheses. They are constructed in interaction with the continuing process of interpretative practice. Quasi-universals may “work” pragmatically, even if embedded in different (meta-)CS’s on both sides.
•A metalanguage, when properly understood, is not a universal language. All words/concepts used in the metalanguage should be quasi-universals deemed suitable for the particular interpretative practice at hand. Hence, the application of a (passing) metalanguage is, like quasi-universals, limited to a few languages or traditions and will be under constant revision in response to the interactive process of interpretation.
•Data and background underdetermine the choice of quasi-universals in every particular case. Because of underdetermination, different interpreters may propose different quasi-universals with respect to a particular alien conceptual scheme, which leads to variations in preferred translations and expositions. For example, translation of classical Chinese into English or Japanese may draw on different quasi-universals. Even different Japanese (or English) interpreters may choose different quasi-universals. For example, the Confucian notion of li 理 may be approached from the side of English exploiting family resemblances with a variety of English concepts.
•Quasi-universals underlie the process of comparing (concepts from) different traditions. One can compare only after having assumed a number of quasi-universals in terms of which the comparata can be investigated.2 Hence the quasi-universals function as tertia comparationes.3 The most basic tertium comparationes is the mutual recognition of human practices (family resemblance of forms of life).
For interpretation to be possible, no universals in the sense of an ideal language sense need to be presupposed. The methodology of empirical research aiming at discovering human universals is disputable.4 Therefore, lists of human universals should not be trusted.5 Even if one would accept that “a small set of words occurs in every language,” such as BODY, PART OF, THING, EAT, one should be aware that this still covers or allows considerable differences of language (games) and forms of life (Moe 2013):
•The semantic range of these “universal” words differ in different traditions.
•Words belong to more than one domain and the salience of the respective domains vary across traditions.
•Domains are linked together differently in meta-CS’s. There are multiple layers of organization (which is difficult to represent in a traditional dictionary).
Given the assumption that the other is a human being and participates in human forms of life, what follows from this with respect to the issue of possibly shared quasi-universals? Like animals, humans eat, sleep, and can be aggressive. However, these biological needs and reflexes are embedded in very different practices and interpretations across traditions.6 Still, such biological universals are a necessary precondition for observing mutually recognizable human practices.
No doubt the most important human universal is that humans have a learnable language and live in communities in which having a language plays an indispensable role.7 Other languages qualify as such, not because of what they put into words (i.e., the content) is something that the interpreter might say, but because of what they do: their communicative function in conveying information and coordinating behavior, or some expressive function of language such as phatic communion. “We can tell that a language is being used, and even a good deal about how it is being used, short of any ability to translate” (Rescher 1980: 327). No concrete linguistic universals need to be presupposed. However, one may start from the working assumption that a language may be said to employ classifications and binaries, including evaluative judgments of what is said and done.8
Putnam writes:
If we succeed in interpreting a bizarre culture at all, we are committed to treat them as persons. To the extent that interpretation works we share objects, concepts, and at least some conceptions of the reasonable, the natural, and so on. (1981: 119)
Which objects and concepts does Putnam have in mind? Putnam’s “we share objects, concepts” is a (perhaps unintended) consequence of the ideal language paradigm. Strictly speaking, people do not share any concepts. The colloquial “sharing concepts” has to be understood as the possibility (and necessity) of both sides extending their FR-concepts to cover practices on the other side. As to conceptions of the reasonable and the natural one may perhaps appeal to family resemblances or quasi-universals across traditions; but one should realize that these resemblances may be rather superficial from a philosophical perspective. For example, compare physis (φύσις), and ziran 自然 in their various meanings (which also shows that meanings are already evanescent within one tradition). If “the reasonable” is assumed to be a quasi-universal, it would have to include not only all varieties of (Western) reasonableness, but also ihuma and Confucian reasonableness.
One can say, as Cheng Chung-ying does: “There is consensus that the human person has many mental functions and abilities recognized as thinking, reasoning, feeling, desiring and willing. If we deny these mental functions and abilities, we would deny humanity” (2000: 34). However, having said this, we may observe that qing 情 in the classical Chinese discourse does not easily map onto Western mental functions. Therefore, one may as well say that denying qing amounts to denying humanity.9 It may be true in today’s English that having mental functions such as “thinking, reasoning, feeling, desiring and willing” is a defining feature of being human. However, it does not follow that this truth can be easily translated in the language of an “other” tradition.
FR-concepts that often allow for FR-extension to other traditions and hence often function as quasi-universals in the practice of interpretation include:10
•quite a few words related to animals, plants, body parts, and basic tools;
•some ordering of space and time (uphill/downhill, last year);
•making distinctions (this, not-this, the other);
•mathematical logic (one, two, three; words such as “and,” “all,” … );11
•some qualities (large/small), including evaluation (right/wrong);
•classifying humans (we, they, people).
Such quasi-universals must be projected in primordial stages of interpretation and it is possible to give preliminary descriptions of human practices or translations of foreign texts in terms of such quasi-universals, provided that one realizes that the specification of every quasi-universal is embedded in the (variable) background of the interpreter (hermeneutic relativity). The interpreter should have a sixth sense for spotting exceptions relative to his or her expectations. The above list of quasi-universals is not one of universals. For every pair of traditions a different list would have to be constructed (and would change over time). Further investigation will quickly show that a particular quasi-universal harbors many differences of detail, if not differences “in the large.” This already starts at the basic level of claiming family resemblance of FR(big) and FR(da 大) or FR(see) and FR(kan 看). Moreover, it is possible that what the interpreter thought to be a clear case of quasi-universal (that works to connect two traditions) cannot be found in yet a third tradition. Further, at later stages of interpretation, differences in assumed quasi-universals may turn up. For example, both in Chinese and in African philosophy there have been discussions about whether causation can be considered to be a quasi-universal.12
We will use the symbols FR(E) and FR(C) as labels for particular FR-concepts in English (and other Western-European languages) and Chinese languages respectively.13 In the simplest case, FR-extension means FR(C) ⇔ FR(E), with ⇔ indicating a relation of family resemblance (not identity or equivalence). That is to say: FR(C) shows similarities with FR(E), as judged from E’s conceptual schemes; FR(E) shows similarities to FR(C), as judged from C’s conceptual schemes. Such FR-extensions support quasi-universals, for example FR(love) ⇔ FR(anpu) and FR(youxi 游戏) ⇔ FR(games). Each of FR(C) and FR(E) can absorb part of the other by FR-extension. The extended FR(C) is the quasi-universal as seen from the Chinese side; and vice versa for FR(E). Rather simple quasi-universals may also occur at a higher level of abstraction, for example, observing the family resemblance of color and yanse 颜色 (a notion from modern Chinese). FR(color) ⇔ FR(yanse).14
A more complicated (and therefore more disputable) example is the pair philosophia (φιλοσοφία) and zhexue 哲学, discussed in chapter 1. Some scholars argue that there is no family resemblance between the extension of these two terms. In saying this, other quasi-universals need to be presupposed. One may also argue that both sides should be capable of constructing a quasi-universal, starting from their own conceptual schemes. The quasi-universal, FR(philosophia) ⇔ FR(zhexue), will supervene on the interplay and extension of conceptual schemes on both sides.
Basic or simple cases are disputable, as it may be argued that important differences have been suppressed, or irrelevant similarities have been highlighted. Sometimes, it may have heuristic value to start a comparative investigation provisionally assuming even implausible quasi-universals.15 However, more often alleged quasi-universals are too easily projected across traditions (see transcendental pretence in chapter 8).
Whatever the variant of FR-extension, the interpreter cannot avoid using a metalanguage that pretends to be universal. This can be seen in almost every publication in intercultural philosophy. For example, Fraser (2011a) provides a detailed discussion of “knowledge and error in early Chinese thought.” In delineating this subject matter, he is assuming that the Western notion “know(ing)” has some similarity to the Chinese notion zhi 知, suggesting: FR(zhi) ⇔ FR(know). In highlighting the differences (and similarities), Fraser employs a metalanguage, which is partly based on terminology of stipulated translations of Chinese expressions such as FR[lei 类 (kind, type)] and FR[fa 法 (model, standard)], and partly on Western terminology (expressions such as mind, competence, error, part-whole relations), which are assumed to be quasi-universals in terms of which the subject matter is addressed.
In slightly more complex cases, FR(C) can be divided into a number of different functions or uses (meanings), each of which has a family resemblance to a particular FR(E). In such a case, we say that FR(C) in Chinese appears as a pseudo-homonym, FR(En), in English where it is assumed that each occurrence of FR(C) has a family resemblance with one of the FR(En). We name it pseudo-homonym because it is only in English, for that matter, that a range of meanings, FR(En), form something like a halo around the head of Buddha, whereas in the original Chinese, FR(C) is simply one word and one concept. A simple example is [FR(shu 鼠)] ⇔ [FR(mouse) and/or FR(rat)]. A philosophically more relevant example is the English verb “to be.” In English to be is one (polysemous) concept, FR(E). From the side of Chinese it may seem that “to be” is a pseudo-homonym, FR(Cn). As we have seen in chapter 8, “to be” is not a universal, but for the different types or functions of English “to be,” FR-extensions seem to be possible.16 For example, FR(existence) ⇔ FR(you) is a simple case of FR-extension.17
There are various ways in which FR(C) can be mapped onto a range of FR(En) or FR(E) onto a range of FR(Cn). In addition to pseudo-homonyms, FR(Cn) or FR(En), we may envisage the possibility that a number of concepts have to be gathered together into one new hybrid concept, FR(ΣCn) or FR(ΣEn).
Most Chinese and foreign language dictionaries suggest that many Chinese characters fit the pseudo-homonym model: FR(C) has different meanings in different contexts correlated with FR(En). The dictionary offers for one Chinese character, FR(C), a number of different English translations, FR(En), applicable to different contexts. From the English perspective it seems that the Chinese character is a homonym. However, it is possible to interpret the dictionary entry differently. Insisting that the Chinese character has one (although complex) meaning, it is speculated that this meaning is a kind of mix of all or many of the English translations offered in the dictionary. That is to say, we may envisage the possibility that the meanings of a number of concepts in the interpreter’s language have to be gathered together and conveyed into one new hybrid concept, not FR(En), but FR(ΣEn).
We believe that the notion of a hybrid concept is a rather undervalued option, although the idea has often been mentioned in passing. Richards (1932: 4) speaks of “blending” of meanings with respect to “ore-like” Chinese characters. In a somewhat different context Hall and Ames speak of concept clusters and write: “Accepting terms as cluster concepts [i.e., hybrid concepts] means accepting their meanings in a single gestalt in which all meanings are potentially foregrounded.”18 Although not stated explicitly, Hall and Ames may consider the following an illustrative example: “the xing [性] of self seems to overlap with ‘character,’ ‘personality,’ or ‘constitution,’ as well as the initial conditions of one’s birth” (190). They see this description of xing as a replacement of logical coherence by aesthetic coherence (124–125). That is to say, they restrict this use of clusters of images to correlative thinking. We propose constructing hybrid concepts as a general feature of good interpretation.
Hybrid concepts may also be considered intermediaries.19 These merged or blended or mingled concepts are like other concepts, except that they have no name yet in the tradition in which they are introduced. They are constructed to get access to a concept in the other tradition for which there is no suitable concept to start an FR-investigation. Merging concepts amounts to making new FR-concepts in the interpreter’s language that can serve as quasi-universals.
An example of blending concepts might be the following. Instead of translating Confucian cheng 诚 in different contexts as sincerity, reality, faithfulness, truth, integrity, earnestness, genuineness, creativity, self-realization, or some other concept from the Western philosophical tradition, one may try to express something “intermediate,” a new hybrid concept from the perspective of the interpreter.20 Hence, FR(cheng) is not a pseudo-homonym, for which FR(cheng) would be FR(En), but is correlated with FR(ΣEn). In such a case one may use “cheng” as a word in the English language, except that the concept cheng is different from the concept cheng 诚 [誠]. The interpreter claims, having provided an elaborate elucidation of FR(chengE), that there is a family resemblance between FR(chengE) and FR(chengC).21
This looks like a case of leaving cheng untranslated. However, it should be distinguished from the case of using cheng in the English language, explaining its meaning in Chinese terms for which “reasonable” quasi-universals (or hybrid concepts) have been secured. The meaning of cheng is the meaning of a Chinese concept to be explained in its own (Chinese) terms, which may or may not be accessible by quasi-universals. In contrast cheng is a newly constructed English concept for which FR(chengE) ⇔ FR(chengC).
Making the distinction between hybrids and pseudo-homonyms is a matter of interpretation and degree.22 For example, there are different accounts of the meaning of qing 情 in classical Chinese. According to Puett (2004), from the beginning the term had a broad semantic range, which gradually developed into a refinement of distinctions. Only by the early Han did these different meanings come to be clearly distinguished. Puett argues that Hansen is wrong to see a “single, unified meaning” (38). One could model the disagreement between Puett and Hansen in terms of FR(qing) ⇔ FR(En) versus FR(qing) ⇔ FR(ΣEn).
Another variant of FR-extension is the use of clusters of concepts. In some cases the interpreter may see a family resemblance between FR-clusters from each side, CS {Cn} and CS {En}, without there being a pairwise connection between individual concepts. An example is: CS {ran 然, dang 當, zhen 真, cheng 诚, … } ⇔ CS {genuinely so, is true, be the case, right … }. This example is discussed later in this chapter. Speaking more generally: a cluster of CS {En} can be seen to have an FR-relation with a cluster of CS {Cn}: CS {Cn} ⇔ CS {En}.23 Comparing the location of ran and “is true” in their respective clusters may help to evaluate FR(ran) ⇔ FR(is true).
The construction of a suitable CS as well as the similarity judgment in saying CS {Cn} ⇔ CS {En} vary among interpreters. Whether a complex case can be sensibly reduced to a simpler case is a matter of discussion among interpreters. Some interpreters may be happy to state for modern Chinese: FR(grief) ⇔ FR(ai 哀). Others may emphasize the difference of grief and ai and may want to compare one cluster including grief and one including ai.24 Similarly, does it suffice to assume that, as regards the Zhuangzi, FR(wuyong 无用) ⇔ FR(useless)?25 Or should one consider a broader cluster in a modern language to cover wuyong; for example, including FR(not necessary), FR(not needed), FR(unprofitable), and perhaps more? On the side of wuyong, should one consider the cluster of FR(wuyong) to include FR(buweiyong 不為用) and/or FR(buyong 不用)? Ideally, because of the holism of language, one should compare clusters as large as possible, but this is practically not possible and interpreters have to decide on a critical mass, which may involve just one FR-extended concept on either side.
Another use of clusters is to embed a foreign concept among a cluster of FR-concepts, which have already been “accessed” by FR-extension. For example,26
qing 情 ∈ CS {xin 心, xing 性, yu 欲; qi 氣, li 理, li 禮; shi 事, shi 實, shi 是, cheng 誠,gan 感, huai 懷, duan 端}
Qing is left untranslated and explained as far as possible on its own terms.27 Qing is situated in a network or conceptual structure of (classical) Chinese concepts, presupposing that the characters listed are accessible by quasi-universals (if needed, to be constructed for the case at hand). For example: FR(xin 心) ⇔ FR(embodied mind), FR(xing 性) ⇔ FR([human] nature), and so on. The meanings of the Chinese characters can be assumed to be all connected (i.e., under one polysemous concept) or may perhaps be divided into two or three groups (i.e., “seeing” two or three different concepts from the interpreter’s side).28
Instead of embedding qing in a cluster of Chinese concepts, we can also embed qing in a cluster of possible translations of qing into English. For example:29
qing 情 ∈ CS {real/basic facts, principles/essentials, passions, virtues, sentiments (basic instinctive feelings), desires, sensibilities ascribed to sages, state of perfect genuineness, ultimate motives, deep convictions, authenticity (of a person), emotional response to situations, distinguishing right from wrong}
Qing is analyzed as a polysemous concept (whether hybrid or homonym) on the basis of proposed translations into English in a variety of contexts and genres.
The phrase “on its own terms” is used frequently in discussions about Chinese and other non-Western traditions, but has rarely been analyzed. We would like to emphasize that there is no such thing as explanation (understanding, letting speak, etc.) on its own terms, strictly speaking. In the case of classical Chinese texts it would mean being able to think and to write in classical Chinese. That is to say, one is expected to explain the meaning of characters or words in the way classical literati or dictionaries such as the Shuowen Jiezi did. However, then the connection with modern discourse is severed. An interpreter working with a modern language, even if familiar with the relevant embedding concepts of an older or dead language, remains tied to this modern language.30 Setting characters among their “own” concepts has to stop somewhere. Explaining qing on its own terms remains relative to the existence of some quasi-universals for which reasonable FR-extensions are hypothesized. Eventually, the connection with the interpreter’s language has to be made directly; for example, by claiming something like: FR(xing 性) ⇔ FR([human] nature), FR(yu 欲) ⇔ FR(desire), and so on. One may set out to explain, say, xing 性 on its own terms, but this explanation will depend on assumed quasi-universals that connect the classical and modern language. This requires a hybrid metalanguage (as used in most publications on Chinese or other non-Western philosophies).31
A slightly different approach is to analyze the meaning of a character in a Chinese text in terms of characters that are used in that very text. For example, Kim (2014) aims to explain Xunzi’s views on the “ideal state for humans” by discussing the meaning and interrelation of, primarily, zhi 指, li 理, he 和, and yi 一; and further concepts such as lun 倫, fen 分, zhidao 治道, dali 大理. The broad context of the discussion is given by (English) phrases such as “ethical thought,” “ideal state for humans,” “personal interactions.” Translations and interpretations of the relevant expressions from the Xunzi are discussed in English, which presupposes comparison of classical Chinese concepts and modern English concepts. It is true that something is said, in English, about the interrelationship (in Chinese) of the characters mentioned. However, in the end, Xunzi’s “own terms” are explained in the “own terms” of modern (philosophical) English.
The focus on relations between Chinese concepts in the original texts is a major improvement compared with interpreting a text already translated in a modern language (which translation is taken for granted), but it would be mistaken to think that one can let these own terms speak for themselves without the interference of quasi-universals. The interpreter must take responsibility for the choice of numerous quasi-universals.
Leaving all alien terms untranslated is not a translation. On the other hand, it is not good either to leave no term untranslated. Being too eager to translate leads to statements such as: “sincerity is the most important moral precept that underpins, motivates, and governs the thought and action of a Sinic soul” (Jung 2013: 473). What the author means is that cheng 诚 “is the moral fiber that produces the fabric of Confucian philosophy” (ibid.). It is not right to suggest that the English concept of sincerity plays such an important role in Confucian philosophy.
The right balance has to be struck between translated and untranslated terms. Translation requires quasi-universals, for which FR-extensions across traditions are or can be made available (perhaps indirectly by providing a good translation: as the text in translation progresses, some English words extend their meanings). A translator always has many options. Here is an example from the European traditions, that illustrates various options for the translator in how to translate Heidegger’s word Ereignis into English: leaving it untranslated (Ereignis), use an everyday word (“event”), trying to grasp Heidegger’s intended meaning in English (“Event of Appropriation”), or introducing a neologism that can absorb whatever Heidegger says about Ereignis (“Enowning”). All these four options have been defended. The frequent occurrence of an untranslated term in explanatory sentences of the interpreter’s language, which can be called a Wittgensteinian perspicuous presentation (übersichtliche Darstellung), embeds the untranslated term by the quasi-universals the interpreter has selected.
On the side of classical Chinese, qi 氣 [气] is a character that is often left untranslated. For example, at one point Graham “defines” qi in the context of Chinese cosmology as “a universal fluid, active as Yang and passive as Yin, out of which bodies condense and into which they dissolve” (1989: 101).32 The alien concept is not translated, but it is given a “best” description in terms of the interpreter’s CS’s. However, Graham assumes concepts such as fluid, bodies, condensation, and dissolve are quasi-universals (which is by no means self-evident), whereas Yin and Yang have already been added to the English language (assuming an educated reader will already have a faint acquaintance with these Chinese concepts).
Alternative translations were discussed in China on a larger scale, when Western terminology was introduced from the late eighteenth century onward. For example, the current translation of the word logic is luoji 逻辑[邏輯], a phonemic loan. As the comprehensive study of Kurtz shows, there have been extensive and sometimes heated discussions in China concerning alternative translations for “logic” that were proposed, first by Jesuits (1623–1683), then in translations of Western works on logic (such as translations of Yan Fu) and in numerous other publications from the period 1869 to 1921. Apart from a variety of phonemic loans, more than twenty-five translations of the word “logic” were proposed. However, most of them were variations drawing on a limited number of Chinese characters, in particular ming 名 (name), li 理 (principle, pattern), and bian 辨 (distinguish, discriminate), and to a lesser extent si 思 (think) and tui 推 (push, prove, etc.).33 Some of these translations, for example bianli 辨理 or libian 理辨, might be considered newly constructed hybrid concepts or intermediaries, “finding or creating a lexical and conceptual ‘in-between’” (Kurtz: 341). Instead of debating alternative translations, which are all unsatisfactory, it may be more fruitful to investigate how the (Western) concept logic might be fitted into a family resemblance with a cluster of Chinese concepts, starting with,
logic ∈ {bian, li, ming, si, tui, … }
to which more characters would have to be added from the numerous proposed translations for logical terminology (see the nineteen tables in Kurtz 2011).
In contemporary English literature on Chinese philosophy it is common (and unavoidable) that numerous quasi-universals are assumed without justifications being given for FR(C) ⇔ FR(E). As Graham remarks: “There are no exact equivalents; if we have translated many important Chinese words without discussion, it is only because the differences cannot be profitably explored within the limits of the present book” (1989: 124). However, it is important to add that Graham does not exclude the possibility that the suppressed differences may become important at a later stage or different focus of the investigation.
Does leaving a term untranslated show that there is no similar concept in another tradition? Perhaps. But note that one cannot claim the absence of a particular concept in another tradition without already knowing a lot about this tradition. Saying that classical Confucianism “had no equivalents for such terms as: freedom, liberty, autonomy, individual, principles, rationality, rational agent, action, objective, subjective, choice, dilemma, duty, or rights,”34 entails that one already knows a lot about Confucianism, and has drawn on numerous FR-relations across traditions.
We suggest that so-called literal translation is to be preferred for “easy” quasi-universals, for example, numbers, because the connotations of a particular number, say “nine,” or “seven,” may possibly be relevant for interpreting more difficult parts of a text. It is not good practice to translate, say, a particular distance by converting the alien standard of measurement, giving the measurement in feet or meters (depending on the intended audience). For example, it is safer to report that the lowest branches of the sacred oak (櫟社樹) in chapter 4 of the Zhuangzi are ten ren 仞 from the ground than to translate it as seventy feet (Watson) or eighty feet (Graham). This is because “ten” may convey meanings other than numerical connotation. In such a case, it is better to leave the alien unit of measurement untranslated so that the numbers do not change. Not only “all translation equivalents of philosophical concepts are inadequate” (Graham 1991: 287), all translation equivalents of everyday concepts are inadequate as well.
In this section we propose the principle of mutual attunement as a further necessary precondition for interpretation. We build on major insights from Davidson concerning the role of the principle of charity in radical interpretation (but do not agree with his major underlying assumptions) and we borrow some formulations from the competing principle of humanity. Therefore, we first critically review the relevant literature concerning these two principles (charity and humanity), which, henceforth, will be referred to by PC and PH in this section.
There is a vast secondary literature on Davidson’s principle of charity, but a generally accepted formulation of the principle is lacking, either in the secondary literature or in Davidson’s own writings.35 LePore and Ludwig gather together more than ten passages in Davidson’s Inquiries into Truth and Interpretation (2001 [1975]), each of which presents the principle of charity in a somewhat different way (LePore and Ludwig 2005, 182–196). They distinguish between four variants: Veracity, Charity, Agreement, and Grace. According to LePore and Ludwig, Charity (or Grace) is the principle Davidson would need as a necessary precondition for his theory of interpretation. On one occasion, Davidson formulates the necessary precondition as “assigning truth conditions to alien sentences that make native speakers right when plausibly possible, according to our own view of what is right” (Davidson 2001: 137). By contrast, LePore and Ludwig formulate their PC (Grace) as:
For any speaker S, time t, belief b, ceteris paribus: b is a belief of S’s at t about and prompted by S’s environment if and only if b is true. (189)
We don’t think one should include the environment in the formulation of the principle. Davidson is right to emphasize that radical translation or interpretation begins with applying PC to observation sentences. But at later stages PC still applies, especially to nonoccasion sentences, unless it is overruled by a principle of humanity (see next subsection).
Our concern is not with the precise formulation of PC in order to make Davidson’s theory of interpretation logically consistent and complete. We are interested in a more intuitive notion of PC, which is sufficiently comprehensive to be applicable to all interpretation (and theories thereof). The intuitive aspect of PC is that one can have disagreements with the other only against a background of a host of agreements; one can misunderstand the other only against a background of understanding in the main. Furthermore, as Davidson remarked, the agreement has to be the right sort of agreement (2001: xvii). Some disagreements are more destructive to understanding than others.36 It is wrong to formulate PC simply as a principle of maximizing agreement or minimizing error,37 because this suggests the misleading idea that (dis)agreement can be quantified.
Often Davidson formulates PC only with respect to truth.
Interpret what the speaker accepts as true as true when we can. (2002: 149)
The principle directs the interpreter to translate or interpret so as to read some of his own standards of truth into the pattern of sentences held true by the speaker. (2002: 148)
According to Davidson, one can “take it as given that most beliefs are correct” (2001: 168), because belief is never singular. There is never just one belief, but always a pattern of beliefs (web of beliefs and conceptual schemes). To identify a false belief, surrounding beliefs have to be correct; only then can the false belief make sense.
Occasionally Davidson uses the word right instead of truth:
Whether we like it or not, if we want to understand others, we must count them right in most matters. (2001: 197).
On some other occasions, he explicitly includes (logical) consistency (in addition to truth) into the PC:38
The basic strategy must be to assume that by and large a speaker we do not yet understand is consistent and correct in his beliefs. (1980: 238)
At one point he even says there are two PCs, namely the Principle of Coherence and the Principle of Correspondence (2002: 211).39 Only on two rare occasions does Davidson include “the good” in PC:
We will try for a theory that finds him consistent, a believer of truths, and a lover of the good (all by our lights, it goes without saying). (1980: 222)
Some kind of basic agreement, not just in beliefs, but also in values, is essential for understanding. (Davidson, in Bergström and Føllesdal: 220)
In Davidson’s writings, and perhaps even more so in the secondary literature, the emphasis always falls on assumptions concerning truth. Sometimes consistency is added, but values are never mentioned. For Davidson himself, this is explainable because he takes “truth” to include “the good” and values.40 But the latter view is not commonly shared. Therefore, we emphasize that “the good” is better explicitly included in the formulation of PC (even if it might be superfluous according to Davidson).
Note that PC does not require that X and Y speak the same language. If they speak different languages, two sets of agreement would be needed, because the two languages may employ different classification systems. On both sides, there has to be “a background of largely unmentioned and unquestioned true beliefs” (2001: 168) on which both sides agree (but usually for different reasons).41
Davidson’s principle of charity has sometimes been associated with Gadamer’s notion of anticipating completeness or perfection (Vorgriff der Vollkommenheit) as a counterforce to the potential arbitrariness or idiosyncratic character of interpretation. But these two notions do not pair up with one another, because Davidson’s principle of charity primarily applies to the primordial “radical” stages of interpretation.42
The most common critique of Davidson is that a charity that forces the others’ beliefs into a mold of the interpreter’s patterns of belief is a false charity, a form of “linguistic imperialism” (Hacking 1975: 149), making it into a “principle of patronisation” (MacDonald Ross 1985: 500). However, this kind of criticism is based on a misunderstanding of the limited range of Davidson’s theory of interpretation. Davidson is looking for the minimally necessary preconditions for interpretation to be possible. He is right to argue that in the situation of radical interpretation, there is no other option than adhering to something like PC.
Another critique of Davidson’s claim that only the principle of charity suffices is that successful communication requires some Gricean principles of cooperation and relevance. Because language can have functions other than communication of truths, translation cannot always be charitable. As Wilson and Sperber remark: “Verbal communication is governed not by expectations of truthfulness, but by expectations of relevance, raised by literal, loose, and figurative uses alike” (583).43 This may be correct for everyday communication between speakers of the same language, but that is not where Davidson’s principle is thought to figure. Moreover, Gricean norms are already part of human practices even before verbal language is understood, as first contacts illustrate. A Gricean norm such as efficiency may well qualify as a biological universal across traditions, as it is included in the necessary conditions for survival (which humans share with animals).
With a number of provisos, Davidson’s PC will be integrated in our proposal for the principle of mutual attunement.
We have pointed out that PC applies solely to the primordial stages of linguistic interpretation. In subsequent stages of interpretation, it would be misplaced to still employ it dogmatically. PC may be overruled by the principle of humanity.44 The assumption of consistency may remain if the (quasi-) universality of logic is presupposed, but the assumption that, on the whole, the other says what is true or right may now be understood in terms of the standards of the other. Note that one needs to have passed the PC stage in order to find out what the standards of the other are. Hence, PH is not applicable in the radical stages of linguistic interpretation, but may overrule PC in later stages, when there is a PC-shared background relative to which one can formulate criteria of rightness in the other’s terms and interpret authors and speakers accordingly.
Formulations of PH differ widely (as to content). A representative formulation is:
Choose truth-conditions in such a way as to minimize inexplicability of attributed beliefs. (Bar-On and Risjord 1992: 171)
This entails, among other things, that X’s beliefs, which are different from Y’s beliefs, are explainable if they conform to the criteria prevailing in X’s community. In the early discussions concerning the correctness of either PC or PH, one issue was how to cover the following types of situations in which truth is not shared (ibid., 172):
•beliefs that are patently false by the interpreter’s lights;
•beliefs that conflict with the rest of the attributed pattern of beliefs;
•beliefs that are radically different from the interpreter’s.
This should not be problematic. Assuming that such situations are well-supported by the evidence, ascribing aberrant beliefs is not excluded by PC, for the simple reason that PC says that shared truth must be assumed on the whole or largely. One can only spot aberrant beliefs against a background of many shared true beliefs. PC has to do its work in order for the odd situations to appear.
Grandy is always cited when the discussion turns to the “origin” of PH, but his formulations are rather vague and general. For example: “The principle of humanity urges us to bear in mind that the speaker is a person and has certain basic similarities to ourselves when we are choosing between alternative translations” (445). What are these “basic similarities”? They may range from “innate similarity standards” to “mutually recognizable human practices.” Some formulations of PH are more ethnocentric than PC, as when Grandy is saying about PH that it requires “the condition that the imputed pattern of relations among beliefs, desires and the world be as similar to our own as possible” (445), which is not very different from the practice of transcendental pretence.
Sometimes, as an alternative to PC or PH, the label “principle of rationality” is advocated. When principles in one way or another appeal to “rationality,” we take it to mean one should be consistent in terms of logic and in terms of having effective behavior in reaching recognizable goals.45 A principle of rationality may also include a principle of explicability: maximizing explanatory cogency (according to the criteria of the interpreter).46 In this sense, principles such as humanity and rationality merge with criteria for good interpretations and may be considered to be (meta-)epistemic virtues.47
Some formulations of the principle of humanity or rationality apply to the very early prelinguistic stage of interpretation. For example, the attributed beliefs must be understandable ones for humans to adopt in the relevant circumstances. At this stage, the criteria of relevance are those of the interpreter. In the prelinguistic stages of interpretation, there is no other choice than assuming that the local Umwelt is partly shared with the other. One has to assume that the other person has mostly reasonable beliefs (by the interpreter’s standards) given its (ascribed) Umwelt. We call this the principle of reasonableness. In the abundant literature on interpreting a language, there is little interest in the prelinguistic stage of interpretation.
As it happens, Davidson himself seems to have changed his mind under the influence of Lewis’ (1985) critique.48
Charity prompts the interpreter to maximize the intelligibility of the speaker, not sameness of belief. This entails, as Lewis says, that interpretation must take into account probable errors due to bad positioning, deficient sensory apparatus, and differences in background knowledge. (2001: xix).
This change of view is apparent from Davidson’s use of the word intelligibility when reminding his readers of his principle of charity; for example, “to make an agent as intelligible as possible, given his actions, his utterances and his place in the world” (2001: 152; cf. 148; 2004: 34). However, distinguishing PC and PH is important for the linguistic stages of interpretation.49 From the start, PC should be understood as “if I were in her or his place” (often presented as a PH-feature).
Perhaps as a consequence of Davidson’s change of mind, the secondary literature is not very consistent in its use of the phrases PC and PH, although they are always seen as alternatives, either the one or the other being favored. For example, Hansen favors a principle of humanity that he formulates as: “most utterances [should] be explainable as warranted for the community” (2007: 481). According to Hansen charity would allow one to assume too much shared truths; with the result, for example, of avoiding the plausibility of concluding that Gongsun Long’s bai ma fei ma is false (but explainable). Hansen is right that one should not assume “too much” agreement and ascribe “too much” truth, but Davidson is right to say that in the radical stages of interpretation, PC is the governing principle. PH should refer to later stages of interpretation, when one has gathered, using PC, knowledge of X’s situatedness in the conceptual schemes of his time and place.
The debate about favoring either PC or PH is misdirected. Both are applicable, though at different stages of interpretation. Davidson is right that at the most fundamental level of interpretation of language, that is to say: in the primordial stages of radical interpretation, one must presuppose agreement in a large number of judgments, as judged by the standards of the interpreter. But in later, more pragmatic stages of interpretation, having a basis of agreement for interpreting the standards of the other, the interpreter can start to judge by the other’s criteria. That is to say, as the interpretation of X advances, Y may also start to interpret by X’s lights (as understood by Y’s lights).
We borrow the phrase “mutual attunement” from Cavell, who uses it when explaining Wittgenstein’s PI §241 and §242, where Wittgenstein says that participating in forms of life entails that this group of people “agree in the language they use” (PI §242). In these paragraphs Wittgenstein is not thinking of the intercultural situation, when different groups of people use different languages, but we can extend the idea as expressed in the citation as follows. Human beings are “in agreement throughout, being in harmony” (Cavell 1979: 32) concerning what it means to be human, to live in communities, to have a language, and to be aware of family resemblances of human forms of life as displayed in the mutual recognition of human practices.50 This “agreement” can be seen as a “thick” extension of the attitude-toward-a-soul principle.
In the prelinguistic stage of interpretation (as in first contacts) the principle of mutual attunement takes the form of the assumption that Y should ascribe those beliefs to X that Y would have were Y in X’s place (principle of reasonableness), which remains in force during linguistic stages of interpretation.
In the primordial “radical” stage of interpretation, the principle of mutual attunement is similar to Davidson’s principle of charity, but it differs on three important points. First, for Davidson, logic, and a behavioristic notion of belief (“holding true”) and some other matters are assumed to be universal across all humans and their languages, and therefore do not need to be interpreted. In our view, nothing need be exempted from interpretation, not even logic.51 No specific universals are needed (as Davidson assumes in his highlighting of triangulation). Second, Davidson’s principle is only concerned with one-way interpretation, and the roles of interpreters and speakers or authors are more or less fixated. Finally, our revised principle of charity includes more than Davidson’s consistency, truth, and rightness. In addition one must presuppose that the other is sincere (or trustworthy), and perhaps more. We revise PC as follows: One must presuppose that the other is, on the whole, sincere and right. Being right includes having true beliefs concerning both “facts” and “values.” A degree of emotional and moral attunement is as necessary for successful communication and interpretation as pragmatic agreement on the happenings in the local Umwelt(s). There must also be some agreement on meta-epistemic virtues such as coherence (“fitting”) and criteria for deductive and inductive inference. The revised principle of charity includes cases of agreeing that two incommensurable statements (in two different languages) are both true.52
Like Davidson’s principle of charity, the principle of mutual attunement refers to “harmony” in a large number of cases. However, these “cases” are not really countable, and the criteria for identifying “single” beliefs are by no means clear. One may easily overlook the “large number,” because many obvious or trivial judgments are included. The large number applies to everyday matters in particular, but it includes FR-similarities for reflective concepts as well.
In later stages of interpretation, attunement is not meant in the sense of aiming to reach consensus, but as mutual recognition of differences against a background of FR-extended quasi-universals. The similarities and differences are not the same on both sides. It is assumed that X and Y are ready to continue the interaction, but this is not the same as aiming at mutual Verständigung, in the strong sense of Habermasian mutual understanding or Gadamer’s fusion of horizons. We agree with Habermas (2000: 346) that the whole point of linguistic communication is to achieve mutual understanding.53 However, to speak of understanding only makes sense provided that it is tied to “local” (meta-) conceptual schemes on both sides, many of which show sufficient family resemblances, but are not identical. When interpreting texts, there happens attunement among interpreters and commentators, but again, one does not primarily aim at reaching consensus. There is no singing together in concert, but fine tuning of hermeneutic circles across traditions relative to an indefinite variety of background CS’s of interpreters and commentators.
In summary, we propose a principle of mutual attunement, which takes on a different form in subsequent stages of interpretation. Earlier items in the list below remain in force through later stages of interpretation.
1.Interpreters start off from the principle of “attitude toward a soul.”
2.In the prelinguistic stage of communication, as in first contacts, it must be assumed that the other has reasonable beliefs (according to the criteria of reasonableness of the interpreter), given the other’s Umwelt (as perceived by the interpreter). This is a revised formulation of a variant of the principle of humanity and might be called principle of reasonableness to distinguish it from PC and PH for the linguistic stages of interpretation.
3.In the primordial “radical” stage of linguistic interpretation, one must presuppose that the other is, on the whole, sincere, consistent, and right. This is a revised formulation of Davidson’s principle of charity. Having to assume that the other is right most of the time includes being right with respect to what is true and what is good.
4.In later stages of interpretation, the other’s beliefs are also judged according to the criteria of the other’s background. The interpreter needs PC to be able to know or guess what the criteria of judgment are in the other tradition. This is a reformulation of the principle of humanity governing linguistic interpretation.
5.In advanced stages of interpretation—the usual situation in intercultural and Chinese philosophy—all four aforementioned principles apply and their relative weight depends on the choice of quasi-universals (i.e., the lexicon of the interpreter’s metalanguage) and epistemic virtues (for evaluating alternative interpretations).
A “thick” principle of mutual attunement (as hinted at in the first paragraph of this subsection is operative in all necessary preconditions, constraints, and other features of interpretation. Such subsidiaries of the principle of mutual attunement include:
1.There is mutual attunement of humans and their local Umwelt (which includes other humans). Invariants constrain the course of events in the system of human(s) cum environment taken together as a whole.54
2.There is mutual attunement in all variants of intra- and intercultural communicative interaction. Interpretation is a dynamic process in which interlocutors interact with one another as described in the XYZ-model. In a face-to-face situation (such as intercultural philosophical dialogue), interaction goes in both directions. Hence, it is a process of mutual attunement (which include contestation and dissensus). If one party is deceased, as is the case with most textual interpretation, the dynamism of attunement is operative in the dialogue among relevant interpreters and commentators.
3.Mutual attunement is operative in the mutually recognizable human practices and in observing family resemblances of human forms of life.
4.Mutual attunement is operative in the choice and revision of quasi-universals and FR-extension; for example, in discussions about FR(philosophia) ⇔ FR(zhexue).
5.There is mutual attunement in the notion of “fitting.”55 The hermeneutic relativity of fitting includes attunement of (meta-)epistemic virtues.
6.The effect of center-periphery forces (or planetarization) may be seen as an enforced and biased mutual attunement of languages in the modern world.
7.Mutual attunement is operative in any reformulation of the subsidiaries of the very principle of mutual attunement. Revision of the formulation of one of the subsidiaries of the principle of mutual attunement for particular interpretative endeavors may be quite substantial.
The concepts used in the formulation of the principle of mutual attunement may not easily allow for FR-extension across traditions. For example, cheng seems to fit with sincerity, but it has also been rendered as truth. This might suggest that a strict separation of truth and sincerity cannot be claimed across traditions.56 As interpretation moves on, or as a result of discussions among scholars, one may reach a stage where a reformulation of (part of) the principle of mutual attunement (the formulation of which is tied to particular CS’s) may be required. It is not assumed that the words used in the formulation of the principle are universals. A possible example of this is not to use “true” in the formulation of the principle of mutual attunement (see next section).
How can one “rationally” revise the formulation of the subsidiaries of the principle of mutual attunement? One can revise one subsidiary of the principle of mutual attunement, while keeping the others constant. This is a variant of Neurath’s boat. As Quine explains (1960: 3): “Neurath has likened science to a boat which, if we are to rebuild it, we must rebuild plank by plank while staying afloat in it. The philosopher and the scientist are in the same boat.” An early version, in terms of the renovation of a vast mansion, can be found in Sima Guang 司马光 (Berthrong: 85).
Hansen (1985) has presented an “argument for urging translators not to use ‘true’ as a translation for any single word of Classical Chinese” (2007: 490n18). In response, Harbsmeier writes: “I have demonstrated through examples that the classical Chinese applied verbs meaning something like ‘be true’ to sentences, and that in this sense they had a concept of semantic truth” (1998: 202; emphasis added). Hansen dismisses this critique, because Harbsmeier’s supporting evidence “consists of supplying a long list of passages of classical texts with his question begging translations” (2007: 490n18). That is to say, Harbsmeier presents translations of passages from classical Chinese using the word true in the English translation. He takes for granted that the reader will agree that his translations are “good” translations. Hansen considers this methodology to be question-begging, because it assumes that “is true” is often a good translation.
Note that the disagreement only concerns semantic truth. Even then the disagreement may not be as large as the rhetoric suggests. A more nuanced Hansen wrote:57
In general, Chinese did reject false propositions and adopt true ones. …
Chinese philosophy seldom employs semantic concepts such as truth [of a sentence], as distinct from pragmatic concepts such as appropriateness of utterance.
Harbsmeier does not address these more modest claims. According to Hansen (504), Chinese has the resources to express the idea of semantic truth: zhen 真 is just such a source. However, there are no theories in classical Chinese in which zhen fulfills that conceptual role. Zhen means “is true,” but in the nonsemantic sense, as in zhen jun 真君 (true gentleman), zhen ren 真人 (true person).58
According to Harbsmeier (1998: 201) and Roetz (1993: 85–96), the range of the truth predicate in modern English is covered in classical Chinese by: ran 然 (that something is so, suchness; be so, be the case), dang 當 (当, adequacy to facts; fit the facts), shi 是 (normative rightness; be this, it, right), shi 實 (accordance with facts; be solid, real, based on fact), xin 信 (veracity; believe; be loyal, trustworthy, reliably so), cheng 誠 (诚, sincerity; be honest, sincere, genuinely so), you 有 (have; exist; there is such a state of affairs), zhen 真 (genuineness; be genuine[ly so]), and perhaps a few more.59 It can be seen that one cannot straightforwardly pair off the European and Chinese FR-concepts. In such a case we may say there is a family resemblance between two clusters: CS(ran, dang, zhen, cheng, …) ⇔ CS(genuinely so, is true, be the case, right …), as already suggested earlier in this chapter. In addition, one may look for one particular FR-extension, which may serve as a quasi-universal to cover (and to compare) both CS’s. Although it is the result of FR-extension, we would consider such a quasi-universal a new hybrid concept. A possible “new” quasi-universal of this kind might be FR(rightness) ⇔ FR(shi 是), to be illustrated in the next subsection.
The fact that there are characters that can be translated in some contexts as “is true” might be taken to support Hansen’s advice that one should never use “is true” in a translation of classical Chinese. Each of the characters in Harbsmeier and Roetz’s lists can perhaps be translated as true in some (rare) contexts, but not one of them has as its core meaning IS TRUE. An alternative translation is always a possibility. Harbsmeier is correct to note that, in classical Chinese texts, one finds theoretical reflection on telling the truth, authenticity, checking and warranting. However, it does not seem to be crucial to use the word true in the translation of supporting passages for these subjects of theoretical reflection.
We think that the debates between Hansen and Harbsmeier (or Roetz) miss important facets of agreement by narrowly focusing on a rigidly defined Western concept. Truth as a property of propositions or sentences was not explicit in Chinese texts, but this did not prevent the occurrence of a wide range of FR-notions, which mean something like right, sincere, authentic, and even “something like ‘be true’” (Harbsmeier 1998: 202). If there is nothing similar at all to any of the various uses of the word true in English, interpretations would surely be next to impossible. But nobody defends such a position.
Harbsmeier remarks concerning his method: “Instead of … taking our notion of truth for granted and asking to what extent the Chinese expressed it, we aspire to start from the Chinese usages surrounding the notion of truth and ask ourselves just how these usages are to be understood” (1998: 201).60 Note that he places the Western notion of truth in the center as a universal reference, whereas Chinese usages merely “surround” it (and can never hit the alleged universal on the head?).61
Some confusion in the discussion may arise because of the relation between the semantic concept of truth and the role of the truth predicate in everyday English. Important features of the role of the truth predicate in everyday English are:62
•It conforms to Aristotle’s intuition.63
•It is used for assent.
•It contributes genuine expressive power when applied to generalizations otherwise inexpressible.
•One expects justification for its use.
However, this is not yet a semantic concept of truth. The latter appeared in Western philosophy only in the 1930s (Tarski 1931). Why then wonder whether it has already occurred in Chinese classics? Graham remarks: “One would not expect another culture, with a language unrelated to the Indo-European family, to have a word with the same range of ambiguities as our ‘true.’ …” (Graham 1991: 296).64 In the West, the word true/truth is extended from the factual to knowledge in general and then to the goal of philosophy. In contrast, the Later Mohists use a number of characters covering the Western being true; and their goal is not Truth, but social order (297). This reinforces our general point that one can always find similarities, that allow FR-extensions, but it makes no sense to look for identity of concepts, in particular not for concepts that already have a stipulated meaning in a Western context.
The predicate “is true” (according to Tarski’s stipulation) applies to sentences. The disagreement between Hansen and Harbsmeier already starts at this level. Harbsmeier says: “Hansen (1985) argues that the ancient Chinese did not have a concept of a sentence. He provides no new philological evidence to support his case” (1998: 183n4).65 Again the disagreement is not as serious as the wording suggests.66 There is a difference between the issue of “having the concept of a sentence” and the issue whether one can parse sentences in the text. The latter is confirmed by virtually all modern translations of classical texts.67 Hansen denies the former.
Harbsmeier’s own philological evidence does not seem to be altogether persuasive. He notes that characters that are sometimes translated as meaning sentence include yan 言, ci 辞, and ju 句, but these characters have other translations as well, as Harbsmeier himself tells us (181–184):68
•ju (phrase, sentence, clause, utterance);
•ci (formulation, sentence);
•yan (speech, sentence, statement).
According to Graham, the Later Mohists use ci (phrase) as a technical term for proposition (or sentence) and he argues that they “discovered the sentence (as Plato did)” (1989: 153–155, 393, 408). Hansen agrees that ci could denote sentences as well as noun-phrases, verb-phrases, and compound terms. Hansen acknowledges that the Mohists did discuss word order and its importance, as Graham shows, but he emphasizes that they did not make a clear distinction between sentences, phrases and compound terms.
Without taking sides in the disagreement between Hansen and Harbsmeier, the fact that such a disagreement is possible and taken seriously is sufficient for suggesting that is true does not easily qualify as a quasi-universal and may require the adjustment of the formulation of our revised principle of charity.
Among the Chinese characters in the cluster that may have a family resemblance with the cluster that contains is true, shifei 是非 in particular, has given rise to conflicting translations by different scholars. Consider the following line from the Xunzi: 69
是是非非謂之知。 非是是非謂之愚。
Reding translates:
To consider what is true as true and what is false as false, that is what is called “knowledge.” To consider what is true as false and what is false as true, that is what is called “ignorance.” (193)
Perhaps one shouldn’t accept this translation. It invites comparison with Aristotle’s definition of truth.70 The latter is not what Xunzi wanted to say, if one takes the context into account. The tendency to translating shi/fei as true/false may be fuelled by the wish to make the text more similar to Western texts, in particular when the subject matter seems to be zhi 知. Regarding Mozi’s three tests, one may also be tempted to translate shi/fei as true/false to make it more similar to Western epistemic virtues. A better translation of Xunzi is (Watson 2003: 27):71
To treat right as right and wrong as wrong is called wisdom. To treat right as wrong and wrong as right is called stupidity.
Perhaps an even better (but more cumbersome) translation is that by Harbsmeier (1993: 12), who speaks of “Xunzi’s puzzling definition”:
Considering this (or what is right (shi)) as this (or: right), and considering not-this (or: what is wrong (fei)) as not-this (or: wrong) constitutes knowing. Considering this as “not-this” and considering what is not-this as “this” is making a fool of oneself.
Note that notwithstanding Harbsmeier’s dismissal of Hansen’s advice, in this passage he avoids (correctly) the use of true/false in the translation. Also note that neither Reding nor Harbsmeier nor Watson raises the issue of alternative translations.
A “fusion” of true/false and right/wrong under the influence of shifei can be found in Wardy’s (97) translation of a passage from the Preface of the Minglitan:72
[Aristotle] distinguished right/true [shi 是] and wrong/false [fei 非], and banished and deluded error so that we might return to the road of unique correctness: it is called “logic”. This speaks of inferential discussion and the principles of names. Its great purpose lies in inferential progression.
Perhaps in this case the motivation to write right/true and wrong/false is to have it both ways: “true” to be true to Aristotle’s text, “right” to acknowledge how Aristotle’s shi 是 would have been understood by the intended audience of the text.
Also consider Roetz’s formulation of his principle of charity as follows:
The very idea that there is a difference between true and false, between right and wrong and the usage of this idea in the theoretical disputes, is the same as ours and certainly constitutes a bridge for communication. (Roetz: 102; emphasis added)
He seems to consider true/right and false/wrong to be the same duality. Perhaps the solution to the problem of translating shifei is to speak of right/wrong throughout and consider true/false a less universal notion, although it plays an important role in the history of Western philosophy. However, perhaps rightness can make a much stronger claim to universality than truth.
Nelson Goodman has argued, “truth is often inapplicable, is seldom sufficient, and must sometimes give way to competing criteria” (1978: 107; cf. 19). That is to say: truth is subordinate to rightness. Goodman gives the following example: The officer ordered the soldier to shoot the prisoners if they were to move. As soon as the officer had finished giving the order, the soldier shot all the prisoners. Being asked to justify his action, he replied that all prisoners were moving around the sun at high speed. In the given situation, the statement of the soldier is true but not right. There are many situations where one speaks of rightness, but cannot speak of truth; for example with respect to orders, validity of (deductive or inductive) inferences, musical performances, and so on (Goodman and Elgin: 155–156).
According to Goodman, a right work of art makes a world, just as a right scientific theory makes a world. Both work of art and theory offer the possibility of new worlds. We add that this is similar for the work of interpretation. An interpretation is also a work that makes a new (Goodmanian) world. Rightness is a matter of fitting and working,
a fitting into a context or discourse or standing complex of other symbols. What counts is not so much the working of what is fitted in as the working of the resultant whole. … The working is also a kind of fitting—fitting into a going operation or process or endeavor. Moreover, even what constitutes fitting and working may undergo change, may itself have to be adjusted in order to fit and work. (Goodman and Elgin: 158–159)
Interestingly, Graham mentions that yi 義 (义 “rightness”) is often explained in terms of yi 宜 (appropriateness, fitting): “Yi [義] is the word cognate with yi 宜 ‘fitting’ which we have elsewhere translated ‘the right.’”73 This suggests that in classical Chinese, as it is the case for Goodman, rightness may be explained in terms of fitting. Graham’s “right as the socially fitting” (1989: 495) fits Goodman’s notion of fitting rather well. When discussing dang 當 as a possible candidate for the function truth has in Western languages, Fraser (2012) also employs the notion of fitting. “Besides utterances, a person’s conduct can be characterized as dang (fitting) or not (both in the sense of fitting the facts and for characterizing a person’s conduct)” (362). Fraser further argues that shi 是 should be understood as a basic conception of correctness that is used when evaluating both empirical facts and moral norms (assuming the latter distinction makes sense in the Chinese context).74 Concerning the Later Mohists, he writes:
Talk of shi–fei distinctions tends to mix descriptive and normative issues. … All judgments—descriptive, prescriptive, or otherwise—[are seen as] as a matter of discriminating shi from fei. These terms refer to right and wrong in an extremely general sense, without distinguishing between … issues falling into areas as diverse as science, politics, ethics, prudence, and etiquette. The Mohists thus seem to be employing a very basic, primitive conception of correctness, of which truth, permissibility, and other normative statuses can be seen as species. (2012: 355–356)
Perhaps even the division of species is mainly a Western projection.
As we have pointed out before, separating truth and rationality from morality and emotionality may be a rather local idiosyncrasy of a few centuries of European history.75 Therefore, it should not be too easily taken for granted that the rightness of shifei is completely different from the rightness of yi 宜.76 The Zhongyong states explicitly: yi zhe yi ye 義者宜也, which Legge (1961) translates as “righteousness is the accordance of actions with what is right.” Later commentators say something like yi (whether 義 or 宜) points to flexible judgment fitting ever-changing situations. We suggest that this applies to shifei as well.
Within the context of Western philosophy we propose the following modest notion of truth to be used in intercultural philosophy:
•Numerous definitions of truth have been proposed in the Western traditions. A satisfying definition of truth cannot be given because which more primitive terms could be used to define true?77 Perhaps in different domains of human activity, different criteria for evaluating the truth or correctness of judgments reflect a locally appropriate definition of truth.
•Truth has something to do with factuality, coherence, relevance to actions and social engagement and similar epistemic virtues, but there is no universal theory that stipulates how to weigh these and other aspects for all possible situations and for all possible groups of humans.
•The FR-concept of being true overlaps with a range of other notions: being right, being so, being correct, being valid, being genuine, being sincere, being real, being appropriate, being essential, being consistent, and so on.
•The stipulated definition of semantic truth as proposed by Tarski (1931) for formal languages is part of an ideal language project, and hence is not useful for intercultural philosophy.
•Rightness and fitting may be the more generic notions and more suitable than the true/false dichotomy for FR-extension across traditions. The features of the “is true” predicate in English, as listed earlier in this section, can easily be reformulated to apply to “is right.”